Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Epigenetic DNA methylation is not associated with natural variation in caregiver-infant physical contact… Quirt, Jill Sutherland 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_quirt_jill.pdf [ 2.41MB ]
Metadata
JSON: 24-1.0105169.json
JSON-LD: 24-1.0105169-ld.json
RDF/XML (Pretty): 24-1.0105169-rdf.xml
RDF/JSON: 24-1.0105169-rdf.json
Turtle: 24-1.0105169-turtle.txt
N-Triples: 24-1.0105169-rdf-ntriples.txt
Original Record: 24-1.0105169-source.json
Full Text
24-1.0105169-fulltext.txt
Citation
24-1.0105169.ris

Full Text

EPIGENETIC DNA METHYLATION IS NOT ASSOCIATED WITH NATURAL VARIATION IN CAREGIVER-INFANT PHYSICAL CONTACT TIME IN A CROSS-SECTIONAL SAMPLE OF HUMAN TODDLERS by Jill Sutherland Quirt B.Sc., Queen’s University, 2009  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Experimental Medicine) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2012 © Jill Sutherland Quirt, 2012  Abstract Background: Early infant experience has been associated with long term epigenetic modifications of gene expression and subsequent health outcomes. We questioned whether daily physical contact time shared between an infant and their caregiver as an early experience was associated with prolonged variations in the epigenetic patterns of toddlers. Methods: We performed a second look analysis of a self-reported 3-day diary kept by 1,055 caregivers of their behaviours and interactions with their infants at 5 weeks of age. We measured the average daily physical contact time shared between the infant and caregiver and compared this across the sample. We defined high and low caregiver-infant contact groups as the upper and lower 16th percentiles. We conducted a descriptive analysis of the associated caregiver and infant behaviours with high and low contact caregiving. We then collected epithelial buccal cells from a subsample of the now 3-5 year old toddlers of the high and low caregiver-infant pairs. We collected samples for 98 toddlers (59 high contact infants, 39 low contact infants). We determined DNA methylation patterns in all toddlers using the Infinium Illumina Methylation Assay and then compared the degree of methylation at 434580 CpG sites. In addition, we compared methylation status of five genes associated with stress response pathways. Results: Average daily caregiver-infant contact time followed a normal distribution over a large range. The mean physical contact time between caregiver-infant dyads was (+ SD) 9h7m (+2h 52m). High and low contact caregiver-infant contact pairs were defined as those who shared >12h 22m/day and those who shared <6h14m/day of physical contact respectively. There were no significant associations between the methylation patterns of  	
    ii	
    the epithelial buccal cells of the toddlers and the differences in physical contact time at 5 weeks of age. There were no significant associations between the methylation status and physical contact time when controlling for gender. No differences were associated with the 5 candidate genes. Interpretation: A difference in caregiver-infant physical contact time of at least 6hrs/day at 5 weeks of age was not associated with individual differences in epigenetic methylation patterns in the epithelial buccal cells of toddlers at 3-5 years of age.  	
    iii	
    Preface The following research has been approved by the UBC Research Ethics Boards as a portion of the study: “The CARE Project: Study 1” (H07-01317). The study presented here is one component of the collaborative project of Dr. Ronald Barr, Dr. Thomas Boyce, Dr. Wendy Robinson, Dr. Michael Kobor and Dr. Moshe Szyf. I was involved in study development and responsible for ethics approval, protocol development and implementation. I conducted and/or supervised participant recruitment and DNA collection. I was responsible for data interpretation. Illumina Infinium Methylation Assay protocol, filtering, imputation and quantile normalization of data was conducted in the Kobor Lab by Sarah Neumann and Lucia Lam using standard ‘R’ 2.12.0 software.  	
    iv	
    Table of Contents Abstract .........................................................................................................................................ii Preface..........................................................................................................................................iii Table of Contents......................................................................................................................... vi List of Tables .............................................................................................................................viii List of Figures .............................................................................................................................. ix Acknowledgements...................................................................................................................... xi Dedication ..................................................................................................................................viii 1  Introduction ......................................................................................................................... 1 1.1 Summary of research ........................................................................................................ 1 1.2 Overview and objectives................................................................................................... 3 1.3 Background and current state of knowledge..................................................................... 5 1.3.1  Individual differences in stress response .............................................................. 5  1.3.2  The long-lasting influence of early life experiences: contributions of animal models ....................................................................................................... 7  1.3.3  Translation to understanding early life influences on human physiology and behaviour...................................................................................................... 10  1.3.4  2  Epigenetic regulation of the genome .................................................................. 11  1.3.5  DNA methylation regulation of the genome............................................................. 11  1.3.6  Epigenetic regulation of candidate genes ................................................................. 13  1.3.7  Early experiential influence on DNA methylation in humans.................................. 18  1.3.8  Summary and introduction to studies ....................................................................... 20  Study 1: Human infant early experience mediated by caregiver-infant interactions............ 22 2.1 Study 1a: Caregiver-infant physical contact time at 5 weeks of age .............................. 23  	
    v	
    2.1.1  Objective ............................................................................................................. 23  2.1.2  Background and rationale ................................................................................... 23  2.1.3  Aim ..................................................................................................................... 24  2.1.4  Methods............................................................................................................... 24  2.1.5  Results................................................................................................................. 25  2.1.6  Discussion ........................................................................................................... 26  2.2 Study 1b: Demographic and associated caregiving behaviour differences in high and low contact caregiving dyads ................................................................................... 28  3  2.2.1  Objective ............................................................................................................. 28  2.2.2  Background and rationale ................................................................................... 28  2.2.3  Aim ..................................................................................................................... 28  2.2.4  Methods............................................................................................................... 28  2.2.5  Results................................................................................................................. 29  2.2.6  Discussion ........................................................................................................... 32  Study 2: DNA methylation patterns associated with high and low caregiver-infant physical contact time............................................................................................................. 34 3.1 Objective ......................................................................................................................... 34 3.2 Background and rationale ............................................................................................... 34 3.3 Aim ................................................................................................................................. 35 3.4 Method ............................................................................................................................ 36  	
    3.4.1  Participants.......................................................................................................... 36  3.4.2  Participant recruitment........................................................................................ 36  3.4.3  Participant sample collection .............................................................................. 37  vi	
    3.4.4  Epithelial buccal swab collection and DNA extraction ...................................... 38  3.4.5  Illumina Infinium Methylation Assay................................................................. 38  3.4.6  Data filtering, normalization and statistical analysis of DNA methylation data ..................................................................................................................... 39  3.5 Results: Demographic and associated caregiving behaviour differences in high and low contact caregiver-infant dyads .......................................................................... 41 3.5.1  Demographic comparison ................................................................................... 41  3.5.2  Infant behavioural characteristics ....................................................................... 44  3.5.3  Caregiver behavioural characteristics ................................................................. 46  3.5.4  Discussion of demographic and associated caregiving behaviour differences in high and low contact caregiver-infant dyads results.................... 47  3.6 Results: DNA methylation pattern differences in high and low contact caregiverinfant dyads..................................................................................................................... 49 3.6.1  Interpretations of Illumina Infinium Methylation Assay results......................... 49  3.6.2  Differential methylation associated with differences in physical contact time between caregiver-infant dyads .................................................................. 53  3.6.3  Differential methylation associated with contact time: linear regression analysis................................................................................................................ 60  3.6.4  Differential methylation associated with contact time: gender specific analysis................................................................................................................ 63  	
    3.6.5  Gene specific analysis......................................................................................... 66  3.6.6  Discussion of DNA methylation results ............................................................. 69  vii	
    3.7 Summary of Study 2: DNA methylation patterns associated with high and low caregiver-infant physical contact time ............................................................................ 69 4  Concluding chapter ............................................................................................................... 70 4.1 Discussion ....................................................................................................................... 70 4.1.1  Convergence with past findings.......................................................................... 70  4.1.2  Age-associated DNA methylation changes in humans....................................... 72  4.1.3  Tissue specific DNA methylation changes in humans ....................................... 74  4.1.4  Definition of early experience ............................................................................ 76  4.2 Limitations ...................................................................................................................... 78 4.3 Future studies .................................................................................................................. 84 4.4 Summary of research ...................................................................................................... 86 Bibliography ......................................................................................................................... 89 Appendix A: Supplemental Illumina Infinium Methylation Assay results for high compared to low physical contact time................................................................................. 97 Appendix B: Supplemental results for all CpG sites associated with targeted candidate genes ..................................................................................................................... 99  	
    viii	
    List of Tables Table 2.1: Demographic comparison of high and low contact participants from PHI study..... 29 Table 3.1: Comparison of consented participants and the non-consented participants from the eligible sample.............................................................................................................. 43 Table 3.2: Comparison of high contact participants and the low contact participants .............. 44 Table 3.3: Comparison of infant and caregiver behaviours found to be significantly higher in high caregiver-infant dyads ......................................................................................... 48 Table 3.4: High contact vs. low contact 6 CpG sites with the lowest p-value by both Ttest and Wilcoxon test................................................................................................................. 54 Table 3.5: High contact vs. low contact 5 CpG sites with the lowest T-test p-value (p < 0.001), and an average beta value > 5% ..................................................................................... 58 Table 3.6: Summary of linear regression and Two-Way ANOVA results for the 7 CpG sites with the lowest regression p-value for caregiver-infant contact time as determined by a regression analyses.............................................................................................................. 61 Table 3.7: Summary of Two-Way ANOVA results for the 12 CpG sites with the greatest gender-contact interaction........................................................................................................... 64 Table 3.8: Summary of T-test results for the CpG site with the smallest p-value by T-test for each of the seven targeted candidate genes........................................................................... 67  	
    ix	
    List of Figures Figure 1.1: A schema outlining the function of the hypothalamic-pituitary-adrenal axis........... 9 Figure 2.1: Histogram of caregiver-infant body contact duration, grouped by 20 minute intervals....................................................................................................................................... 25 Figure 2.2: Histogram of caregiver-infant body contact duration, grouped by 20 minute intervals with high and low contact groups defined ................................................................... 26 Figure 2.3: Comparison of rodent LG and human caregiver-infant contact distributions......... 27 Figure 2.4: PHI high and low contact groups infant behaviours frequencies............................ 30 Figure 2.5: PHI high and low contact groups infant behaviours durations ............................... 31 Figure 2.6: PHI high and low contact groups caregivers mean infant pick-ups per day ........... 31 Figure 2.7: PHI high and low contact groups caregiver durations of holding the infant........... 32 Figure 3.1: Selection of participants for inclusion in the Study 2 ............................................. 42 Figure 3.2: High and low contact groups infant behaviours frequencies .................................. 45 Figure 3.3: High and low contact groups infant behaviours durations...................................... 46 Figure 3.4: High and low contact groups caregivers average infant pick-ups per day.............. 46 Figure 3.5: High and low contact groups caregiver durations of holding their infant............... 47 Figure 3.6: Comparison of the top 6 cytosine-guanine dinucleotide (CpG) sites associated with low and high caregiver-infant physical contact time as determined by Ttest (p<0.0001) and Wilcoxon analyses (p<0.0001) ................................................................... 55 Figure 3.7: Comparison of the top 5 cytosine-guanine dinucleotide (CpG) sites with a DNA methylation difference of >5% (>0.05 average beta) between low and high Contact groups, and the lowest p-value (p<0.001) as determined by a T-test ......................................... 58  	
    x	
    Figure 3.8: Comparison of the 7 cytosine-guanine dinucleotide (CpG) sites with the lowest p-value for regression analyses by contact time.............................................................. 62 Figure 3.9: Comparison of the top 12 cytosine-guanine dinucleotide (CpG) sites with significant contact group and gender interaction as determined by two-way ANOVA (p<0.0001)................................................................................................................................... 65 Figure 3.10: Comparison of the top cytosine-guanine dinucleotide (CpG) site for each targeted gene of interest .............................................................................................................. 68 Figure 4.1: Glucocorticoid receptor gene exon 1F promoter region .......................................... 82	
    	
    xi	
    Acknowledgements I offer my gratitude to the faculty, staff and my fellow students at the University of British Columbia who have inspired me to continue my work in the field of Experimental Medicine. I owe particular thanks to Dr. Ronald Barr for his commitment to teaching me, for his patience, and for his ongoing support throughout this study as my primary supervisor. I thank Dr. Thomas Boyce, Dr. Wendy Robinson, and Dr. Michael Kobor for their guidance, insight and continued help throughout my Masters degree. A special thanks is due to Lucia Lam and Sarah Neumann who patiently guided me through the new field of epigenetics, helping me sort through and interpret endless numbers of results. Thank you to all the past and present members of the Barr Lab. It is a pleasure working with you each day. I’d like to thank my parents for their endless support throughout all my years of education. Finally, a deep felt thank you to Duncan for the love and support he shows me each day.  	
    xii	
    Dedication  I would like to dedicate this thesis to my wonderful family.  	
    xiii	
    1.	
    Introduction  1.1  Summary of research Early experiential factors can have a profound influence on behaviour, physiology  and health of the individual as a child and adult. A mechanism through which early experience can affect long-term health outcomes is through epigenetic modification of gene expression. Epigenetics is a stable modification of gene expression through changes in DNA structure that does not involve the underlying nucleotide base sequence. There are several examples of early experiential influences having long term effects on behavioural and physiological outcomes that have been mediated by epigenetic modification of gene expression. In rodent studies, naturally occurring differences in maternal care were associated with epigenetic DNA methylation differences and subsequent endocrine and behavioural responses in the offspring. Similarly, changes in DNA methylation have been associated with adverse early experiences, such as childhood abuse, in humans. These findings and others have contributed to our understanding of possible relationships between developmentally early experiences and subsequent epigenetic patterns, and specifically DNA methylation. In this study, we demonstrate that the daily physical contact time a human caregiver and infant share has a normal distribution with a large range in a human sample. Furthermore, we describe the typical differences in caregiver-infant interactions and behaviours associated with differences in physical contact time. We demonstrate that distribution of daily physical contact time in humans is similar to an analogous distribution described in rodent studies investigating DNA  	
    1	
    methylation patterns associated with variation in mother-offspring interactions. We capitalize on the analogous nature of our daily physical contact time distribution amongst human caregiver-infant pairs to explore the DNA methylation patterns associated with higher and lower amounts of daily physical contact time in early human life. We investigate whether these variations in caregiver-infant contact time are associated with lasting differences in epigenetic regulation of gene expression in peripheral epithelial buccal cells. Specifically, we determine whether normal variation in caregiver-infant physical contact time at 5 weeks of age is reflected in prolonged individual differences in epigenetic methylation patterns, as evidenced in the epithelial buccal cells of toddlers at 3-5 years of age. We capitalize on a previously collected, detailed phenotypic description of an early caregiving behaviour (namely, caregiver-infant physical contact time) and our ability to characterize large numbers of epigenetic markers of genes. Our study will be performed under the assumption that early experience in the form of differences in daily physical contact time is associated with epigenetic change in multiple tissues and sites throughout the genome. Thus, associated DNA methylation patterns with physical contact time are likely to be observable in various tissues including peripheral epithelial buccal cells. In this study, we test a hypothesis that a difference in early life experience, namely average daily physical contact time between caregivers and infants, will be associated with differences in the methylation patterns later in life as a toddler. In addition, we test a secondary hypothesis that a difference in caregiver-infant physical contact time will be associated with differences in the methylation status of genes  	
    2	
    implicated in human stress response pathways; namely NR3C1, DRD4, 5HTT, COMT and BDNF. We also assess whether variation in physical contact time in early life results in differential methylation in males and females. Both positive and negative findings will contribute to increasing our understanding of the relationship between early life experiences and later epigenetic methylation patterns in humans.  1.2  Overview and objectives The long-term objective of this study is to contribute to the understanding of the  influence of early experience on human infant development. Early developmental experiences can influence later childhood and adult behaviour and health. For example, low quality parental care can lead to the dysregulation of neurobiological stress systems in offspring1. The hypothalamic-pituitary-adrenal (HPA) axis has been studied as a responsive system that is affected by differences in early experience, and as a mediator in the relationship between early environmental exposures and long term health outcomes in adulthood2. Activation of the HPA axis due to exposure to stress results in the release of stress hormones and subsequent alterations in metabolic, cognitive and emotion regulatory processes and engagement of the hypothalamic negative feedback loop to restore the HPA pathway to basal levels3. A major mechanism by which environment may cause long term cellular changes within an individual is through epigenetic programming. Epigenetics refers to the stable modification of DNA and chromatin that can affect gene expression, but that does not involve changes to the underlying nucleotide base sequence of the DNA. The best understood epigenetic modification is DNA methylation, which in mammals occurs  	
    3	
    primarily at cytosines located 5’ to guanosine in a CpG dinucleotide4. Extensive DNA methylation is associated with chromatin structure, chromosome stability, transcriptional silencing of imprinted genes, and the maintenance of the transcriptionally inactive X chromosome in females5. There are several examples of early developmental experiences that have been shown to be associated with epigenetic modification of genetic expression in both human and nonhuman studies. For example, natural variations in certain maternal caregiving behaviours in rodents has been associated with variation in DNA methylation status of the glucocorticoid receptor gene of their offspring and the development of behavioural and endocrine responses to stress in adult life6. Human studies have demonstrated similar relationships between early experience and DNA methylation. For example, a decreased level of glucocorticoid receptor expression mediated by increased methylation of the NR3C1 promoter region has been observed in victims of childhood abuse7. Previous human and nonhuman epigenetic studies such as the ones mentioned briefly above suggest relationships between early experience and later DNA methylation. These studies suggest that there may be significant differences in the methylation patterns that lead to differences in gene expression and offspring phenotype, and that these differences may be elicited by differences within the normal range of some caregiverinfant interactions. However, -­‐  We know little in terms of the normal range of variability in human caregiverinfant interactions and the caregiver and infant behaviours that are associated with differences in physical contact.  	
    4	
    -­‐  It is unknown if caregiver-infant physical contact time is associated with prolonged changes in epigenetic methylation patterns.  -­‐  It is unclear whether epigenetic changes associated with early experiences are observable in peripheral epithelial buccal cells. Therefore, in this study, we propose to investigate whether the amount of physical  contact time between caregiver and infant are associated with later differences in epigenetic patterns among toddlers. The findings will contribute to understanding whether, and if so, in what way early caregiving experiences may be associated with later DNA methylation in humans.  1.3  Background and current state of knowledge  1.3.1  Individual differences in stress response Responses to stress are usually adaptive. The perception of a stressor results in  neural signals that increase the release of stress hormones from the adrenal glands and sympathetic nervous system into the bloodstream. The actions of these hormones that include glucocorticoids and catecholamines (i.e. norepinephrine) act to increase the availability of energy substrates in an effort to maintain normal cellular and organ function. For example, catecholamines act in the brain to increase vigilance and provoke a state of fear, avoidance learning and fear conditioning8, thereby reducing the likelihood of subsequent encounters with the same stressor. In short, defensive responses are the logical outcomes of stress-induced change in brain and associated endocrine activity, and are defined by an increase in synthesis and release of stress hormones8,	
  9. Although protective in theory, chronic stimulation of the defensive response system can result in negative outcomes. Persistent activation of these responses may  	
    5	
    result in chronically enhanced emotional arousal, sustained increases in blood sugars and fats, potential for hyperinflammation, and disruption of sleep and normal cognitive function among other effects10. This can predispose to disease such as diabetes, obesity, mood disorders, and heart disease11. Similarly, insufficient activation of defensive responses in conditions of threat can compromise health outcomes and be associated with chronic fatigue, chronic pain, posttraumatic stress disorder, and hyperinflammation12,	
  13. Therefore there seems to be a balance between the positive and negative effects of the stress response in terms of eventual good or bad outcomes for an individuals’ health. ‘Stress diathesis’ models have been used to describe the interactions between development and the prevailing level of stress in predicting health outcomes14,	
  15. Such models explore the origins of illness and the nature of underlying vulnerabilities, linking psychological and biological determinants of health. The core concept is that adversity in early life modifies the development of neural and endocrine responses to stress in such a way as to predispose an individual to later disease or adverse health outcomes16. The stress diathesis concept postulates that infants may not be simply passive recipients of their early environmental experience, but that early experience as an infant may predispose them to respond to later life stressful experiences in ways that may have good or bad outcomes. Early postnatal life is a period when external stimuli, whether permissive or aversive, can influence emotional and cognitive development in human offspring15,	
  17. While dramatic early adversity such as sexual, emotional and physical abuse are reasonably seen to have significant long-term consequences, lasting influences on the life course have also been reported from co-occurring day-to-day early childhood experiences, such as low parental investment18. The quality of early experience is largely  	
    6	
    mediated by parental investment, such as behavioural interaction and nutrient supply. While the forms of variation of parental investment will differ, there may be common effects on phenotypic plasticity that in each case can affect the development of defensive responses and reproductive strategies6,	
  19. Evidence with respect to adaptation of the stress response to differences in early developmental experiences comes from both human and nonhuman research. A brief review of relevant literature as it relates to both human and nonhuman research follows. 1.3.2  The long-lasting influence of early life experiences: contributions of animal  models Psychologists Levine and Denenberg were among the first to demonstrate the potential for differences in early experiences to influence the subsequent development of rudimentary defensive responses when they reported that postnatal handling of infant rodents dampened the magnitude of behavioural and hypothalamic-pituitary-adrenal responses to stress in adulthood20,	
  21. More recently, Meaney and colleagues have observed that laboratory rats exhibit a range of maternal caregiving behaviours of archedbacked nursing (ABN) and licking and grooming (LG). Simply examining the typical range in pup LG among lactating rats without any experimental manipulation, they reported considerable differences in the frequency of pup LG, and that these differences remain stable throughout the reproductive life of the female22. These individual differences within the normal range influence the development of the neural substrates that underlie the phenotypic behavioural and endocrine responses to stress in the offspring6,	
  17,	
  19. Compared with adult offspring who received high levels of maternal LG in the postnatal period, the adult offspring of mothers who provided lower levels of LG  	
    7	
    showed heightened behavioural stress reactivity, including less open-field exploration and longer delays in eating food when presented in a novel environment6,	
  19. Furthermore, these behavioural differences were accompanied by corresponding neuroendocrine responses such as increased plasma adrenocorticotropic hormone levels, and decreased sensitivity to the inhibitory effects of glucocorticoids during conditions of acute stress17. Conversely, offspring of mothers exhibiting high levels of pup LG demonstrated more modest behavioural and endocrine responses to stress as compared to those reared by low-LG mothers6. The activation of behavioural, emotional, autonomic, and endocrine responses to stressors is mediated in part by the corticotrophin releasing factor (CRF) system of the Hypothalamic-Pituitary-Adrenal (HPA) axis23. CRF acts at CRF1 receptors to stimulate the release of norepinephrine in various corticolimbic structures23. In a negative feedback loop, circulating glucocorticoids act at receptor sites in corticolimbic structures, inhibiting hypothalamic CRF expression. Therefore, circulating glucocorticoids have a significant influence on the regulation of HPA activity. Figure 1.1 is a schematic representation of the HPA axis and corticotrophin-releasing factor (CRF) feedback loop.  	
    8	
    	
   Figure 1.3:A schema outlining the function of the hypothalamic-pituitary-adrenal axis. CRF is released into the portal system of the anterior pituitary glad, stimulating the synthesis and release of adrenocorticotrophin (ACTH). ACTH then stimulates adrenal glucocorticoid release. Glucocorticoids act on glucocorticoid receptors throughout regions of the brain, including the hippocampus, to inhibit the release of CRF in a negative feedback loop. Note: As shown schematically on the right, in adult offspring of high-LG mothers compared to low-LG, Meaney and colleagues observed (i) increased glucocorticoid receptor expression, (ii) enhanced negative-feedback sensitivity to glucocorticoids, (iii) reduced CRF expression in the hypothalamus and (iv) more modest pituitary-adrenal response to stress17,	
  19,	
  24,	
  25. This schema is adapted from a previous publication of Meaney26.  Alterations in the CRF systems in select brain regions were observed in rats relating to the effects of maternal care on the development of defensive responses to stress. Offspring of high-LG mothers demonstrated decreased CRF expression in the hypothalamus, decreased levels of CRF1 receptor and reduced plasma adrenocorticotropin (ACTH), as well as increased hippocampal glucocorticoid receptor expression6,	
  17,	
  19. Therefore, high-LG offspring had enhanced negative-feedback sensitivity, decreased CRF levels and, as a result decreased glucocorticoid responses to acute stress as adults compared to the offspring of low-LG mothers17,	
  19. Low-LG offspring had significantly higher stress-induced increases in norepinephrine stimulated by CRF26. In short, increased pup LG was associated with increased effectiveness of the negative-feedback system. This is reflected in differences in offspring behavioural responses to stress. As adults, offspring of high-LG mothers showed a decreased startle response, substantially less fearfulness in the presence of stressors such as a novel 	
    9	
    environment, and increased willingness to move out and explore a new environment without fear6,	
  19. In cross-fostering studies with rat pups, postnatal development of behavioural and HPA stress response pathways of the offspring were associated with the maternal behaviour of the rearing mother as opposed to the biological mother27. 1.3.3  Translation to understanding early life influences on human physiology and  behaviour The research program of Meaney and his colleagues has demonstrated the longterm regulatory differences of high levels of early maternal care: pups who received higher levels of LG demonstrated enhanced ability to cope with later life stress. There is some evidence that analogous relationships between maternal care and stress response exist in humans. Human adults who reported extremely low-quality relationships with their parents were found to have significantly higher release of dopamine during stressful events as compared to those that reported extremely high-quality relationships with their parents28. Furthermore, elevated concentrations of salivary cortisol levels have been associated with a disorganized attachment system between mother and infant29. Hane and Fox reported that infants who experienced low-quality maternal care behaviour (relative to those receiving high quality care) displayed significantly more fearfulness during the presentation of novel stimuli and less positive joint attention (an index of infants’ positive sociability)1. Such evidence suggests that early human caregiving may affect the later development of stress reactivity systems. A consistent theme reflected in these studies is that various experiential signals can alter phenotypic outcomes through parent-offspring interactions. The relevant form  	
    10	
    of the variation in parental-offspring interaction varies among species, however the principle idea is that of early experientially mediated effects on phenotypic outcome. What remains to be determined are the mechanisms by which environment influences phenotype of the individual. 1.3.4  Epigenetic regulation of the genome Physical modifications of structural and chemical properties of DNA result in  changes in genomic expression. Genetic expression may be altered through modifications to the DNA and chromatin structure without changes at the nucleotide level. This layer of additional regulation constitutes the Epigenome. The Epigenome is the “stable alterations in gene expression by non-genomic mechanisms, resulting in stable alterations in phenotypes”30. Epigenetics is the study of this level of gene regulation, its interaction with the genome, and the translation of the genome and epigenome into phenotypic variability. The essential features of epigenetic mechanisms are (i) structural modification to chromatin (at the level of histone proteins or DNA), (ii) regulation of chromatin structure and function, (iii) effects on gene expression, and (iv) these effects occur without any change to the nucleotide sequence26. DNA methylation is perhaps the best understood epigenetic modification that modifies genetic expression. 1.3.5  DNA methylation regulation of the genome DNA methylation is an active and stable biochemical modification associated  with the silencing of gene transcription31-­‐33. In mammals, methylation is achieved through the actions of DNA methyltransferases that actively target cytosine nucleotide bases. The addition of a methyl group at cytosine-guanine dinucleotide (CpG) sites (over represented in the CpG-regions of promoter regulatory regions of many genes) displaces  	
    11	
    proteins that results in gene silencing4. Methylation DNA modification is a bidirectional enzymatic reaction34. The balance of methylation and demethylation results in fluctuations in gene silencing and expression. Gene silencing by DNA methylation occurs in one of two ways. First, widespread methylation of a DNA region precludes transcription factor binding thereby silencing the expression of genes in that region of DNA. Second, selective methylation of cytosines in the promoter regions of genes attracts methylated-DNA binding proteins35, which then attract repressor complexes that actively mediate gene silencing. For example, gene silencing through selective methylation can be caused by histone deacetylase complexes (HDAC: a component of the repressor complexes) that prevent the acetylation of histone proteins that results in the chromatin favoring a closed state preventing transcription factor binding and therefore gene expression. One common example of gene silencing by DNA methylation is the silencing of the X-chromosome. The inactivation of one copy of the X-chromosome occurs in all mammalian females and is mediated by DNA methylation36. The inactivation of the Xchromosome in females occurs very early in life and remains stable throughout the life of that female. As a result, researchers in the field of epigenetics believed that DNA methylation occurred early in embryonic development during periods of cell division and that DNA methylation was likely irreversible. However, although as with the Xchromosome silencing, some genomic regions may undergo prenatal DNA methylation, DNA methylation can also occur later during postnatal life24,	
  37,	
  38. In recent studies, DNA methylation patterns have been described as being actively modified in mature cells and these modifications appear to occur in response to cellular signaling induced by  	
    12	
    environmental factors39-­‐41. For example, mature lymphocytes42 and neurons37 have shown changes in DNA methylation patterns associated with environmental stimuli that stably altered cellular function. Epigenetic DNA modification is therefore a candidate mechanism by which early experiences may influence gene expression. 1.3.6  Epigenetic regulation of candidate genes DNA methylation of genes associated with the HPA stress response pathway are  of particular interest given their direct role in the regulation of individual stress responses, and the previously described association with early developmental experiences. The following is a brief literature review of five genes associated with the HPA pathway and DNA methylation epigenetic regulation. 1.3.6.1 Glucocorticoid Receptor Gene (NR3C1) Levels of hippocampal glucocorticoid receptor regulate the HPA system response to stress through a negative feedback relationship with higher levels of glucocorticoid receptor associated with attenuated stress responsivity43. The glucocorticoid receptor gene in the rat is similar to that in the human44. In rodents, the glucocorticoid receptor (Nr3c1) gene contains 9 exons. Exon 2-9 are coding regions for mRNA while there are multiple alternative exon 1s within the promoter region, the activation of which alters the level of gene transcription45. This structure is similar in the human glucocorticoid receptor gene (NR3C1)44. The different promoter regions (associated each with exon 1) are associated with the activation of the NR3C1 gene in different cell types45. Therefore, the NR3C1 gene is potentially active in various cells of the human body, with different functions across cell types. Furthermore, it is possible to increase the expression of the  	
    13	
    NR3C1 gene in one tissue type while its expression in another type is unaltered, or even decreased. Weaver et al. reported that high-LG in rats was associated with decreased methylation at the Nr3c1 17 promoter site in rodents24. The exon 17 promoter region for the rat Nr3c1 gene binds the transcription factor: nerve growth factor-inducible factor A26 (NGFI-A). Offspring of low-LG mothers had reduced binding in the hippocampus of NGFI-A to the promoter region24. Through a cross-fostering study in which the offspring of low-LG mothers were raised by high-LG mothers and vice-versa, it was determined that the variations in maternal care corresponded with the methylation status of the offspring independent of the LG behaviour of their biological mother24. This suggested that the maternal behaviour served as a mechanism for the nongenomic transmission of differences in stress reactivity across generations. In a different study in humans, the methylation status of the NR3C1 gene in cord blood samples of the infants of mothers who had suffered depression while pregnant was compared to that of the infants of mothers who had not suffered from depression during pregnancy. Oberlander and colleagues observed that an increase in methylation of the exon 1F region of glucocorticoid receptor gene in newborns was associated with prenatal exposure to maternal depression as well as with increased cortisol stress responses in the infants at age 3 months46. The human exon 1F region of the NR3C1 gene in humans is analogous exon 17 region in rodents, and includes the NGFI-A transcription factorbinding site. Considering the role of the glucocorticoid receptor in the regulation of the HPA pathway, as well as the work by Meaney and colleagues regarding the methylation status  	
    14	
    of glucocorticoid receptor 17 promoter region and variations in maternal care in rodents, the NR3C1 is a clear candidate gene for DNA methylation regulation by early developmental experiences. 1.3.6.2 Serotonin Neurotransmitter (5-HT, SLC6A3, SLC6A4) In another line of evidence, a number of authors have described results consistent with the involvement of methylation mechanisms in the development of fear responses in children. The development of emotion and fear regulatory regions of the brain, including the amygdala, are associated with later development of affective disorders47,	
  48. Regulation of emotional states is a primary function of the serotonin neurotransmitter, 5HT49. The promoter region of the SLC6A4 gene encoding for the 5-HT transporter (5HTTP) is polymorphic, and is at least weakly associated with disorders of emotional function such as depression and anxiety50,	
  51. 5-HTTP polymorphisms and methylation status has been associated with variation in infant temperament and novel reactions52 and long term mediation of the expression of fear53. For example, in an Iowa adoptee sample by Philibert and colleagues have shown that early abusive experiences are associated with epigenetic modifications54,	
  55. Specifically, methylation status of the CpG islands at the 5HTT was related to reports of child abuse56 and the level of transporter expression was inversely related to the degree of methylation55. Furthermore, higher levels of 5HTTLPR methylation were associated with increased risk of unresolved responses to loss or similar trauma in those who were carriers of the 5 HTTLPR ll allele, a usually protective variant57. In addition, prenatal maternal depression has been demonstrated to be associated with methylation patterns of SLC6A4 promoter encoding the transmembrane serotonin transporter58,	
  59. Observations  	
    15	
    such as these suggest that there may be an association between early experience and the methylation status of genes encoding the serotonin neurotransmitter. 1.3.6.3 Dopamine Receptor gene (DRD4) The catecholamine neurotransmitter dopamine has been related to social interaction and brain development. Dopamine dysfunction in the frontal lobe of the brain is associated with problems in attention, hedonic activities, cognitive processes, working memory and social functioning60. Furthermore, the D4 dopamine receptor (DRD4) gene has been considered as a candidate gene for infant attachment behaviour as it is preferentially expressed in the brain regions of the mesocorticolimbic dopamine pathway mediating reward related to social interactions, including mother-infant attachment61. In a study by Wong et al., monozygotic twins were observed to have high degrees of discordance in DNA methylation when raised in different environments62. The results of this study suggest that variation in DRD4 DNA methylation is at least in part attributable to environmental factors. 1.3.6.4 Catechol-O-methyltransferase gene (COMT) Catechol-O-methyltransferase (COMT) regulates the homeostatic levels of neurotransmitter dopamine in the synaptic cleft63. Although two known isoforms exist [membrane-bound COMT (MB-COMT) and soluble COMT (S-COMT)], MB-COMT is the predominant form associated with the degradation of synaptic dopamine in the human brain64,	
  65 and is regulated by a distinct promoter region64. MB-COMT hyperactivity results in increased synaptic dopamine degradation and is associated with disturbances in attention, executive cognitive and working memory performance both in normal samples and in psychopathologies such as schizophrenia66. Abdolmaleky and colleagues have  	
    16	
    provided some of the first comprehensive molecular explanations at the genetic and epigenetic levels for a relationship between risk for schizophrenia and bipolar disorder and hyperactivity of MB-COMT67. They observed MB-COMT hypomethylation/overexpression in the left brain of schizophrenia and bipolar disorder patients67. Hyperactivity of MB-COMT could lead to a significant increase in the rate of dopamine degradation, resulting in hypodopaminergic state and frontal lobe hypoactivity67. Furthermore, Abdolmaleky and colleagues found that while schizophrenia and bipolar disorder patients had hypomethylated MB-COMT promoter regions, there was a predominant unilateral partial methylation in control subjects67. This may occur in response to the effects of other neuronal pathways and genes involved in brain laterality, or may be the result of a physiological feedback response to environmentally mediated MB-COMT over-expression67. It is possible that the variation seen in psychopathologies of the methylation state of the MB-COMT promoter region could be a consequence of poorly established methylation patterns during critical stages of development, or it may be due to changes in physiological feedback modulation of the MB-COMT. 1.3.6.5 Brain derived neurotrophic factor (BDNF) Brain derived neurotrophic factor (BDNF), encoded by the BDNF gene, contributes to neuronal activity-dependent processes, and is a leading candidate cellular factor for associative memory function68. BDNF gene expression in the hippocampus is induced following contextual and spatial learning, the mechanism being essential for normal learning and memory development69,	
  70. Further implicating a role for BDNF signaling in the stress response pathway, it has been shown that decreases in hippocampal BDNF levels are correlated with stress-induced depressive behaviours71. Within the  	
    17	
    BDNF gene, neuronal activity-dependent regulation of specific promoter regions dictates the spatial and temporal expression of specific transcript isoforms72, which can then regulate the subsequent trafficking and targeting of the transcript73. Alteration of the methylation state of the BDNF promoter region is associated with contextual fear in the conditioning of adult rats68. The polymorphic nature of the BDNF gene suggests a natural flexibility in expression in response to environmental stimuli that could be important in the development of learning and memory in children. 1.3.7  Early experiential influence on DNA methylation in humans As a result of the stability of DNA methylation, methylation patterns may endure  beyond the period of differential early experiences and therefore may provide a molecular basis for stable environmental effects on the phenotype of the offspring. Therefore, prenatal and early postnatal epigenetic programming resulting from environmental signals could define life-long gene expression trajectories that continue throughout life and affect health outcomes of the individual. An example of this is a study completed on individuals born during the Dutch Hunger Winter of 1944-45. Due to nutritional restriction during gestation, those who were prenatally exposed to the famine had less DNA methylation of the imprinted Insulin-like growth factor II (IGF2) gene compared with their unexposed, same-sex siblings74. While this was likely in preparation for the environment of food scarcity and hardship that the offspring were entering into, the abundant nutritional environment of the postwar years in fact led to decreased IGF2 expression and a corresponding increase in the prevalence of cardiovascular disease in adulthood74.  	
    18	
    Another example in humans comes from study of the DNA methylation status of the glucocorticoid receptor in adult victims of childhood abuse that was compared to adults who did not suffer from childhood abuse in a study of postmortem hippocampi. In this study, samples were selected on the basis of psychological autopsies and patient history. Victims of childhood abuse were matched with non-victims who had similar psychiatric disorders diagnoses. For example, those who suffered from depression were matched with others who had suffered depression. Non-suicidal control subjects were selected based on age and gender for comparison. Neuronal cells from the hippocampus of suicide victims who had suffered abusive experiences in childhood were found to have a decreased level of glucocorticoid receptor expression compared to victims who had not suffered from childhood abuse7. Furthermore, they found increased levels of cytosine methylation of the promoter region of the glucocorticoid receptor gene7. A third example described global and gene-specific DNA methylation associated with prenatal smoke exposure that was observed in the epithelial buccal cells of kindergarten and first grade children75. Breton et al. found that prenatal tobacco smoke exposure was associated with small yet significant decreases in global DNA methylation in a cohort of 348 children as well as differences in the methylation status of detoxification genes (i.e. AXL and PTPRO)75. Most recently a longitudinal study considered DNA methylation in adolescents associated with parental reports of adversity during the adolescent’s infancy and childhood. Essex and colleagues found that maternal stressors in infancy and paternal stressors in early childhood were most strongly predictive of later differential methylation, and that the patterning of epigenetic methylation marks varied by the child’s  	
    19	
    gender38. Specifically, DNA methylation in the epithelial buccal cells of adolescent offspring was associated with maternal stress during infancy. DNA methylation differences were described in genes that have plausible linkages to stress (i.e. NEUROG1, a BHLH transcription factor gene that is involved in neuronal differentiation and celltype specification in the developing nervous system) and genes that are not known to be associated with stress in humans38 (i.e. PGAM2). This is one of the first long term prospective studies of the association between early childhood experiences and the later epigenetic status of children’s genomic DNA. 1.3.8  Summary and introduction to studies The convergent evidence from the reviewed literature on the association of early  experiences with later behaviour and with epigenetic regulation of gene expression in both human and nonhuman species suggests that early experience can have a lasting effect on development of offspring, and that DNA methylation may be a mechanism mediating these effects. There are a number of early experiential factors that may influence epigenetic regulation of gene expression. We propose to consider one potentially influential early experience; namely, daily physical contact time shared between an infant and its caregivers. We propose to address the following questions that have yet to be investigated. First, as an early experiential variable in human development, are differences in early caregiver-infant physical contact times associated with differences in DNA methylation in toddlers? Second, are differences in physical contact time associated with differences in the methylation status of genes associated with the HPA stress  	
    20	
    response pathway? Third, are such changes in DNA methylation detectable in peripheral epithelial buccal cells? In this study, we determine whether a specific feature of early caregiver-infant experience (namely, caregiver-infant contact time) can lead to lasting differences in epigenetic regulation of gene expression in human toddlers. We capitalize on a detailed phenotypic description of an early caregiving differences amongst caregiver-infant dyads at 5 weeks of age, and our ability to characterize large numbers of epigenetic markers of appropriate genes. We study this under the assumption that early experiences in the form of differences in contact time affect DNA methylation in multiple tissues and sites throughout the genome, and thus are likely to be observed in a variety of tissues including epithelial buccal cells regardless of the tissue type of specific gene expression. We hypothesize that differences in caregiver-infant physical contact time within the normal range at 5 weeks of age are associated with prolonged individual differences in DNA methylation patterns as evident in the epithelial buccal cells of toddlers at 3-5 years of age.  	
    21	
    2.  Study 1: Human infant early experience mediated by caregiver-  infant interactions. 	
    The reviewed literature suggests that early human experience may be associated  with epigenetic patterns in later life. Differences in caregiver-infant interactions may lead to variations in DNA methylation gene regulation and ultimately to differences in gene expression and offspring phenotype. To study early human experience and its associated epigenetic marker patterns, we first must identify and describe a clearly definable aspect of early human infant experience. We therefore propose to examine early experience in human infants to define a naturally variable caregiver-infant interaction. We propose to describe caregiver-infant physical contact time as an early experience that exhibits a large range and for which there are a collection of associated behaviours of caregivers that contribute to defining the early experience of the infant. We then propose to test the hypothesis that normal variation in caregiver-infant physical contact time within the normal range in early infancy is reflected in prolonged individual differences in DNA methylation patterns. The following studies aim to address these objectives. First, to study the epigenetic patterns associated with early experience in human offspring, we required a naturally variable, clearly defined characteristic of caregiverinfant early experience. We were able to capitalize on a previous study that supplied a detailed description of caregiver-infant behaviours and interactions. The first part of a two-part study, Study 1: Naturally occurring variation in the early experience of human infants, defines an early experience of human infants that may prove to be associated with variations in epigenetic patterns.  	
    22	
    2.1  Study 1a: Caregiver-infant physical contact time at 5 weeks of age.  2.1.1  Objective The objective of this study was to define caregiver-infant contact time in early  infancy. This study was a second look analysis of a previous data set for the purposes of defining early infant experience in a sample of young infants. 2.1.2  Background and rationale In a previous study entitled ‘Parents Helping Infants’ (PHI) by Barr et al.,  reported in: Barr RG, Barr M, Fujiwara T, et al. (2009) Do educational materials change knowledge and behaviour about crying and shaken baby syndrome? A randomized control trial, caregivers were asked to complete a diary called the Baby’s Day Diary for 4 consecutive days while their infant was 5 weeks old. Although mothers were the most likely recorders, any caregiver that was caring for the baby for a period of time was asked to record for that time period. The Baby’s Day Diary has been described previously, is used widely, and has been tested for reliability and validity76-­‐78. Caregivers recorded continuous 24-hour periods for six infant behaviours (awake and content, fussing, crying, unsoothable crying, feeding, and sleeping) and caregiver carrying/holding with body contact. We propose physical contact time as a candidate descriptor of an early experience in human infants. Physical contact time between the caregiver and infant is an early experience that crosses a number of behaviours of both the infant and caregiver, such as breastfeeding and infant crying behaviours. We used the previously collected data from this study to (1) derive measures of physical contact time in caregiver-infant dyads, and  	
    23	
    (2) determine what the ‘cluster’ of caregiving activities that were associated with differences in contact time. 2.1.3  Aim To derive measures of physical contact time in caregiver-infant dyads at 5 weeks  of age. 2.1.4  Methods The PHI study included 1,279 parents and their infants who were recruited from  maternity wards in and around the greater Vancouver Area. Participants were mothers of infants born at >34 weeks gestation. Mothers of infants with serious medical conditions or who could not speak or read English were excluded. In our second look analysis, we considered all PHI participants for which there were 3 complete days of the Baby’s Day Diary. With a 3-day diary requirement, 1,055 caregiver-infant dyads were eligible for physical contact time analysis. Calculation of physical contact time between infants and caregivers was determined based on the addition of all caregiver diary-reported time (hours/day) spent holding or carrying the infant. Individual measures of physical contact time for each caregiver-infant dyad were then averaged over the number of days for which there were complete diary entries by the caregiver. Furthermore, to define “high” and “low” contact time caregivers, we adopted the approach of Meaney and colleagues22 that observed differences in DNA methylation patterns of offspring associated with having high-LG or low-LG mothers. High-LG mother-infant dyads were defined as those mother-offspring pairs that were 1 standard  	
    24	
    deviation above the mean frequency distribution of cumulative LG, while low-LG were those mother-offspring pairs 1 standard deviation below the mean22. 2.1.5  Results The distribution of physical contact time in hours/day amongst caregiver-infant  pairs is represented in Figure 2.1.  Figure 2.1: Histogram of caregiver-infant body contact duration, grouped by 20 minute intervals.  Mean daily physical contact time shared by caregivers and infants followed a unimodal normal distribution with a slight skew. The mean hours/day (+ SD) of physical contact time for caregiver-infant dyads was 9 hours/day (+ 3hrs/day). Following Champagne et al.79 we defined high contact and low contact as greater than one standard deviation above and below the mean daily physical contact time of the 1,055 caregiver-infant dyads observed respectively. Of the 1,055 caregiver-infant dyads, 155 were defined as high contact caregiver-infant dyads and 152 were defined as low contact caregiver-infant dyads. High contact caregiver-infant dyads had a daily average 	
    25	
    of >12h 22min of contact, while low contact caregiver-infant dyads had a daily physical contact average of < 6h 14m of contact (figure 2.2). High and low contact groups differed by a separation of over 6 hours of mean daily physical contact time between caregivers and their infants.  Physical Contact Time Between Caregiver-Infant Dyads  Figure 2.2: Histogram of caregiver-infant body contact duration, grouped by 20 minute intervals with high and low contact groups defined.  The range of mean daily physical contact time between caregivers and infants for the observed population was broad, ranging from 3 hours/day to 23 hours/day. 2.1.6  Discussion As expected, the distribution of caregiver-infant physical contact time followed a  unimodal normal distribution with a slight skew. This distribution of physical contact time is analogous to the observed variation in maternal LG in rodent studies performed by Meaney and colleagues22. Figure 2.3 compares the distribution of rodent LG as 	
    26	
    published by Champagne et al.79 and the distribution of physical contact time in our sample. a  b  	
    Figure 4.3: Comparison of rodent LG and human caregiver-infant contact distributions. (a) The frequency distribution of cumulative LG during the first 6 days postpartum in rats with superimposed computer-generated normal distribution curve22. (b) The distribution of mean daily physical contact time between human caregiver-infant dyads in minutes/day.  When employing the same methods as Meaney and colleagues to define high and low caregiver-infant dyads, we observed a greater than 6 hour difference in caregiverinfant mean daily physical contact time. The large range of daily physical contact time and the significant difference between high and low contact caregiver-infant pairs in daily contact time ( > 6 hrs/day) makes it more likely that these two groups will provide useful samples to see whether differences in this form of caregiver experience are associated with later differences in methylation patterns.  	
    27	
    2.2  Study 1b: Demographic and associated caregiving behaviour  differences in high and low contact caregiving dyads. 2.2.1  Objective The objective of this analysis was to consider what other caregiving behaviours  were associated with high and low contact times among caregiver-infant dyads so as to better understand the ‘cluster’ of caregiving experiences that these infants would be receiving. 2.2.2  Background and rationale In the Baby’s Day Diary, caregivers recorded the frequency and duration of a  number of behaviours in addition to physical contact time. These behaviours included: durations and frequencies of infant fussing, crying and unsoothable crying; durations of caregivers holding their infant while feeding, while the infant was distressed and while the infant was neither feeding nor was distressed; and infant pick-ups throughout the day. We propose to define the frequencies and durations of these behaviours that are associated with high and low contact caregiving. 2.2.3  Aim Describe caregiver and infant behaviours associated with high and low contact  caregiving. 2.2.4  Methods Demographic variables of high and low contact caregiver-infant dyads were  compared using t-test or chi-squared statistics. High and low contact caregiver-infant dyads were also compared on the following infant behaviours: mean daily frequencies of infant fussing, crying and unsoothable crying, as well as mean daily durations of infant  	
    28	
    fussing, crying and unsoothable crying. In addition, the following caregiver behaviours were compared: mean infant pick-ups per day, mean daily durations of caregiver-infant contact while the infant was feeding, while the infant was distressed, and while the infant was neither feeding nor distressed. 2.2.5  Results 2.2.5.1 Demographic comparisons Mothers of the high contact caregiver-infant dyads had a higher level of education  than low contact caregiver-infant pair mothers (17yrs vs. 15yrs), had higher family incomes (30%<$60,000 vs. 49%<$60,000) and were more likely to breastfeed their infant (84% vs. 59%). Infants of high contact caregiver-infant pairs were more likely to be male (61% vs. 49%), and more likely to be an only child (68% vs. 39%). High and low contact caregivers did not differ significantly in terms of the primary caregiver’s marital status. Table 2.1: Demographic comparison of high and low contact participants from PHI study High Contact Caregivers (n=155)  Low Contact Caregivers (n=152)  17+ 3 30% Under $60,000  15 + 3 49% Under $60,000  Marital Status** • Single • Common-law • Married  7 (5%) 22 (15%) 121 (80%)  Infant Sex • Female • Male  Characteristics Maternal Education Family Income*  Significance t= -5, p<0.01 2  ✓  X =21, p<0.01  ✓  9 (6%) 17 (12%) 120 (82%)  X2 =0.8, p=0.6  X  61 (39%) 94 (61%)  77 (51%) 74 (49%)  X2 =4, p<0.05  ✓  Birth Order • Only Child • 1+ Siblings  106 (68%) 49 (32%)  59 (39%) 92 (61%)  X2 =29, p<0.01  ✓  Type of Feed*** • Breastmilk • Formula  125 (84%) 24 (40%)  87 (59%) 59 (40%)  X2 =22, p<0.01  ✓  *High n=138; Low n=135 ** High n=150; Low n=146 *** High n=149; Low n=146  	
    29	
    2.2.5.2 Infant behavioural characteristics The frequency of infant fussing bouts (7 vs. 5 times/day) and infant unsoothable crying bouts (0.5 vs. 0.3 times/day) were significantly higher in infants of high contact caregiver-infant dyads than those of low contact caregiver-infant dyads. The frequency of infant crying bouts (not defined as unsoothable) did not significantly differ between the two groups. All frequencies are represented in figure 2.4.	
   *t=-5.7, p<0.01  t=-1.6, p=0.11  *t=-2.7, p<0.01  	
   Figure 2.4: PHI high and low contact groups infant behaviours frequencies. Error bars represent standard deviations.  Infants of high contact caregiver-infant dyads also had significantly more fussing (1h 55mins vs. 1h 25mins) per day than infants of low contact caregiver-infant dyads (figure 2.5). However, the duration of unsoothable crying and non-unsoothable crying per day were not significantly different between the two contact groups.  	
    30	
    *t=-4.5, p<0.01  t=-1.4, p=0.154  t=-1.6, p=0.113  	
   Figure 2.5: PHI high and low contact groups infant behaviours durations. Error bars represent standard deviations.  2.2.5.3 Caregiver behavioural characteristics High contact caregivers picked up their infant on average more times per day than did low contact caregivers (4.3 vs. 3.2 pick-ups per day; figure 2.6). 	
    *t=-3.4, p<0.01  	
    Figure 2.6: PHI high and low contact groups caregivers mean infant pick-ups per day. Error bars represent standard deviations.  In addition, duration of daily contact holding the infant while feeding, while the infant was distressed, and while the infant was neither distressed nor feeding is examined below in figure 2.7.  	
    31	
    *t=-25, p<0.01 *t=-11.4, p<0.01  *t=-11, p<0.01  	
   Figure 2.7: PHI high and low contact groups caregiver durations of holding their infant. Error bars represent standard deviations.  On average, high contact caregivers fed longer than low contact caregivers (4h 20min vs. 3h 0min per day), they held their infants longer while the infant was distressed (2h 20min vs. 1h 3min per day), and they held the infant for longer periods of time while the infant was neither distressed nor feeding (7h 5min vs. 1h 9min per day). 2.2.6  Discussion A descriptive analysis of the high and low caregiver-infant dyads of our sample  demonstrates that there are differences in a number of infant and caregiver behaviours associated with high and low contact caregiving. Comparing demographic characteristics of these populations, high contact caregiving was associated with a higher level of maternal education and greater family income. Infants who experienced high contact caregiving were more likely to be males and only children. In addition, high contact caregivers were more likely to be breastfeeding their children at 5 weeks of age, a behaviour that elicits very close contact with the mother at times of feeding.  	
    32	
    Furthermore, the behavioural results of infants suggest that high contact caregiving was associated with increases in some infant behaviours; namely, more frequently distressed infants, more frequent bouts of unsoothable crying and longer durations of fussing when the infant is distressed. There was an associated increase in some caregiving responsive behaviours of high contact caregivers including significantly longer feeding times, longer durations of holding the infant while it is distressed, and most notably longer daily durations of holding the infant while it is neither distressed, nor feeding. These observations suggest that high contact caregiving is associated with a cluster of both infant and caregiving behavioural differences. The difference in the mean daily duration of holding while the infant was neither feeding nor distressed was approximately 6 hours between caregivers of high contact and low contact caregiver-infant dyads. This difference of 6 hours is equivalent to the difference in average total daily physical contact time between the high and low contact groups (figure 2.7). Therefore, the difference between caregivers in the average daily time spent holding the infant while the infant was neither feeding nor distressed is a prominent feature of longer average daily physical contact time between caregiver-infant dyads. In summary, studies 1a and 1b describe caregiver-infant physical contact time as an early experience that exhibits a large range for which there are a cluster of associated behaviours of caregivers with high and low contact caregiving that contribute to defining the early experience of the infant. In the following study, we assess whether differences in this early experiential caregiving feature (and/or its associated caregiving differences) are associated with different epigenetic methylation patterns later in life.  	
    33	
    3.  Study 2: DNA methylation patterns associated with high and low  caregiver-infant physical contact time. The differences in caregiver-infant contact time suggests that caregiver-infant daily contact time may be an early infant experience for which there may be associated distinguishable DNA methylation patterns. We propose to explore the potential associations between caregiver-infant contact time and DNA methylation in Study 2. 3.1  Objective The objective of this study is to see if differences in caregiver-infant physical  contact time at 5 weeks of age within the normal range of caregiving is reflected in prolonged individual differences in epigenetic methylation patterns, as evidenced in the epithelial buccal cells of toddlers at 3-5 years of age. 3.2  Background and rationale In this study, we will measure DNA methylation patterns between the toddlers of  high and low contact caregiver-infant dyads at 3-5 years of age. Through collaboration with Dr. Micheal Kobor’s lab, we were able to characterize large numbers of DNA methylation markers of genes using the Illumina Infinium Human Methylation Assay (described in Methods). For the purposes of this study, epithelial buccal cells were used for the evaluation of DNA methylation patterns in study participants. Epithelial buccal cells have been demonstrated to be useful as samples in which methylation patterns in humans can be studied38,	
  80. Talens et al. compared the methylation status of candidate loci in both mononuclear blood cells and epithelial buccal cells for stability over time and tissue correlations for the methylation status of select gene loci. They found that for 50% of the  	
    34	
    tested loci, DNA methylation measured in blood was significantly correlated with DNA methylation in the epithelial cells of the same participant80. Furthermore, human epithelial buccal cell DNA methylation patterns have been associated with social contextual-related epigenetic changes38, as described previously in the study of Essex et al in section 1.3.7. These studies suggest that there can be meaningful variation in the human epithelial cells epigenome, stability of these variations over years of time, and overlap between epithelial buccal cell and mononuclear blood cell epigenetic markers. We hypothesized that DNA methylation patterns would observable in the epithelial buccal cells of toddlers. The non-invasiveness of epithelial buccal cell collection makes them a viable method of genetic sampling for epigenetic studies in humans, especially in toddlers. In the previous literature review of epigenetics and early experience there was evidence that DNA methylation of stress related genes may be associated with early experience in both human and nonhuman studies. As a sub-inquiry in this study, we specifically consider the methylation status of the following genes for which there is evidence that their methylation status is associated with early experience in either human or nonhuman studies: NR3C1, DRD4, 5HTT, COMT, and BDNF (See section 1.3.6 for review). A review of the literature that supports the inclusion of these genes can be found in section 1.3.6. 3.3  Aim We test the hypothesis that differences in caregiver-infant physical contact time  within the normal range at 5 weeks of age is reflected in prolonged individual differences in epigenetic methylation patterns, as evidenced in the epithelial buccal cells of toddlers  	
    35	
    at 3-5 years of age. We also assess whether variation in physical contact time in early life results in differential methylation in males and females. In addition, we test a second hypothesis that mean daily caregiver-infant physical contact time will be reflected in the methylation status of genes associated with the human stress response pathways; namely NR3C1, DRD4, 5HTT, COMT, and BDNF. 3.4  Methods 3.4.1  Participants  The previously defined high (n = 155) and low (n = 152) contact caregiver-infant dyad groups were eligible for participation in this study. Eligible participants included toddlers and their parents of high and low contact caregiver-infant dyads (previously described in section 2.2). Studies 1a and 1b further describe the demographic and behavioural phenotypes of the high and low contact samples that were contacted for recruitment for this study. Ethics approval was obtained by the Behavioural Research Ethics Board (H07-01317). 3.4.2  Participant recruitment  Prospective participants were sent a letter explaining the study and inviting their participation. In recognition of their previous involvement with PHI, a copy of the research paper entitled “Do educational materials change knowledge and behaviour about crying and shaken baby syndrome? A randomized controlled trial.”81 was included in the initial contact package. The initial invitation package was followed-up with a phone call approximately two weeks later to determine whether the prospective participants were interested, to answer any questions and to review eligibility criteria. Eligibility required that the infant and mother had not suffered and were not now suffering from any major  	
    36	
    health concern. For those who had since developed minor medical conditions eligibility was determined on a case-by-case basis. An appointment was made at this time for obtaining signed consent and sample collection. In addition, all previously completed PHI demographic questionnaires and Edinburgh Postnatal Depression Scores were extracted for the eligible participants for future analysis. The PHI questionnaire results were used for the purposes of demographic analyses since they were completed at the time when the infant was subject to the difference in caregiving experience, and therefore would be a better indication of any additional differences between the experiences of the high and low contact toddlers. The Edinburgh Postnatal Depression Scale has been described previously, is used widely, and has been tested for reliability and validity82. 3.4.3  Participant sample collection  Sample collection occurred at the participant’s convenience and location of choice (either at their home or the BC Women & Children’s Health Centre). Prior to sample collection, the child’s parent(s) were asked to review and sign two copies of the written informed consent for participation in the study. Research staff then collected buccal samples and saliva samples from the child (detailed description follows). Participants were also administered a general demographic questionnaire. Questionnaires were administered to both parents of the child. In the event that one parent was missing, or time did not allow, the questionnaire was left to be completed at the parent’s convenience with a self-addressed and postage-paid envelope, or an appointment was made for their completion over the phone by a research assistant.  	
    37	
    3.4.4  Epithelial buccal swab collection and DNA extraction  Before collection, participant parents were asked to limit their child’s consumption of food and beverage to water for the hour prior to the appointment time. If the child had eaten or drank (other then water) within the hour, they were asked to rinse their mouth thoroughly with water twice. Buccal samples were collected using Isohelix Buccal Swabs (Cell Projects, Harrietsham, Kent, England). They were stabilized with stabilization reagents for storage at room temperature. Samples were then transported to the Kobor lab and stored until the time of DNA extraction. Immediately prior to DNA methylation analysis, Genomic DNA was isolated using Isohelix Buccal DNA Isolation Kits (Cell Projects, Harrietsham, Kent, England) and was purified and concentrated using DNA Clean & Concentrator (Zymo Research, Irvine, CA). Sample yield and purity were assessed spectrophotometrically using NanoDrop ND-1000 (thermo Scientific, Wilmington, DE). 3.4.5  Illumina Infinium Methylation Assay  Quantitative DNA methylation measurements of purified genomic DNA were performed with the Infinium HumanMethylatrion450 BeadChip Assay (Illumina, San Diego, CA). This microarray platform allows for quantitative measurement of the DNA methylation status of over 450,000 cytosine-guanine dinucleotide (CpG) sites within and outside of CpG islands, covering 99% of all available Reference Sequence (RefSeq) Genes. The RefSeq collection provides quite a comprehensive well-explained set of sequences, including genomic DNA, transcripts, and proteins83. A 1000ng amount of DNA was used for Bisulfite Conversion using the EZ DNA Methylation Kit (Zymo Research, Ircine, CA). Bisulfite deaminates unmethylated cytosine to uracil, while  	
    38	
    methylated cytosine is protected from deamination, thereby converting epigenetic information to sequence-based information. This can then be measured using similar methods to those used to distinguish single nucleotide polymorphisms. A sample of 160ng of bisulfite converted DNA was whole-genome amplified and fragmented by an enzymatic process and hybridized to BeadChip arrays. Following extension with DNPlabeled and biotin-labeled dNTP, each array was stained with Cy5 labeled anti-DNP antibodies and Cy3 labeled streptavidin and scanned with the Illumina HiScan on a twocolour channel. This detects the Cy5 signal on the red channel and the Cy3 signal on the green channel. Using the IlluminaGenomeStudio software package, average beta values were calculated by dividing the methylation probe signal intensity by the sum of methylated and unmethylated probe signal intensities. Average beta values therefore range from 0 (completely unmethylated) to 1 (fully methylated) and provide a quantitative readout of the relative DNA methylation for each CpG site within the cell population interrogated. In terms of reproducibility, replicates across runs achieved r>0.99, indicating a high reproducibility. 3.4.6  Data filtering, normalization and statistical analysis of DNA  methylation data Background and control normalization steps were taken using the standard settings of the GenomeStudio software package. All normalization, calculations and statistical analysis were done using R 2.11.0 (http://www.R-project.org/) and bioconductor 2.7 (http://www.bioconductor.org/). No significant outliers were identified through cluster analysis. All samples had the expected values for internal controls available on the arrays. Average background intensities, measured by negative  	
    39	
    background probes present on the array, were subtracted for raw data from both the green and red channels separately to adjust for varying background signals across the samples, and for Cy3 and Cy5 differences. Confirmation of sample gender was achieved using the chromosome X data. Females would exhibit close to 50% methylation in mode chromosome X sites due to the inactive X, whereas males would be unmethylated. The sex of all samples was confirmed. Individual data points were removed following failure of the manufacturer’s standard quality control (> 3 beads represented and detection o values < 0.01). CpG sites missing > 25% of it’s data points were determined to be bad, and were removed (n = 95) to ensure that only high-confidence probes were included in the subsequent analyses. All CpG probes with a detection p-value > 0.05 were removed to ensure that only high-confidence probes were included in the subsequent analysis. The detection p-value is an assessment of whether the signal intensity for the average beta value is significantly different from the negative probes, as well as consistent across all oligonucleotides interrogating a given CpG site on the array. To increase statistical power, CpG sites that were completely unmethylated (all data points were < 0.05 average beta; n = 34354) or completely methylated (all data points were > 0.95 average beta; n = 4834) were removed, so that all sites would have between-sample differences in methylation > 5% of the average beta value. In addition, all genotyping probes (n = 65) were removed from the dataset. Finally, all probes known to be associated with the X-chromosome were removed prior to analysis. These probes were removed due to the ~50% methylation of these CpG sites associated with silencing of the second X-chromosome in females. Failure to remove these probes would result in an inaccurate representation of the  	
    40	
    differences in methylation among toddlers. This resulted in a final data set of 434580 CpG sites that were used in all subsequent analyses. Technical replicates were collapsed by averaging the available data points. Any missing data points were imputed using a knearest mean imputation. For gene-specific analyses, CpG sites in promoter regions of the glucocorticoid receptor (NR3C1), dopamine receptor (DRD4), serotonin transporter (5 HTT), catecholO-methyltransferase (COMT) and brain-derived neurotrophic factor (BDNF) were used for DNA methylation analyses. 3.5  Results: Demographic and associated caregiving behaviour differences in  high and low contact caregiver-infant dyads 3.5.1 	
    Demographic comparison  Of those eligible for participation (n = 307), 98 participants successfully  completed Study 2:DNA methylation patterns associated with high and low caregiverinfant physical contact time. A total of 155 high contact caregiver-infant dyads were contacted, 59 participants completed the DNA sample collection and were included in our analysis. Of the 152 low contact caregiver-infant dyads eligible, 39 completed the DNA sample collection and were included in our analysis. Figure 3.1 depicts the breakdown of participant eligibility and recruitment.  	
    41	
    Figure 3.1: Selection of participants for inclusion in the Study 2.  	
    42	
    Table 3.1 compares the 98 families that were successfully completed to those that were either ineligible or declined participation (n = 209). Table 3.1: Comparison of consented participants and the non-consented participants from the eligible sample.  16+ 3 25% Under $60,000  NonConsented Participants (n=209) 15 + 3 55% Under $60,000  Marital Status** • Single • Common-law • Married  3 (5%) 11 (10%) 83 (85%)  Infant Sex • Female • Male  Characteristics Education Family Income*  Consented Participants (n=98)  Significance t= -3, p<0.01 2  ✓  X =27, p<0.01  ✓  13 (6%) 28 (14%) 160 (80%)  X2 =2, p=0.4  X  40 (41%) 58 (59%)  99 (47%) 111 (53%)  X2 =1, p=0.3  X  Birth Order • Only Child • 1+ Siblings  56 (57%) 42 (43%)  109 (59%) 101 (48%)  X2 =2.4, p=0.7  X  Type of Feed*** • Breast milk • Formula  72 (74%) 25 (26%)  142 (71%) 58 (29%)  X2 =0.3, p=0.6  X  *Consented n=92; non-consented n=183 ** Consented n=97; non-consented n=201 *** Consented n=97; non-consented n=200  Compared to the PHI participants that did not consent to participate in Study 2, consented participants were only significantly different in the level of maternal education (16yrs vs. 15yrs) and gross family income (25% vs. 55% <$60,000). All other demographic characteristics recorded, namely marital status, infant sex, birth order and type of infant feed, did not significantly differ between the consented participants and those that declined or were ineligible at the time of Study 2 completion. The final study sample was comprised of 59 toddlers from high contact caregiverinfant dyads and 39 from low contact caregiver-infant dyads. Their demographic characteristics are summarized in table 3.2. A comparison of the consented high contact and low contact caregiver-infant dyads shows there were a significantly larger percentage 	
    43	
    of male infants in the high contact group compared to the low contact group (68% vs. 46%). Table 3.2: Comparison of high contact participants and the low contact participants. “High” Contact CaregiverInfant Dyads (n=59)  “Low” Contact CaregiverInfant Dyads (n=39)  17+ 2 16% Under $60,000  16 + 3 39% Under $60,000  Marital Status** • Single • Common-law • Married  3 (5%) 7 (12%) 48 (83%)  Infant Sex • Female • Male  Characteristics  Education Family Income*  Significance  t= -2, p=0.3  X  X2 =16, p=0.01  X  0 (0%) 4 (10%) 35 (90%)  X2 =2.2, p=0.3  X  19 (32%) 40 (68%)  21 (54%) 18 (46%)  X2 =4.5, p<0.05  ✓  Birth Order • Only Child • 1+ Siblings  39 (66%) 20 (34%)  17 (44%) 22 (56%)  X2 =6.0, p=0.1  X  Type of Feed*** • Breast milk • Formula  46 (79%) 12 (21%)  26 (67%) 13 (33%)  X2 =1.9, p=0.2  X  *High n=56; Low n=36 ** High n=58; Low n=39 *** High n=58; Low n=39  3.5.2  Infant behavioural characteristics  The frequencies and durations of infant behaviours that may be associated with high and low contact caregiving are compared in the following results. First, the frequency of infant fussing bouts (8 vs. 5 times/day) was significantly higher in high contact infants than those of the low contact group. However, infant crying (not defined as unsoothable) and infant unsoothable crying did not significantly differ between the two groups. All frequencies are represented in figure 3.2. 	
    	
    44	
    *t=-3.2, p<0.01  t=-1.6, p=0.1  t=-2, p=0.04  	
   Figure 3.2: High and low contact groups infant behaviours frequencies. Error bars represent standard deviations.  	
    	
    High contact infants also had significantly longer periods of fussing (2h 7mins vs. 1h 38min) per day than low contact infants, as evident in figure 3.3. However, the duration of unsoothable crying and non-unsoothable crying were not significantly different between the two contact groups.  	
    45	
    *t=-2.6, p<0.01  t=-2.1, p=0.04  t=-1.6, p=0.1  	
   Figure 3.3: High and low contact groups infant behaviours durations. Error bars represent standard deviations.  3.5.3  Caregiver behavioural characteristics  The following is a descriptive analysis of the associated behaviours of caregivers with high and low contact caregivers for the completed study 2 participants (n = 98). High contact caregivers did not pick-up their infants significantly more times per day than low contact caregivers (figure 3.4). 	
    t=-0.8 p=0.4  	
   	
   Figure 3.4: High and low contact groups caregivers average infant pick-ups per day. Error bars represent standard deviations.  	
    46	
    In addition, duration of daily contact from holding the infant while feeding, while the infant was distressed, and while the infant was neither distressed nor feeding is examined below in figure 3.5.  *t=-14, p<0.01 *t=-6.4, p<0.01  *t=-8, p<0.01  	
   Figure 3.5: High and low contact groups caregiver durations of holding their infant. Error bars represent standard deviations.  High contact mothers fed on average longer than low contact mothers (4h 35min vs. 2h 55m), they held their infants longer while it was distressed (2h 33min vs. 1h 5min), and they held the infant for longer periods of time while the infant was neither distressed nor feeding (7h 4min vs. 1h 15min). All of these were significantly different between high and low contact caregiver-infant dyads.  	
    47	
    3.5.4  Discussion of demographic and associated caregiving behaviour differences in high and low contact caregiver-infant dyads results  The clustering of behavioural characteristics of high and low contact caregivers and infants of the consented Study 2 participants were similar to those of the larger PHI population (n = 307; see table 3.3). Table 3.3: Comparison of infant and caregiver behaviours found to be significantly higher in high caregiver-infant dyads  Contact associated infant and caregiver behaviours Infant Behaviours Frequency of infant fussing Frequency of infant unsoothable crying Frequency of infant crying Duration of infant fussing Duration of infant unsoothable crying Duration of infant crying Caregiver Behaviours Average pick-ups Duration of feeding Duration of holding while infant is distressed Duration of holding while infant is not distressed nor feeding  Significant findings for infant and caregiver behaviours PHI High vs. Study 2 Low High vs. differences Low differences √ √  √  √  √  √ √ √  √ √  √  √  Frequency of unsoothable crying and the number of times the caregiver picked up the infant were no longer significantly greater in the high contact caregiver-infant group of study 2. However, all other behaviours found to be significantly different between high and low contact groups in the PHI population were similarly significant in the study 2 subsample. The greatest behavioural difference between the high and low contact caregivers of study 2 is likewise the significantly longer periods of time caregivers held their infants while the infant was neither crying nor distressed. The differences in these 	
    48	
    caregiving behaviours show similar differences between high and low contact caregiverinfant dyads as was previously seen in the larger PHI high and low contact caregiverinfant pairs. 3.6  Results: DNA methylation pattern differences in high and low contact  caregiver-infant dyads 3.6.1  Interpretations of Illumina Infinium Methylation Assay results  Analysis of methylation patterns associated with differences in caregiver-infant physical contact time was assessed using a variety of statistical interrogations. Here I describe the steps in the analytic strategy followed in the evaluation of the methylation assay results. The described terms will be used throughout the methylation assay results section. 3.6.1.1 Methylation difference and average Beta values Methylation of each CpG site was evaluated for each toddler’s DNA sample. Infinium HumanMethylation450 BeadChips provide a quantitative measure of DNA methylation denoted as Average Beta (β) calculated as the ratio of methylated to total DNA. Average beta value ranges from 0-1, with 0 indicating that the given CpG site is completely un-methylated. Methylation difference compares the mean average beta of high and low contact caregiver-infant dyads at that CpG site. Methylation difference can be likened to a percentage (i.e. 0.0158 would suggest that at this particular CpG site in high contact toddlers is 1.58% more likely to be methylated than in low contact toddlers). For the purposes of our analyses, low contact is used as the reference point; therefore a positive methylation difference indicates a greater average beta in the high contact than low contact.  	
    49	
    3.6.1.2 Benjamini-Hochberg correction A Benjamini-Hocherg (B-H) correction method was used to test for false positive difference due to multiple testing. B-H is a method of controlling the expected false discovery rate for our results. We controlled for a 5% false discovery rate. Although B-H is arguably a less conservative approach for correcting for false discovery than other statistical methods, it provides a sufficient restriction on our results for a first look at the DNA methylation associations with caregiver-infant physical contact time. 3.6.1.3 T-test and T-test adjusted p-values The T-test p-values are automatically computed using the ‘R’ software program. The T-test p-values are corrected for multiple testing issues by the B-H correction resulting in a list of adjusted p-values for each CpG site. This calculates which of the CpG sites are significantly different for high contact and low contact caregiver-infant dyads allowing for a 5% false discovery rate, and is calculated in consideration of all the CpG sites represented on the Illumina array (n = 434,580). 3.6.1.4 Wilcoxon test and Wilcoxon test adjusted p-values Considering that a t-test is a parametric test that applies only when the test statistic follows a normal distribution, using only a t-test analysis may be of concern if CpG site methylation does not follow a normal distribution. Previously, when the 27K array Illumina analyses were used in the Kobor Lab, they found that half of the CpG sites were not normally distributed (evaluated previously by the Kobor lab but unpublished to date). Therefore, in addition to a t-test, a Wilcoxon test was applied to compare methylation difference between high and low contact toddlers for each CpG site. This is a non-parametric test that does not assume a normal distribution of values. For the majority  	
    50	
    of our results, we used the t-test p-values to rank the CpG sites in order of potential significance. However, as a precaution, we also performed Wilcoxon tests for each analysis and have included these results throughout. 3.6.1.5 Linear regression analysis From the PHI study, we were able to calculate mean daily physical contact time between caregivers and infants for each of the study participants. In addition to grouping high and low contact caregivers together as was described previously in section 2.1, we did a linear regression against methylation status where contact time was not used to define groups but rather as a continuous variable to see if there was a correlation between physical contact time and methylation status. The range within high and low caregiver contact groups was fairly substantial (10h 50m and 3h 6m respectively) which lead us to believe that considering contact time as a continuous variable (rather than a grouping variable) might provide a more sensitive way to assess possible correlations between methylation and contact time. As a result, the specific physical contact time for each infant was determined and used as a predictor variable in a regression analysis. We compared the distribution of average beta values for all caregiver-infant dyads to see if there were any apparent trends in average beta values between high and low contact caregiver toddlers. A linear regression analysis may make apparent any differences that are specific to the tail ends of the contact groups that are hidden by categorizing pairs as high or low caregiver contact toddlers. Furthermore, a linear regression analysis may make any associations with sex of the participant or sex-and-contact interaction associations with differential methylation of a CpG site more apparent. In summary, considering the variability in physical contact time amongst human caregiver-infant  	
    51	
    dyads at 5 weeks of age, we performed a linear regression against methylation status where contact time was considered as a continuous variable to see if there were any correlations between physical contact time and methylation status in our sample of toddlers as this may be a more sensitive approach with which to explore whether DNA methylation differences are associated with contact time. 3.6.1.6 Gender-specific analysis To determine whether any associations were due to gender, methylation differences associated with high and low caregiver-infant physical contact time were assessed while controlling for gender. These results are of contact group comparisons within females only or males only. T-statistics, t-test adjusted p-values and methylation differences for each CpG site were determined for males and females only. The mean average beta for high contact males and low contact males for each CpG site were calculated and compared. The same was done for female toddlers. The methylation differences for the female and male only populations were the mean of the high contact toddlers minus the mean of the low contact toddler group within their gender groups, therefore indicating the magnitude of the difference for that gender at that site. 3.6.1.7 Candidate gene analysis As described in our literature review, there are a few candidate genes that are of particular interest as candidates for DNA methylation regulation associated with variations in early human experience. The genes of interest are: the glucocorticoid receptor (NR3C1), dopamine receptor (DRD4), serotonin transporter (5 HTT), catecholO-methyltransferase (COMT), and brain-derived neurotrophic factor (BDNF). We compared the average beta values of all CpG sites represented on the Infinium Illumina  	
    52	
    MethylationBeadChip450 array within regions of our candidate genes for high and low contact toddlers. We considered all of the 239 CpG sites located in the regions of one of the candidate genes that are represented on the array. High and low contact caregiver toddlers were compared for differences in methylation at these sites. 3.6.2  Differential methylation associated with differences in physical  contact time between caregiver-infant dyads The methylation status of each of the 434,580 CpG sites (as previously defined in: Data filtering, normalization and statistical analysis of DNA methylation data) on the Illumina array was evaluated and compared between high contact and low contact toddlers. Using a simple T-test, we compared the p-values across all CpG sites and ranked in ascending order. 14,827 CpG sites had a p-value of < 0.05, 196 CpG sites had a p-value of < 0.001, and only 8 CpG sites were found to have a p-value of < 0.0001. Following B-H correction for multiple testing, none of the 434,580 CpG sites were found to have a significant methylation difference between high and low contact groups. Therefore, high and low physical contact toddlers did not show significant differences in DNA methylation patterns of their epithelial buccal cells. The results of the Wilcoxon test were similar to the T-test p-values. They were consistent with the previous parametric analysis in showing that there were no CpG sites that were significantly different between high and low contact toddlers following B-H correction. Although none of the methylation differences of the considered CpG sites (n = 434,580) were found to be significant between high and low contact toddlers following B-H correction, we took a secondary look at the CpG sites for which the methylation  	
    53	
    difference between high and low contact toddlers showed the greatest statistical by t-test and Wilcoxon test prior to B-H correction. We considered the top 8 CpG sites by T-test p-value of significance and the top 10 CpG sites by Wilcoxon test p-value of significance (Appendix A; Common sites for which there was a low p-value from both the Wilcoxon and T-test analyses are highlighted in grey). When comparing the results of these two tests, six CpG sites overlapped. Nonetheless, there were no significant differences in methylation (see Methylation Difference, table 3.4) between high and low contact toddlers at any of these sites. That was true when adjusted T-test was used, and when adjusted Wilcoxon test was used.	
  The six CpG sites with the smallest p-values as determined by both T-test and Wilcoxon test were evaluated. The results are summarized in Table 3.4 Table 3.4: High contact vs. low contact 6 CpG sites with the lowest p-value by both T-test and Wilcoxon test. Gene  CpG site  Tstatistic  TCEB3 ESPL1 HINT1 BRF1 ZBTB20 (unknown)  cg21773314 cg25976786 cg20640983 cg22496859 cg17460368 cg01710670  -4.62 4.41 -4.41 -4.29 4.68 4.10  Tstatistics p-value 1.2E-05 2.7E-05 2.9E-05 5.6E-05 1.1E-05 9.6E-05  Methylation difference -0.013 0.018 -0.012 -0.010 0.016 0.246  T-test adjusted p-value 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999  Wilcoxon statistics  Wilcoxon p-values  560 1730.5 592 596 1724 1705  1.8E-05 2.6E-05 5.1E-05 5.8E-05 3.2E-05 5.8E-05  Wilcoxon adjusted p-value 1 1 1 1 1 1  In addition, we plotted the distribution of average beta values for the six CpG sites with the smallest p-values as determined by both T-test and Wilcoxon. These are shown in figure 3.6.  	
    54	
    	
   Figure 3.6: Comparison of the top 6 cytosine-guanine dinucleotide (CpG) sites associated with low and high caregiverinfant physical contact time as determined by T-test (p<0.0001) and Wilcoxon analyses (p<0.0001). Blue circles indicate female infants, green stars indicate male infants, and with best fit lines in the corresponding colours.  Figure 3.6 represents the distributions of average beta values for the entire study sample (n = 98). Blue circles indicate female subjects, green stars indicate male subjects, and their corresponding best-fit lines are blue and green accordingly. For these sites, the male and female subjects are so tightly clustered that it is difficult to identify males from females due to their extensive overlap. Looking at the distribution for example of TCEB3-cg21773314, we see first that values the female and male subjects are tightly clustered together with no significant differences between males and females in average beta values at any of these sites. This suggests little variation in the methylation status of  	
    55	
    these sites associated with gender. Second, comparing high and low contact toddlers, the values for all subjects have clustered on a very horizontal plane for each CpG site except cg-01710670. For the other 5 sites, there were no outstanding differences between the distribution of average beta values amongst high contact toddlers and low contact toddlers. In summary, there appears to be very little difference in the average beta values for these five sites associated with either gender or physical contact time at 5 weeks. However, as suggested, there is an interestingly different distribution of the cg01710670 CpG site. Of the sites with the greatest statistical significance as defined using the T-test and Wilcoxon test, the CpG site cg01710670 appears to have a larger methylation difference between high and low contact toddlers as well as a more distinct difference between males and females. From the distribution of beta averages seen in figure 3.6, there seems to be two distinct groupings of the methylation status of this CpG site. One possibility is that this distribution of average beta values is a characteristic of the cg01710670 site (perhaps functioning bimodally); another possibility is that an error has occurred during analysis. In interpreting this result, a number of considerations are relevant. The cg01710670 site is not found to be within any known genes as determined by GenomeStudio Software from the Illumina Infinium Methylation Assay. It is possible that the apparent grouping of average beta values for cg 01710670 is due to confounding factors that we are unaware of at this time. This could be related to the structural nature of the gene itself: cg01710670 could be a single nucleotide polymorphism (SNP) or be in close range of one that would result in different alleles, and therefore result in a bimodal distribution. We were unable to conclude that this site was near a known SNP; however,  	
    56	
    we are still unclear as to the precise location of this CpG site in terms of a possible associated gene. In addition, while all procedures have been verified for the possibility of technical error including batch effects, chip damage, collection and initial DNA levels without resulting in any evidence of technical error, it remains possible that the irregular distribution of cg01710670 is a product of unknown variable analyses technique or error. Thus the significance of the apparent correlation observed here between cg01710670 and the contact group and gender may or may not be spurious but cannot be considered conclusive. As this study is an exploratory study of the potential epigenetic patterns associated with caregiver-infant contact time, we also considered the results using less restrictive analyses. With a less stringent p-value cut-off, the possibility of reporting false positives increases, therefore we focused our analysis on CpG sites with a minimum 5% difference (methylation difference of either > 0.05 or < -0.05) between high and low contact caregiver toddlers. Using a less restrictive t-test p-value cut-off of < 0.001 (instead of the p < 0.0001 used previously) and an absolute methylation difference of > 0.05, we nonetheless observed there were no significant differences in methylation (see Methylation Differences, table 3.5) between high and low contact toddlers at any of these sites. That was true when adjusted t-test was used, and when adjusted Wilcoxon test was used. The top 5 sites using this less restrictive t-test p-value of < 0.001 analyses are listed in Table 3.5. Interestingly, there was no overlap with our previous analysis except for CpG site cg01710670 (highlighted in Table 3.5 in grey) that is thought to be inconclusive in regard to the objectives of our study as previously discussed.  	
    57	
    Table 3.5: High contact vs. low contact 5 CpG sites with the lowest T-test p-value (p < 0.001), and an average beta value > 5%. Gene  CpG site  Tstatistic  (unknown) FAM132A SMYD3 RNF215 LRRC15  cg01710670 cg24345856 cg08887025 cg03868770 cg14180696  4.10 -3.64 -3.56 3.55 -3.49  Tstatistics p-value 9.6E-05 4.9E-04 6.3E-04 7.0E-04 7.3E-04  Methylation difference 0.246 -0.099 -0.119 0.059 -0.055  T-test adjusted p-value 0.9999 0.9999 0.9999 0.9999 0.9999  Wilcoxon statistics  Wilcoxon p-values  1705 937 926 1606 745  5.8E-05 0.1222 0.1040 9.6E-04 0.0033  Wilcoxon adjusted p-value 1 1 1 1 1  Figure 3.7: Comparison of the top 5 cytosine-guanine dinucleotide (CpG) sites with a DNA methylation difference of >5% (>0.05 average beta) between low and high Contact groups, and the lowest p-value (p<0.001) as determined by a T-test. Females are denoted as blue circles, while males are green stars. The best-fit line for each sex is in the corresponding colour.  The distributions of average beta values for the top 5 CpG sites using our less restrictive analyses (p < 0.001, 5% methylation difference) between high and low contact toddlers are plotted in Figure 3.7. Females are denoted as blue circles, while males are  	
    58	
    green stars. The best-fit line for each sex is in the corresponding colour. These 5 sites have been selected for a 5% methylation difference between high and low contact toddlers which explains the less horizontal clustering, or the somewhat greater differences in the distribution of average beta values of high compared to low toddler groups. However, none of these sites were found to be statistically significantly different following B-H correction. What stands out in these distributions are sites cg14180696 (LRRC15) and cg03868770 (RNF215) for the apparent opposite trends of the differences for males and females. Sites cg01710670, cg24345856, and cg08887025 share an unusual clustering of average beta values. Cg01710670 has been previously discussed as one of the six CpG sites with the smallest p-values as determined by both T-test and Wilcoxon test in the non-gender specific analysis. It is likely that the apparent bimodal distribution of average beta values for both cg24345856 and cg08887025 are bimodal due to a single nucleotide polymorphism (SNP) and with 2 alleles. Cg24345856 has a SNP at the CpG site with a minor allele frequency of 17.8%, so this likely accounts for the observed distribution. Cg08887025 has a SNP with a minor allele frequency of ~30% within 5 nucleotide base pairs of the CpG site. Depending on the probe orientation, this likely also accounts for the bimodal distribution observed. In summary, there are no statistically significant differences in the methylation status of these CpG sites.  	
    59	
    3.6.3  Differential methylation associated with physical contact time: linear  regression analysis Using a linear regression of methylation vs. contact time with contact time as the predictor, only seven CpG sites demonstrated a p-value significance of < 0.0001. Following a further restriction to a 5% methylation change, the number of significant CpG sites was reduced to 1. Following Benjamini-Hochberg correction of p-values no CpG sites were found to be significant. Table 3.6 summarizes the results of statistical analyses for the top 7 cytosine-guanine sites with the smallest p-values as determined by a linear regression. Methylation difference indicates the mean methylation level of those of the high contact group compared to the mean methylation level of those of the low contact group. Adjusted p-values were those following Benjamini-Hochberg correction for false discovery. No sites retained significance following this correction. Additionally, two-way ANOVA with interaction assessed the correlation of these sites to gender and a contact time X gender interaction. No sites were significantly correlated to gender, nor did controlling for gender affect the initial findings of no statistical significance.  	
    60	
    Table 3.6: Summary of linear regression and Two-Way ANOVA results for the 7 CpG sites with the lowest regression p-value for caregiver-infant contact time as determined by a regression analyses. Gene  CpG site  Regression p-value  Methylation difference  T-test adjusted p-value  Gender pvalues  Contact time pvalues  SH3BP5 NA TSKS PRNP NLRP6 PSG9 DENND4B  cg17252884 cg27560687 cg22525294 cg03704756 cg16359142 cg11320553 cg19292953  3.0E-05 3.2E-05 5.8E-05 7.0E-05 9.4E-05 9.4E-05 9.6E-05  -0.016 0.031 0.107 -0.022 -0.037 -0.037 -0.030  0.999969 0.999969 0.999969 0.999969 0.999969 0.999969 0.999969  0.5629 0.1156 0.1896 0.1896 0.9994 0.8624 0.2484  0.0000 1E-04 2E-04 2E-04 1E-04 1E-04 2E-04  Gender and contact time interaction p-values 0.769 0.053 0.846 0.846 0.930 0.251 0.906  The distribution of the average beta values for the 7 CpG sites with the lowest regression p-value are found in figure 3.8. Females are denoted as blue circles and males green stars. The lines of best fit are in their associated colour for sex. Contact time shared between caregiver and infant is along the x-axis in minutes/day. The average beta values for these CpG sites have quite a horizontal distribution, indicative of no apparent association between the methylation status of the CpG site and physical contact time. Looking at the average beta values of males and females, there is no apparent difference in their distribution at these sites.  	
    61	
    Figure 3.8: Comparison of the 7 cytosine-guanine dinucleotide (CpG) sites with the lowest p-value for regression analyses by contact time. Note that contact time is the average daily minutes/day of physical contact between caregiver and infant for each dyad of the current study. Females are denoted as blue circles, while males are green stars. The bestfit line for each sex is in the corresponding colour.  In summary, there are no apparent associations between DNA methylation status and contact time as a result of the linear regression analyses. Furthermore, there appear to be no gender-contact time interaction associated with differences in methylation at these sites.  	
    62	
    3.6.4  Differential methylation associated with physical contact time: gender  specific analysis Specific analyses for male and female toddlers produced patterns of methylation that were distinctive of those observed with the full sample. A two-way ANOVA comparing high and low caregiver contact groups in all 434,580 CpG sites with gender control resulted in only 9 sites surviving a p-value cut-off of <0.0001, only one of which met the 5% methylation change criteria, and none survived following BenjaminiHochberg correction. The 12 CpG sites with the lowest p-value (p < 0.0001) are described in table 3.7. Eleven of the twelve CpG sites were novel sites, while only one was previously investigated without gender control. Table 3.7 summarizes the results of statistical analyses for the 12 cytosineguanine dinucleotide (CpG) sites with a p-value of < 0.0001 in gender and contact group interaction. The methylation differences listed for the female and male only samples are the mean of the high contact toddlers minus the mean of the low contact toddler group, indicating the magnitude of the difference for that gender at that site. The T-statistic values for males and females are opposite (i.e. negative methylation difference for one, while positive for the other) indicating that the methylation difference in males and females due to contact is unique and opposite in nature at these sites. As expected of the interaction term, these are sites where the methylation difference in one sex between contact groups is most different from the difference in methylation between contact groups for that of the other sex.  	
    63	
    Table 3.7: Summary of Two-Way ANOVA results for the 12 CpG sites with the greatest gender-contact interaction.  Gene  CpG site  GAK  cg04220779  FAM20C  cg09630706  (unknown)  cg03553758  FAM174A  cg21759811  (unknown)  cg16164062  IGFBP7  cg00884221  RNF215  cg17583390  TMEM30C  cg06073355  OR12D2  cg08930944  (unknown)  cg27442616  TTC23L  cg07895169  (unknown)  cg09030829  Gender & contact group interaction PAdjusted values p-value 2.3E06 4.6E06 2.3E05 5.5E05 6.0E05 6.0E05 6.0E05 6.2E05 6.7E05 7.4E05 8.1E05 9.9E05  Female only  T-test statistic  Male only  0.9905  -3.84  T-test adjusted p-values 5E-04  Methylation difference  T-test statistic  Methylation difference  3.37  T-test adjusted p-values 1.4E-03  -0.027  0.9905  4.04  2E-04  0.019  -2.7  9E-03  -0.011  0.9905  -3.80  5E-04  -0.012  2.27  0.027  0.006  0.9905  -3.24  2.5E-03  -0.012  2.54  0.014  0.007  0.9905  4.32  1E-04  0.016  -2.20  0.031  -0.010  0.9905  -2.90  6.1E-03  -0.009  2.97  4.4E-03  0.007  0.9905  -3.82  5E-04  -0.025  2.55  0.013  0.021  0.9905  2.76  8.8E-03  0.017  -3.13  0.004  -0.145  0.9905  -2.84  7.E-033  -0.023  3.03  0.004  0.020  0.9905  -4.17  2E-04  -0.014  1.79  0.079  0.006  0.9905  3.57  1E-03  0.008  -2.11  0.040  -0.004  0.9905  3.79  5E-03  0.024  -1.66  0.103  -0.008  0.023  No particular site was found to have a large methylation difference between high and low contact groups, as is evident in the average beta distribution graphs plotted in figure 3.9.  	
    64	
    Figure 3.9: Comparison of the top 12 cytosine-guanine dinucleotide (CpG) sites with significant contact group and gender interaction as determined by two-way ANOVA (p<0.0001). Blue circles indicate female infants and green stars indicate male infants. Best-fit lines are drawn for the females and males in their respective colours.  Each of the 12 CpG sites where the methylation difference in one sex between contact groups is most different from the difference in methylation between contact groups for that of the other sex have little difference in the distribution of average beta values. This confirms our non-significant results for differences between male and female toddlers for high and low contact associations with methylation status of CpG sites. However, such sites as RNF215-cg17583390 have discernible distinct opposing trends observed in the gender-contact interaction analyses suggesting that it is possible that males and females have different DNA methylation patterns associated with physical contact time in infancy relative to this gene. With a larger sample, the difference in 	
    65	
    caregiver-infant physical contact time might be more likely to be statistically significant. If replicated with a different sample, it is possible that we would see a larger difference between the males and females (a stronger opposing methylation change in response to variations in contact time) that would make this interaction more statistically significant, helping it to pass multiple test correction. 3.6.5  Gene specific analysis  For our proposed candidate genes, namely the glucocorticoid receptor (NR3C1), dopamine receptor (DRD4), serotonin transporter (5 HTT), catechol-O-methyltransferase (COMT), and brain-derived neurotrophic factor (BDNF) a limited number of CpG sites located in the regulatory regions of these genes are represented on the Illumina Infinium HumanMethylatrion450 BeadChip Assay. For these genes, there are a total of 239 CpG sites on the micro-array. The t-test results for all 239 CpG sites associated with targeted candidate genes can be found attached as supplemental results in “Appendix D: Supplemental Results all CpG sites associated with targeted candidate genes”. A summary of the statistical analyses for the cytosine-guanine dinucleotide (CpG) site with the lowest t-test p-value for each of the targeted candidate genes is found below in table 3.8. As in previous analyses, there were no significant differences in methylation (see Methylation Differences, Table 3.8) between high and low contact toddlers at any of these sites. That was true when t-tests and adjusted t-tests were used, and when Wilcoxon and adjusted Wilcoxon tests were used.  	
    66	
    Table 3.8: Summary of T-test results for the CpG site with the smallest p-value by T-test for each of the seven targeted candidate genes. Gene  CpG site  Tstatistic  BDNF COMT DRD4 HTT NR3C1 SLC6A3 SLC6A4  cg05818894 cg17810098 cg20411756 cg12573747 cg17860381 cg12516187 cg06841846  -1.93 3.28 -1.27 2.54 -2.25 -2.85 -1.54  Tstatistics p-value 5.9E-02 1.4E-03 0.2086 1.3E-02 2.8E-02 5.9E-03 0.1275  Methylation difference 0.006 -0.005 0.022 -0.004 0.004 0.034 0.004  T-test adjusted p-value 0.993 0.427 0.993 0.993 0.993 0.870 0.993  Wilcoxon statistics  Wilcoxon p-values  962 1501 947 1468 864 786 963  0.1956 0.0086 0.1605 0.0169 0.4456 0.0099 0.1981  Wilcoxon adjusted p-value 0.989 0.827 0.989 0.844 0.989 0.827 0.989  All of the CpG sites associated with our candidate genes failed to be significant following multiple testing correction of B-H. The distributions of the average beta values for the single top CpG site for each targeted gene are plotted in figure 3.10. Blue circles indicate female infants and green stars indicate male toddlers. As expected, the distributions of these CpG sites do not show a significant difference between high and low contact toddlers.  	
    67	
    Figure 3.10: Comparison of the top cytosine-guanine dinucleotide (CpG) site for each targeted gene of interest. Females are denoted as blue circles, while males are green stars. The best-fit line for each sex is in the corresponding colour.  The only CpG site that might be of interest is DRD4-cg20411756 that appears to have different trends for males and females of high and low contact caregivers. Although there is little difference in the male toddlers between high and low contact (if anything a slightly positive trend), there is a much more negative trend amongst high and low contact female toddlers at this site. However, in the previously presented gender specific analysis, no sites were found to be significantly associated with sex and contact group, and this site was not among the sites nearing significance (section 3.6.4). Therefore, we  	
    68	
    conclude that there are no significant methylation differences between the high and low contact toddlers at these sites. 3.7  Summary of Study 2: DNA methylation patterns associated with high and  low caregiver-infant physical contact time. In summary, the presented results are an exploratory analysis of the methylation patterns of epithelial buccal cells from toddlers who experienced differences in the amount of physical contact time shared between them and their caregiver as infants. We sought to fully explore the potential differences within our study population. To do so, we used more liberal statistical cut-offs to ensure any potential positive associations were not missed, as well as tested for gender differences and interactions, and specifically looked at the methylation patterns of regulatory regions of candidate genes. Regardless of the strategy used, the overall finding was that there were no cytosine-guanine dinucleotide sites of the 434,580 sites interrogated for which an average daily variation in caregiverinfant physical contact time of 6 hours/day at 5 weeks of age were associated with lasting differences in the methylation patterns of the epithelial buccal cells of these infants at the age of 3-5 years.  	
    69	
    4  Concluding chapter  4.1  Discussion Interrogating nearly 450,000 CpG sites in regulatory regions representing all 23  chromosomes of human DNA84, we found no CpG site for which toddlers who experienced either higher or lower levels of caregiver-infant physical contact time showed significantly different levels of DNA methylation. Specifically, there were no significant methylation differences in the epithelial buccal cells of toddlers associated with higher or lower caregiver-infant physical contact time defined at 5 weeks of age as a daily average difference of + 6hrs. To our knowledge, this study is the first report of the methylation patterns associated with caregiver-infant physical contact time in humans. 4.1.1  Convergence with past findings  The literature supporting the epidemiologic associations between early life experiences as a child and long-term consequences for health and development is ever expanding. The findings in our study are not convergent with past findings of DNA methylation patterns associated with early experiences in humans. This study is an exploratory study of a specific early experience, namely caregiver-infant physical contact time at 5 weeks of age and associated prolonged DNA methylation patterns in the epithelial buccal cells of toddlers. Our study differs from previous studies in terms of study population (3-5 year old toddlers), tissue type used in methylation analysis (epithelial buccal cells) and our early experience of interest (caregiver-infant physical contact time). The following section will briefly review the relevant human studies that have shown epigenetic associations with early experience in humans, and address a few key concepts that need consideration in the interpretation of this particular study.  	
    70	
    To date, there are only a handful of studies that have considered the possible associations between early experience and DNA methylation patterns in humans. First, in a study of monozygotic twins, Fraga et al. established that epigenetic markers of peripheral lymphocytes were more distinct in monozygotic twins who were older, had different lifestyles, and had spent less of their lives together85. Furthermore, Fraga et al. observed that the older monozygotic twins ( > 28yrs of age) exhibited significantly higher variance between twins than younger monozygotic twins (Pearson test, P < 0.05)85. Second, McGowan et al. observed DNA methylation patters to be associated with a history of childhood abuse in the postmortem hippocampi of adults with an average age of 34 years at time of death7. Third, prenatal exposure to maternal depression has been associated with increased DNA methylation of the glucocorticoid receptor gene in the cord blood samples of infants46. Similarly, prenatal exposure to tobacco smoke was associated with small yet significant changes in global DNA methylation of epithelial buccal cells of kindergarten and first grade children75. Finally, in a longitudinal study of 15 year-old adolescents, Essex et al. found epithelial buccal cell DNA methylation patterns to be associated with maternal stress during infancy and paternal stressors during preschool years38. The above studies do not comprise the entire background literature in this subject. However, they are representative summary of the current understanding of the associations between early human experience and prolonged epigenetic patterns. Of the reviewed literature, two observations can be made. First, most of these studies considered the DNA methylation status of individuals once the individual had reached adolescence (> 15 years old), and second the majority of these studies did not use epithelial buccal  	
    71	
    cells in their evaluation of DNA methylation patterns. The following sections consider age-associated DNA methylation changes, tissue specific methylation changes, and the definition of early human experience as they relate to our study. 4.1.2  Age-associated DNA methylation changes in humans  Most human studies that have demonstrated an association between an early experiential variable and prolonged epigenetic patterns have evaluated the epigenetic profiles of their subjects at a much later age then we have in this study. There is evidence that humans experience a genome-wide decrease in DNA methylation with age86, while gene-specific CpG dinucleotides located in the promoter regions of some genes become increasingly methylated with age87. Although all of the previously reviewed studies have focused on adults, there are a few studies that have considered how much normal age-associated DNA methylation variation occurs during childhood and what the implications may be. First, in the study by Fraga et al., global and locus-specific methylation patterns between monozygotic twins found that the twins were epigenetically indistinguishable during the early stages of life, while older monozygotic twins exhibited remarkable differences in their global and loci-specific distribution of DNA methylation suggesting that DNA methylation is highly coordinated throughout early development85. Fraga and colleagues found a direct association between the epigenetic differences observed and the age of the monozygotic twins. The results suggest that older twins (defined as > 28yrs of age) are epigenetically more different between pairs than younger twins (defined as < 28yrs of age)85.  	
    72	
    In another study, Alisch et al. examined the methylation status of 27,578 CpG sites in the peripheral blood DNA from a cross-sectional study of 398 males ages 3-17 years88. They observed significant age-associated changes in the DNA methylation status at 2,078 CpG sites. This is one of the first reports of age-related DNA methylation changes from a sizable pediatric population. The findings by Alisch et al. suggest that in pediatric epigenetic studies it is especially important to closely match comparative groups for age and to adjust for age as a covariate. With regard to our study, the age range is so minimal (3-5 years), this likely is not a confounding factor that would result in a significant difference due to ageassociated methylation changes. In addition, the mean age of high and low contact caregiver-infant dyads was the same (4.4 years), leading us to believe that this difference is sufficiently small not to affect our results. However, the report by Alisch et al. provides evidence that the dynamics of DNA methylation variation are more pronounced throughout childhood than perhaps previously appreciated, and that age-associated DNA methylation changes relatively rapidly in children. Therefore, it is possible that given the highly variable nature of a child’s epigenome and our short-term follow-up time from the early experience of caregiver-infant physical contact time to the evaluation of the DNA methylation patterns, any potential associations between DNA methylation patterns and caregiver-infant physical contact time would not be detectable. It is possible that the variability of the epigenome in young children during early development would preclude us from finding any significant associations in the DNA of our sampled toddlers at the age of 5 years.  	
    73	
    Breton et al. observed differences in the global methylation levels of epithelial buccal cells in kindergarten and first grade children75. We can assume these children to be aged 4-7 years and therefore that our study samples overlap in age span. However, tobacco smoke is known to generate oxidative stress that can cause DNA lesions that can interfere with the binding of DNA methyltransferases to DNA resulting in global hypomethylation89. Therefore, given this property of tobacco, prenatal tobacco exposure is not comparable to caregiver-infant contact time as an early human infant experience that might affect epigenetic methylation patterns. Although the findings of Breton et al are supportive of epithelial buccal cell use in epigenetic studies, the findings do not preclude the possibility that epigenetic patterns associated with at least some early human experiences are not observable at the relatively early age of 5 years. We therefore suggest that as a future study this same cohort of children be evaluated following a more prolonged period--perhaps at 10-15 years of age--for epigenetic differences associated with early caregiver-infant physical contact time. Furthermore, a future study of this nature would provide an opportunity to compare the global and loci-specific DNA methylation of this cohort at two time points in their development which in itself would be an interesting study of the changes in DNA methylation in developing children. 4.1.3  Tissue specific DNA methylation changes in humans  Although the neural tissue epigenome is closest to the neurobiological processes that drive human development, human epigenetic research is necessarily limited to using peripheral tissues such as circulatory cells and biopsy tissues. Byun et al. found that DNA methylation patterns are largely conserved between tissue types and individuals, and that  	
    74	
    intra-individual DNA methylation patterns exhibit greater similarity than inter-individual DNA methylation patterns, consistent with tissue-specific DNA methylation patterns90. In this study, they collected 11 tissue types (bladder, colon, esophagus, liver, lung, pancreas, stomach, brain, heart, kidney, and spleen) from six individuals post-mortem. All of the collected tissues were from anatomically normal areas. Comparing 905 autosomal gene loci in the 66 tissue samples, Byun et al. found that the similarity of DNA methylation between tissues was greater than between individuals90. For the purposes of our study, caregiver-infant early experience and DNA methylation associations were analyzed using buccal epithelial cells. This is a histological cell type with limited previous studies examining these cells as a target tissue for social exposure-related epigenetic changes. We note that, to our knowledge, our study constitutes one of few studies of DNA methylation patterns in buccal epithelial cells associated with early human experience38,	
  75. Talens et al. reported meaningful variation in human epithelial buccal cell epigenome, stability over prolonged periods of time, and significant overlap between epigenetic markers of epithelial buccal cells and leukocytes80. In addition, as described previously in the literature review, Essex and colleagues reported measurable epigenomic change associated with early life experience (parental stress) and DNA methylation patterns in epithelial buccal cells of adolescents38. DNA methylation patterns of epithelial buccal cells, while likely not identical to epigenetic patterns of neural tissues, nonetheless should offer a reliable medium for exploratory studies of early experience associated with later changes in the epigenome.  	
    75	
    4.1.4  Definition of early experience  Our method for defining the cut-offs high and low caregiver-infant dyads was the same method used in the work of Meaney and colleagues that found DNA methylation patterns differences associated with offspring of either high- or low-LG mothers in rodents (see section 2.1 for a review). Our study was an analogous exploratory study of variations in human caregiver interactions with offspring to see if there were associations between the early caregiver-infant experience and prolonged epigenetic patterns in the offspring. We chose to consider physical contact time between the caregiver and infant as our early experience. However, it is possible that caregiver-infant contact time as an early experience was (1) not analogous to LG in rodents, and (2) not an early experience with which long-term methylation pattern outcomes in offspring are associated when within the normal range of this experience. However, regardless of the closeness of our analogy to that of Meaney and colleagues, this study did not aim to repeat the rodent study in humans, but rather to use the LG rodent studies as a reference to define two groups of subjects with a significant difference in an early experience in humans. Furthermore, the range and large difference in average daily contact time between our two caregiver-infant dyad groups encouraged us to consider daily contact time as an early experience for which there was a significant difference within the normal range amongst our human sample. There are, of course, additional early experiences that may prove interesting to explore in association with later life DNA methylation patterns, and these will likely be the basis of future studies. In our study, the difference of high and low caregiver-infant mean daily physical contact time has was > 6 hrs/day. Although this is a significant difference of contact  	
    76	
    time, our distribution of caregiver-infant physical contact time is only representative of a normal variation of a cultural norm within a metropolitan North American city. Therefore, our sample is limited in its ability to represent caregiver-infant contact time as reflected in the human experience more generally. It might be beneficial to expand the sampled experience to include a greater diversity of cultures as this would better represent variations in caregiver-infant interaction and likely expand the range of daily physical contact time observed. For example, !Kung San caregiving includes almost constant physical contact during the day and night. Konner reports that daytime caregiver-infant contact during the first 3 months of life occurs >80% of the time, compared to <25% of the time in Western cultures91. In addition, we evaluated the methylation differences associated with contact time at 5 weeks of age in the epithelial buccal cells of toddlers aged 3-5 years. Between these ages, there is a significant length of time for which we have no observations of the experience of the child. Given this study aimed to capitalize on a previous study that was not originally designed as a longitudinal behavioural study or as an epigenetic study, this limitation was unavoidable. It is therefore important to note that there is an extended period of time for which we have no information about the infant’s experience that may have had a significant affect on the developing child. In future studies, it will be important to have timely evaluations of the early experiential environment of the infant and toddler to ensure a more complete description of the early life experience of the infant.  	
    77	
    4.2  Limitations There are some limitations associated with the design and execution of our  epigenetic study exploring early experience and the associations in epigenetic patterns in humans. First, the participant sample of the study is a subset of a larger cohort. Our subset of high and low contact caregiver-infant dyads was not found to differ significantly from the larger cohort of PHI high and low contact caregiver-infant dyads (except for small although significantly different maternal education level and family income indices – see section 2.3.5). Furthermore, the associated behaviours of caregivers of high contact caregiver-infant dyads appear to cluster with similar behaviours to the entire PHI high contact caregiver-infant dyad group. Similarly, the associated behaviours of caregivers of low contact caregiver-infant dyads appear to cluster with similar behaviours to the entire PHI low contact caregiver-infant dyad sample. Therefore, a similar caregiving experience associated either high or low contact caregiving was experienced by the infants of both the PHI and our subsample groups of infants. Although it is unlikely that our subsample is substantially different from the larger sample, there remains the possibility that our sample is systematically different from the larger cohort from which they were drawn. Second, our total sample for which we performed DNA methylation analysis was only 98. Although this is a fairly significant sample size for human study, the magnitude of a DNA methylation analysis that considers close to 450,000 CpG sites may require much larger sample sizes to find associations of significance. Comparing only 98 samples for associations in 434,580 CpG sites may cause potentially significant associations to be ‘washed out’ since one is testing for significance across such a large pool of CpG sites.  	
    78	
    On the other hand, given the number of the CpG sites represented on the array (450,000 potential methylation sites) it is possible that Type 1 error might have occurred, although unlikely given our steps to minimize this. An inherent problem of analyzing such a high throughput data set is the detection of both false positives and false negatives. In this study, we considered the methylation status of 434,580 CpG sites in 98 toddlers, therefore we effectively had a test size of almost 42.6 million. The magnitude of a test of this nature is so large that type 2 error requires significant consideration. The probability of not finding any positive results purely due to the sample size is a real challenge in this area of science. As of yet we are unsure of how to address this possibility of type 2 error. One possibility is to consider the application of confidence intervals to these estimates, but how to approach the application of confidence intervals for a high throughput study of this nature remains a statistical challenge. We will attempt to address this in future studies. There are a few assumptions were made in this study that may not have been accurate. First, it is possible that 3 days of caregiver-infant physical contact data at one specific age (5 weeks) was not a sufficiently extended period of observation to be associated with variations in DNA methylation amongst toddlers. For example, Essex et al. followed the participating families from pregnancy, and measured the parental stress levels during child infancy and preschool38, an observation period of perhaps 6 years. With a longer observation of physical contact time between caregivers and infants, a more representative sample of mean daily physical contact over a longer period may be a determining factor that is associated with later DNA methylation patterns. Therefore,  	
    79	
    although challenging given the commitment required of participants, a future study that involved a more prolonged period of Baby’s Day Diary recording could be of interest. A further limitation of our study was the use of only one tissue type for DNA methylation analysis; namely, epithelial buccal cells. This was reviewed in a previous section (4.1.3). To summarize, although there is support for the use of human epithelial buccal cells for studies similar to our own, additional tissue types may provide a better sample of the epigenetic patterns associated with caregiver-infant physical contact time. In a future study, we would propose the use of multiple tissue types (i.e. epithelial buccal cells and mononuclear blood cells) to compare the associated epigenetic patterns in a tissue specific manner and provide us with an opportunity to compare the effective usefulness of epithelial buccal cells in epigenetic pediatric studies. The glucocorticoid receptor gene (Nr3c1 rodent/NR3C1 human) has been the focus of a number of human and nonhuman DNA methylation studies. In the work of Meaney and colleagues, elevated levels of maternal LG was found to be associated with decreased methylation of the glucocorticoid receptor 17 promoter site that corresponded with an elevated level of receptor expression24. Specifically, the binding site for NGFI-A (a transcription factor induced by nerve growth) was differentially methylated in the offspring of high- and low-LG mothers with offspring of low-LG mothers having reduced binding of NGFI-A to its promoter region24. In support, McGowan et al. found that methylation interfered with the NGFI-A transcription factor binding to the NR3C1 exon 1F promoter region (the human homologue), inhibiting NR3C1 promoter activity in individuals with treatment-resistant forms of major depression and was accompanied by increased HPA activity7. Considering that NR3C1 gene expression has been known to  	
    80	
    affect stress regulatory function92, this was one of the primary foci of our target gene analysis. In previous studies that have considered the methylation status of the NR3C1 gene, each study described significant methylation changes at different sites within the exon 1F promoter region of this gene. Oberlander et al. (2008) found differences in the methylation status at two CpG sites in the exon 1F promoter region to be associated with exposure to prenatal maternal depression, while McGowan et al. observed methylation differences at three different CpG sites to be associated with early experiences of child abuse7. Furthermore, Perroud et al. observed a significant association between childhood sexual abuse and methylation status of the NR3C1 gene in peripheral blood samples93. Perroud et al. found victims of sexual abuse to be more highly methylated than those that did not suffer sexual abuse 93 again at different sites than those observed by the Oberlander or McGowan groups. To illustrate, Figure 4.1 shows the nucleotide base sequence of the exon 1F region of the NR3C1 gene. The sites described in the Oberlander, Perroud and McGowan studies noted above are indicated separately in Figure 4.1. The ‘cg’ sites highlighted in yellow along the base sequence are those for which the methylation status has been considered in one of these listed studies. By comparison, there are four sites represented by probes on the Illumina 450K array associated with the 1F region and these ‘cg’ sites are underlined in figure 4.1. Sites for which there were significant methylation differences found by Oberlander, Perroud and McGowan are bolded and underlined in their respective rows. The NGFI-A binding site is also highlighted in green.  	
    81	
    Figure 4.1: Glucocorticoid receptor gene exon 1F promoter region. Adaptation of schema by Robinson et al (unpublished).  It is evident when comparing the results of these studies that there is no one CpG site of the exon 1F promoter region for which changes are consistently observed even amongst those studies that have reported an association with methylation changes with the NR3C1-1F promoter region. In fact, there is no overlap in the specific sites where methylation differences were observed among these three studies. The incomparability of these studies suggests could reflect that methylation changes associated with early experience in humans may be highly variable. Alternatively the inconsistency may reflect type I error and some may not be reproducible associations. This also makes comparability challenging between studies. In addition, an apparent limitation inherent to the Illumina 450K array itself is that it has probes for only 4 ‘cg’ sites within the exon 1F region of the glucocorticoid receptor gene. One site for which a methylation difference was observed in previous studies was the upstream site of the “cg-gggg-cg” basepair sequence of NGFI-A transcription factor binding site94. In the rodent analogue, Meaney and colleagues observed that there was no methylation difference found at the downstream “cg” location for rodent licking and grooming94. Oberlander et al. observed significant methylation differences at the human analogue of this site in infants who had been exposed to maternal depression in utero46. In the current Illuminia 450K Array, the downstream CpG site is represented  	
    82	
    (cg15910486) but the CpG site equivalent to the site previously found to have methylation status associated with LG is not represented. Although the transcription factor binding site is partially represented on the microarray, the specific site described in the previous rodent studies is absent. As a result, it would be necessary to use an alternative target approach, such as bisulfite-pyrosequence, to determine if these specific CpG sites differ in methylation status amongst high and low contact caregiver-infant children. This may be a focus of future studies. In reviewing the literature that supports our study, it is evident that the previous studies have reported methylation changes associated with early experiences at very specific sites along the genome. To explore the methylation differences associated with variation in caregiver-infant contact time at 5 weeks of age, we performed a targeted analysis on five candidate genes of interest. We considered the CpG sites represented on the Illumina 450K array for NR3C1, SLC6A4, DRD4, COMT, and BDNF. We found that none of the associated CpG sites (n = 239) for these genes had significant methylation differences between high and low contact toddlers. We targeted these genes for their involvement in the human HPA stress response pathway and the observations of previous DNA methylation studies in humans associated with early life experience (see section 1.3.6 for a review). As with the NR3C1 gene, the Illumina array may not be entirely representative of the promoter regions of each of these genes. Given the specificity of observations of human epigenetic studies associated with early life experience, it may be necessary to specifically consider the methylation status of each CpG site associated with a promoter region rather than be limited to methylation sites present on the Illumina array. This could be achieved through pyrosequencing. Alternatively, it is possible that,  	
    83	
    although we chose these genes for their association with the human HPA stress response pathway, early caregiver-infant contact time is not associated with methylation differences in the promoter regions of these specific genes. Even taking into account these limitations, the reported research contributes to the growing field of epigenetics as a first exploratory look at the later life DNA methylation patterns associated with differences in caregiver-infant physical contact time. Due to the continual advancements in genomic research and technology, there continue to be greater opportunities for enhanced study of development, behaviour, health and their relationships. These findings may be viewed as a contribution to the investigation of gene-environment interactions and ways in which experience and genomic variation interact to conjointly affect development. 4.3  Future Studies Relevant to our area of interest, namely caregiver-infant early experiences, there  are a few studies that could contribute to our understanding the relationships between early experiential environment of children and long-term health outcomes. Building on this study, we have suggested a number of future investigations that could improve on the study design and expand on our findings. First, it would be beneficial to collect and analyze blood samples of the high and low contact children to investigate whether and, if so what, epigenetic patterns may be evident from DNA of mononuclear blood cells. It would also be a benefit to have a study comparing the epigenetic patterns of epithelial buccal cells to mononuclear blood cells, and how they might be correlated in a relatively young pediatric population.  	
    84	
    Second, as discussed in the limitations (section 4.2), the current Infinium Illumina MethylationBeadChip450 array lacks the CpG site for which there is evidence of early experience associated methylation in rodents. As a result, a future study could include pyrosequencing all the collected samples to specifically examine the methylation status of the upstream site of the “cg-gggg-cg” basepair sequence of NGFI-A transcription factor binding site of the glucocorticoid receptor gene, as well as those found to have significant differences by Oberlander et al. (2008), McGowan et al. (2009) and Perroud et al (2011) to compare the status of high and low contact infants. Third, participants of the PHI study recorded a number of day-to-day behaviours and interactions that may be candidates for early experiences that could be associated with prolonged epigenetic variation in toddlers. Therefore the PHI study data set provides other early experiences (such as breast vs. formula feeding) that could be explored to determine if other experiences are related to later epigenetic modification of gene expression. Finally, it could be of interest to re-evaluate the DNA methylation patterns of the toddlers of this study in 5-10 years to see if there are changes in their DNA methylation patterns. As discussed, it is possible that as the child’s epigenome stabilizes, differences associated with early experience may become more apparent. In addition, it would provide us with an opportunity to compare the epithelial buccal cell DNA methylation patterns of a pediatric cohort over an interesting period of child development. Finally, ultimately aim is to understand the relationship between differences in early human experience and later life differences in behavioural or health outcomes of these infants. This study compared the methylation differences in toddlers associated with  	
    85	
    prior differences in caregiver-infant physical contact time. However we have yet to consider the associated behavioural differences in toddlers that previously experienced high and low contact caregiving. For a more complete understanding, it will be necessary to describe candidate toddler behaviours that might be expected to be related to earlier differences in caregiver-contact experience. In future studies, we may want to evaluate the later behavioural stress responses of toddlers who previously experienced high and low contact caregiving to see if behavioural differences between our two contact groups persist. These studies are just a few of the various additional investigations that could be pursued. As with most research, this study presents as many questions as it attempts to answer. 4.4  Summary of research To summarize, Study 1: Human infant early experience mediated by caregiver-  infant interactions, described caregiver-infant physical contact time as an early experience that exhibits a large range and which is associated with a cluster of behaviours of caregivers with high and low contact caregiving. We observed that the greatest behavioural difference between the high and low contact caregivers was the amount of daily time the caregivers spent holding their infant while their infant was neither feeding nor fussing. We then completed an exploratory analysis of the methylation patterns of epithelial buccal cells from toddlers who experienced differences in the amount of physical contact time shared between them and their caregiver as infants. We sought to fully explore the potential differences within our study sample. To do so, we used more liberal statistical cut-offs to ensure that any potential positive associations were not  	
    86	
    missed, as well as testing for gender differences and interactions, and specifically looked at the methylation patterns of regulatory regions of candidate genes. Regardless of the approach, the overall finding was that there were no cytosine-guanine dinucleotide sites of the 434,580 sites interrogated for which an average daily difference in caregiver-infant physical contact time of at least 6 hours/day at 5 weeks of age were associated with lasting differences in the methylation level in epithelial buccal cells of these infants at the age of 3-5 years. It is possible that DNA methylation regulation of genetic expression associated with early human experience is tissue specific, and therefore our analysis of epithelial buccal cells could result in an inaccurate evaluation of epigenetic patterns that are in fact present in the neural cells of toddlers who experienced high or low caregiver-infant contact time. Furthermore, it is possible that given the highly variable nature of a child’s epigenome, any potential associations between DNA methylation patterns and caregiverinfant physical contact time are not detectable in toddlers at five years of age. For these reasons, we may have missed DNA methylation patterns in epithelial buccal cells of toddlers that are in fact associated with differences in caregiver-infant physical contact time at 5 weeks of age. However, this study is a thorough look at the differences in caregiver-infant physical contact time within the normal range at 5 weeks of age and the associated later life differences in epigenetic methylation patterns as evident in the epithelial buccal cells of toddlers at 3-5 years of age. Therefore our finding of no difference in the methylation patterns of high and low contact toddlers might well stand up on future replications. Although this is a “negative” study, this project contributes a  	
    87	
    fairly definitive observation to the available studies of epigenetic markers in human developmental research. 	
    	
    88	
    Bibliography 1. Hane AA, Fox NA. Ordinary variations in maternal caregiving influence human infants' stress reactivity. Psychol Sci. 2006;17:550-556. 2. Meaney MJ. Environmental programming of phenotypic diversity in female reproductive strategies. Adv Genet. 2007;59:173-215. 3. McEwen BS. Physiology and neurobiology of stress and adaptation: central role of the brain. Physiol Rev. 2007;87:873-904. 4. Jaenisch R, Bird A. Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet. 2003;33 Suppl:245-254. 5. Robertson KD. DNA methylation and human disease. Nat Rev Genet. 2005;6:597-610. 6. Caldji C, Tannenbaum B, Sharma S, Francis D, Plotsky PM, Meaney MJ. Maternal care during infancy regulates the development of neural systems mediating the expression of fearfulness in the rat. Proc Natl Acad Sci U S A. 1998;95:5335-5340. 7. McGowan PO, Sasaki A, D'Alessio AC, et al. Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat Neurosci. 2009;12:342-348. 8. Dallman MF, Akana SF, Strack AM, Hanson ES, Sebastian RJ. The neural network that regulates energy balance is responsive to glucocorticoids and insulin and also regulates HPA axis responsivity at a site proximal to CRF neurons. Ann N Y Acad Sci. 1995;771:730-742. 9. Sapolsky RM. Stress hormones: good and bad. Neurobiol Dis. 2000;7:540-542. 10. Chrousos GP, Gold PW. The concepts of stress and stress system disorders. Overview of physical and behavioral homeostasis. JAMA. 1992;267:1244-1252. 11. Harkonmaki K, Korkeila K, Vahtera J, et al. Childhood adversities as a predictor of disability retirement. J Epidemiol Community Health. 2007;61:479-484. 12. Raison CL, Miller AH. When not enough is too much: the role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders. Am J Psychiatry. 2003;160:1554-1565. 13. Yehuda R, Bierer LM. Transgenerational transmission of cortisol and PTSD risk. Prog Brain Res. 2008;167:121-135.  	
    89	
    14. McEwen BS. Early life influences on life-long patterns of behavior and health. Ment Retard Dev Disabil Res Rev. 2003;9:149-154. 15. Meaney MJ. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu Rev Neurosci. 2001;24:1161-1192. 16. Heim C, Nemeroff CB. The role of childhood trauma in the neurobiology of mood and anxiety disorders: preclinical and clinical studies. Biol Psychiatry. 2001;49:1023-1039. 17. Liu D, Diorio J, Tannenbaum B, et al. Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal responses to stress. Science. 1997;277:1659-1662. 18. Kishiyama MM, Boyce WT, Jimenez AM, Perry LM, Knight RT. Socioeconomic disparities affect prefrontal function in children. J Cogn Neurosci. 2009;21:1106-1115. 19. Francis DD, Champagne FA, Liu D, Meaney MJ. Maternal care, gene expression, and the development of individual differences in stress reactivity. Ann N Y Acad Sci. 1999;896:66-84. 20. DENENBERG VH. Critical Periods, Stimulus Input, and Emotional Reactivity: a Theory of Infantile Stimulation. Psychol Rev. 1964;71:335-351. 21. Levine S. The ontogeny of the hypothalamic-pituitary-adrenal axis. The influence of maternal factors. Ann N Y Acad Sci. 1994;746:275-88; discussion 289-93. 22. Champagne FA, Francis DD, Mar A, Meaney MJ. Variations in maternal care in the rat as a mediating influence for the effects of environment on development. Physiol Behav. 2003;79:359-371. 23. Bale TL, Vale WW. CRF and CRF receptors: role in stress responsivity and other behaviors. Annu Rev Pharmacol Toxicol. 2004;44:525-557. 24. Weaver IC, Cervoni N, Champagne FA, et al. Epigenetic programming by maternal behavior. Nat Neurosci. 2004;7:847-854. 25. Weaver IC, Champagne FA, Brown SE, et al. Reversal of maternal programming of stress responses in adult offspring through methyl supplementation: altering epigenetic marking later in life. J Neurosci. 2005;25:11045-11054. 26. Meaney MJ. Epigenetics and the biological definition of gene x environment interactions. Child Dev. 2010;81:41-79.  	
    90	
    27. Caldji C, Diorio J, Meaney MJ. Variations in maternal care alter GABA(A) receptor subunit expression in brain regions associated with fear. Neuropsychopharmacology. 2003;28:1950-1959. 28. Pruessner JC, Champagne F, Meaney MJ, Dagher A. Dopamine release in response to a psychological stress in humans and its relationship to early life maternal care: a positron emission tomography study using [11C]raclopride. J Neurosci. 2004;24:2825-2831. 29. Hertsgaard L, Gunnar M, Erickson MF, Nachmias M. Adrenocortical responses to the strange situation in infants with disorganized/disoriented attachment relationships. Child Dev. 1995;66:1100-1106. 30. Kramer DA. Commentary: Gene-environment interplay in the context of genetics, epigenetics, and gene expression. J Am Acad Child Adolesc Psychiatry. 2005;44:19-27. 31. Bestor TH. Gene silencing. Methylation meets acetylation. Nature. 1998;393:311-312. 32. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev. 2002;16:6-21. 33. Bird AP, Wolffe AP. Methylation-induced repression--belts, braces, and chromatin. Cell. 1999;99:451454. 34. Ramchandani S, Bhattacharya SK, Cervoni N, Szyf M. DNA methylation is a reversible biological signal. Proc Natl Acad Sci U S A. 1999;96:6107-6112. 35. Klose RJ, Bird AP. Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci. 2006;31:89-97. 36. Hellman A, Chess A. Gene body-specific methylation on the active X chromosome. Science. 2007;315:1141-1143. 37. Champagne FA, Weaver IC, Diorio J, Dymov S, Szyf M, Meaney MJ. Maternal care associated with methylation of the estrogen receptor-alpha1b promoter and estrogen receptor-alpha expression in the medial preoptic area of female offspring. Endocrinology. 2006;147:2909-2915. 38. Essex MJ, Thomas Boyce W, Hertzman C, et al. Epigenetic Vestiges of Early Developmental Adversity: Childhood Stress Exposure and DNA Methylation in Adolescence. Child Dev. 2011. 39. Meaney MJ, Szyf M. Environmental programming of stress responses through DNA methylation: life at the interface between a dynamic environment and a fixed genome. Dialogues Clin Neurosci. 2005;7:103123.  	
    91	
    40. Jirtle RL, Skinner MK. Environmental epigenomics and disease susceptibility. Nat Rev Genet. 2007;8:253-262. 41. Bird A. Perceptions of epigenetics. Nature. 2007;447:396-398. 42. Murayama A, Sakura K, Nakama M, et al. A specific CpG site demethylation in the human interleukin 2 gene promoter is an epigenetic memory. EMBO J. 2006;25:1081-1092. 43. Sapolsky RM, Krey LC, McEwen BS. The adrenocortical axis in the aged rat: impaired sensitivity to both fast and delayed feedback inhibition. Neurobiol Aging. 1986;7:331-335. 44. Turner JD, Muller CP. Structure of the glucocorticoid receptor (NR3C1) gene 5' untranslated region: identification, and tissue distribution of multiple new human exon 1. J Mol Endocrinol. 2005;35:283-292. 45. McCormick JA, Lyons V, Jacobson MD, et al. 5'-heterogeneity of glucocorticoid receptor messenger RNA is tissue specific: differential regulation of variant transcripts by early-life events. Mol Endocrinol. 2000;14:506-517. 46. Oberlander TF, Weinberg J, Papsdorf M, Grunau R, Misri S, Devlin AM. Prenatal exposure to maternal depression, neonatal methylation of human glucocorticoid receptor gene (NR3C1) and infant cortisol stress responses. Epigenetics. 2008;3:97-106. 47. Davis M. Neural systems involved in fear and anxiety measured with fear-potentiated startle. Am Psychol. 2006;61:741-756. 48. Perez-Edgar K, Fox NA. Temperament and anxiety disorders. Child Adolesc Psychiatr Clin N Am. 2005;14:681-706, viii. 49. Lucki I. The spectrum of behaviors influenced by serotonin. Biol Psychiatry. 1998;44:151-162. 50. Canli T, Lesch KP. Long story short: the serotonin transporter in emotion regulation and social cognition. Nat Neurosci. 2007;10:1103-1109. 51. Hariri AR, Drabant EM, Weinberger DR. Imaging genetics: perspectives from studies of genetically driven variation in serotonin function and corticolimbic affective processing. Biol Psychiatry. 2006;59:888897. 52. Lakatos K, Nemoda Z, Birkas E, et al. Association of D4 dopamine receptor gene and serotonin transporter promoter polymorphisms with infants' response to novelty. Mol Psychiatry. 2003;8:90-97.  	
    92	
    53. Gross C, Hen R. Genetic and environmental factors interact to influence anxiety. Neurotox Res. 2004;6:493-501. 54. Beach SR, Brody GH, Todorov AA, Gunter TD, Philibert RA. Methylation at SLC6A4 is linked to family history of child abuse: an examination of the Iowa Adoptee sample. Am J Med Genet B Neuropsychiatr Genet. 2010;153B:710-713. 55. Philibert R, Madan A, Andersen A, Cadoret R, Packer H, Sandhu H. Serotonin transporter mRNA levels are associated with the methylation of an upstream CpG island. Am J Med Genet B Neuropsychiatr Genet. 2007;144B:101-105. 56. Beach SR, Brody GH, Todorov AA, Gunter TD, Philibert RA. Methylation at SLC6A4 is linked to family history of child abuse: an examination of the Iowa Adoptee sample. Am J Med Genet B Neuropsychiatr Genet. 2010;153B:710-713. 57. van IJzendoorn MH, Caspers K, Bakermans-Kranenburg MJ, Beach SR, Philibert R. Methylation matters: interaction between methylation density and serotonin transporter genotype predicts unresolved loss or trauma. Biol Psychiatry. 2010;68:405-407. 58. Oberlander TF, Grunau R, Mayes L, et al. Hypothalamic-pituitary-adrenal (HPA) axis function in 3month old infants with prenatal selective serotonin reuptake inhibitor (SSRI) antidepressant exposure. Early Hum Dev. 2008;84:689-697. 59. Devlin AM, Brain U, Austin J, Oberlander TF. Prenatal exposure to maternal depressed mood and the MTHFR C677T variant affect SLC6A4 methylation in infants at birth. PLoS One. 2010;5:e12201. 60. Lawford BR, Young R, Noble EP, Kann B, Ritchie T. The D2 dopamine receptor (DRD2) gene is associated with co-morbid depression, anxiety and social dysfunction in untreated veterans with posttraumatic stress disorder. Eur Psychiatry. 2006;21:180-185. 61. Gervai J, Novak A, Lakatos K, et al. Infant genotype may moderate sensitivity to maternal affective communications: attachment disorganization, quality of care, and the DRD4 polymorphism. Soc Neurosci. 2007;2:307-319. 62. Wong CC, Caspi A, Williams B, et al. A longitudinal study of epigenetic variation in twins. Epigenetics. 2010;5:516-526.  	
    93	
    63. Abdolmaleky HM, Thiagalingam S, Wilcox M. Genetics and epigenetics in major psychiatric disorders: dilemmas, achievements, applications, and future scope. Am J Pharmacogenomics. 2005;5:149-160. 64. Tenhunen J, Salminen M, Lundstrom K, Kiviluoto T, Savolainen R, Ulmanen I. Genomic organization of the human catechol O-methyltransferase gene and its expression from two distinct promoters. Eur J Biochem. 1994;223:1049-1059. 65. Chen J, Lipska BK, Halim N, et al. Functional analysis of genetic variation in catechol-Omethyltransferase (COMT): effects on mRNA, protein, and enzyme activity in postmortem human brain. Am J Hum Genet. 2004;75:807-821. 66. Bruder GE, Keilp JG, Xu H, et al. Catechol-O-methyltransferase (COMT) genotypes and working memory: associations with differing cognitive operations. Biol Psychiatry. 2005;58:901-907. 67. Abdolmaleky HM, Cheng KH, Faraone SV, et al. Hypomethylation of MB-COMT promoter is a major risk factor for schizophrenia and bipolar disorder. Hum Mol Genet. 2006;15:3132-3145. 68. Lubin FD, Roth TL, Sweatt JD. Epigenetic regulation of BDNF gene transcription in the consolidation of fear memory. J Neurosci. 2008;28:10576-10586. 69. Hall J, Thomas KL, Everitt BJ. Rapid and selective induction of BDNF expression in the hippocampus during contextual learning. Nat Neurosci. 2000;3:533-535. 70. Linnarsson S, Bjorklund A, Ernfors P. Learning deficit in BDNF mutant mice. Eur J Neurosci. 1997;9:2581-2587. 71. Duman RS, Monteggia LM. A neurotrophic model for stress-related mood disorders. Biol Psychiatry. 2006;59:1116-1127. 72. Lauterborn JC, Rivera S, Stinis CT, Hayes VY, Isackson PJ, Gall CM. Differential effects of protein synthesis inhibition on the activity-dependent expression of BDNF transcripts: evidence for immediateearly gene responses from specific promoters. J Neurosci. 1996;16:7428-7436. 73. Aranda-Abreu GE, Hernandez ME, Soto A, Manzo J. Possible Cis-acting signal that could be involved in the localization of different mRNAs in neuronal axons. Theor Biol Med Model. 2005;2:33. 74. Heijmans BT, Tobi EW, Stein AD, et al. Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A. 2008;105:17046-17049.  	
    94	
    75. Breton CV, Byun HM, Wenten M, Pan F, Yang A, Gilliland FD. Prenatal tobacco smoke exposure affects global and gene-specific DNA methylation. Am J Respir Crit Care Med. 2009;180:462-467. 76. Lam J, Barr RG, Catherine N, et al. Electronic and paper diary recording of infant and caregiver behaviors. J Dev Behav Pediatr. 2010;31:685-693. 77. Barr RG, Kramer MS, Boisjoly C, McVey-White L, Pless IB. Parental diary of infant cry and fuss behaviour. Arch Dis Child. 1988;63:380-387. 78. St James-Roberts I, Plewis I. Individual differences, daily fluctuations, and developmental changes in amounts of infant waking, fussing, crying, feeding, and sleeping. Child Dev. 1996;67:2527-2540. 79. Champagne FA, Francis DD, Mar A, Meaney MJ. Variations in maternal care in the rat as a mediating influence for the effects of environment on development. Physiol Behav. 2003;79:359-371. 80. Talens RP, Boomsma DI, Tobi EW, et al. Variation, patterns, and temporal stability of DNA methylation: considerations for epigenetic epidemiology. FASEB J. 2010;24:3135-3144. 81. Barr RG, Barr M, Fujiwara T, Conway J, Catherine N, Brant R. Do educational materials change knowledge and behaviour about crying and shaken baby syndrome? A randomized controlled trial. CMAJ. 2009;180:727-733. 82. Holden JM. Postnatal depression: its nature, effects, and identification using the Edinburgh Postnatal Depression scale. Birth. 1991;18:211-221. 83. Pruitt K, Brown G, Tatusova T. The reference sequence (RefSeq) database. In: McEntyre J, Ostell J, eds. The NCBI Handbook [Internet]. Bethesda (MD): National Center for Biotechnology Information; 2002. 84. Sandoval J, Heyn H, Moran S, et al. Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome. Epigenetics. 2011;6:692-702. 85. Fraga MF, Ballestar E, Paz MF, et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A. 2005;102:10604-10609. 86. Bollati V, Schwartz J, Wright R, et al. Decline in genomic DNA methylation through aging in a cohort of elderly subjects. Mech Ageing Dev. 2009;130:234-239. 87. Issa JP. Epigenetic variation and human disease. J Nutr. 2002;132:2388S-2392S.  	
    95	
    88. Alisch RS, Barwick BG, Chopra P, et al. Age-associated DNA methylation in pediatric populations. Genome Res. 2012. 89. Franco R, Schoneveld O, Georgakilas AG, Panayiotidis MI. Oxidative stress, DNA methylation and carcinogenesis. Cancer Lett. 2008;266:6-11. 90. Byun HM, Siegmund KD, Pan F, et al. Epigenetic profiling of somatic tissues from human autopsy specimens identifies tissue- and individual-specific DNA methylation patterns. Hum Mol Genet. 2009;18:4808-4817. 91. Konner MJ. Maternal care, infant behavior, and development among the !kung. In: Lee RB, DeVore I, eds. Kalahari Hunter-Gatherers, Studies of the !Kung San and their Neighbors. Cambridge, MA: Harvard University Press; 1976:218. 92. van Ijzendoorn MH, Bakermans-Kranenburg MJ, Belsky J, et al. Gene-by-environment experiments: a new approach to finding the missing heritability. Nat Rev Genet. 2011;12:881; author reply 881-c1. 93. Perroud,N., Paoloni-Giacobino,A., Prada,P., et al. Increased methylation of glucocorticoid receptor gene (NR3C1) in adults with a history of childhood maltreatment: a link with the severity and type of trauma. . 94. Weaver IC. Epigenetic programming by maternal behavior and pharmacological intervention. Nature versus nurture: let's call the whole thing off. Epigenetics. 2007;2:22-28. 95. Pollak SD, Nelson CA, Schlaak MF, et al. Neurodevelopmental effects of early deprivation in postinstitutionalized children. Child Dev. 2010;81:224-236. 95  	
    96	
    Appendix A: Supplemental Illumina Infinium Methylation Assay results for high compared to low physical contact time. The results for the 8 CpG sites with the lowest t-test p-value for methylation status associated with high and low physical contact time. No CpG sites were found for which there is a significant methylation difference following correction for false discovery using Benjamini-Hochberg (B-H). Supplemental table 1 includes both T-test statistical results and the Wilcoxon test results for the 8 CpG sites with the lowest p-value for T-test p-value prior to false discovery correction. Table 2 represents the 10 CpG sites with the lowest p-value according to Wilcoxon statistical results, none of which remained significant following B-H correction. Highlighted in grey are those that overlap with the top sites according to both T-test and Wilcoxon p-values. These are discussed in the results section 2.3.6. Supplemental Table 1: Top 8 CpG sites with the lowest p-value by T-test for Contact Group Gene  CpG site  Tstatistic  ZBTB20 TCEB3 ESPL1 HINT1 BRF1 PRNP C19orf30 (unknown)  cg17460368 cg21773314 cg25976786 cg20640983 cg22496859 cg03704756 cg07456645 cg01710670  4.68 -4.62 4.41 -4.41 -4.29 -4.16 4.09 4.10  	
    TMethylation statistics difference p-value 1.1E-05 1.2E-05 2.7E-05 2.9E-05 5.6E-05 8.9E-05 8.9E-05 9.62-05  0.016 -0.013 0.018 -0.012 -0.009 -0.011 0.011 0.246  T-test adjusted p-value 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999  Wilcoxon Wilcoxon Wilcoxon statistics p-values adjusted p-value 1724 560 1730.5 592 596 619 1596 1705  3.2E-05 1.8E-05 2.6E-05 5.1E-05 5.8E-05 1.2E-04 1.2E-03 5.8E-05  1 1 1 1 1 1 1 1  97	
    Supplemental Table 2: Top 10 CpG sites with the lowest p-value by Wilcoxon Test Gene  CpG site  Tstatistic  CYFIP1 TCEB3 ESPL1 ZBTB20 (unknown) (unknown) (unknown) HINT1 BRF1 (unknown)  cg10329699 cg21773314 cg25976786 cg17460368 cg16422814 cg16010458 cg02989453 cg20640983 cg22496859 cg01710670  3.91 -4.62 4.41 4.68 3.86 3.83 3.73 -4.41 -4.29 4.10  	
    TMethylation statistics difference p-value 1.7E-04 1.2E-05 2.7E-05 1.1E-05 2.0E-04 2.9E-04 3.6E-04 2.9E-05 5.6E-05 9.6E-05  0.023 -0.013 0.018 0.016 0.011 0.015 0.013 -0.016 -0.009 0.246  T-test adjusted p-value 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999  Wilcoxon Wilcoxon Wilcoxon statistics p-values adjusted p-value 1759 560 1730.5 1724 1719 1717 1711 592 596 1705  1.0E-05 1.8E-05 2.6E-05 3.2E-05 3.7E-05 4.0E-05 4.8E-05 5.1E-05 5.8E-05 5.8E-05  1 1 1 1 1 1 1 1 1 1  98	
    Appendix B: Supplemental results for all CpG sites associated with targeted candidate genes The Infinium HumanMethylatrion450 BeadChip from Illumina contains 450,000 CpG sites. Of these sites, 239 are associated with the cytosine-guanine dinucleotide sites of our seven targeted genes, namely, genes for the glucocorticoid receptor (NR3C1), dopamine receptor (DRD4), serotonin transporter (5 HTT), catechol-O-methyltransferase (COMT), and brain-derived neurotrophic factor (BDNF). The following table presents the statistical results of both T-test and Wilcoxon test, as well as methylation differences and Benjamini-Hochberg (B-H) adjusted p-values. The specific gene associated with this site is listed. They are in accordance with lowest T-statistic p-value. None of the 239 sites were found significant following B-H correction for false discovery. Supplemental Table 3: Results for all CpG sites associated with the targeted gene analyses.  CpG Site cg17810098 cg12516187 cg12573747 cg00997378 cg17860381 cg05874888 cg14506366 cg17988021 cg05818894 cg15014679 cg14994247 cg18773129 cg04111177 cg07624479 cg00525874 cg26128129 cg09691393  	
    Tstatistics 3.2775 -2.8510 2.5388 -2.4735 -2.2511 2.1807 -2.0385 1.9178 -1.9303 -1.9110 1.8701 1.8684 1.8408 -1.8380 -1.8267 1.7793 -1.7645  T statistic p-values 0.0015 0.0059 0.0130 0.0156 0.0283 0.0318 0.0448 0.0590 0.0590 0.0594 0.0654 0.0661 0.0687 0.0696 0.0718 0.0796 0.0816  Methylation difference -0.0053 0.0336 -0.0040 0.0139 0.0041 -0.0134 0.0168 -0.0146 0.0062 0.0281 -0.0097 -0.0045 -0.0045 0.0163 0.0349 -0.0786 0.0048  T-test adjusted p-value 0.4274 0.8697 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name COMT;TXNRD2 SLC6A3 HTT SLC6A3 NR3C1 HTT SLC6A3 HTT BDNF BDNF;BDNFOS SLC6A3 COMT NR3C1 HTT SLC6A3 HTT SLC6A3  Wilcoxon statistics 1501 786 1468 792 864 1448 900 1453 962 887 1355.5 1400 1355.5 922 989 1365 876  Wilcoxon p-values 0.0086 0.0100 0.0169 0.0113 0.0446 0.0250 0.0808 0.0227 0.1956 0.0656 0.1170 0.0585 0.1170 0.1128 0.2725 0.1017 0.0547  99	
    Wilcoxon adjusted p-values 0.8267 0.8267 0.8440 0.8267 0.9892 0.9140 0.9892 0.9140 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892  CpG Site cg12122041 cg10022526 cg03901281 cg08786822 cg11032634 cg15544235 cg12175949 cg16526509 cg23426002 cg08289189 cg15313332 cg14515049 cg25328597 cg06841846 cg18595174 cg14588307 cg08730070 cg07525751 cg07238832 cg15593398 cg20753294 cg02613510 cg27521571 cg20411756 cg10860553 cg03855291 cg23268677 cg13723431 cg08726248 cg10635145 cg12848639 cg17413943 cg12484332 cg25363289 cg15909473 cg07704699 cg21161187 cg16392193 cg05189570 cg16703956 cg14291693 cg27193031 cg04672351  	
    Tstatistics 1.7405 1.7231 -1.7207 1.7137 -1.6741 1.6754 -1.6757 -1.6529 1.6512 1.6357 -1.6305 -1.6284 -1.6138 -1.5441 -1.4788 -1.4434 1.3490 1.3401 1.3359 -1.3305 -1.3182 -1.3009 1.2955 -1.2674 1.2589 1.2490 -1.2412 -1.2393 1.2211 -1.2205 -1.2201 -1.2070 -1.2020 -1.2027 -1.1940 -1.1819 -1.1717 -1.1525 1.1399 -1.1332 -1.1240 -1.0978 -1.0948  T statistic p-values 0.0857 0.0882 0.0895 0.0948 0.0982 0.0986 0.0995 0.1018 0.1037 0.1059 0.1075 0.1076 0.1108 0.1275 0.1430 0.1530 0.1807 0.1842 0.1860 0.1884 0.1934 0.1973 0.2000 0.2086 0.2119 0.2152 0.2186 0.2187 0.2257 0.2258 0.2260 0.2321 0.2325 0.2333 0.2359 0.2413 0.2446 0.2521 0.2574 0.2614 0.2645 0.2753 0.2781  Methylation difference -0.0040 -0.0045 0.0069 -0.0245 0.0032 -0.0113 0.0033 0.0049 -0.0060 -0.0188 0.0031 0.0076 0.0025 0.0044 0.0048 0.0063 -0.0039 -0.0040 -0.0191 0.0028 0.0099 0.0204 -0.0109 0.0222 -0.0093 -0.0264 0.0025 0.0114 -0.0054 0.0135 0.0048 0.0045 0.0050 0.0105 0.0030 0.0136 0.0019 0.0061 -0.0093 0.0020 0.0043 0.0021 0.0028  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name HTT BDNF SLC6A3 HTT COMT;TXNRD2 HTT COMT;TXNRD2 SLC6A3 BDNF;BDNFOS COMT BDNF HTT BDNF SLC6A4 BDNF SLC6A3 COMT SLC6A3 BDNF;BDNFOS HTT NR3C1 BDNF COMT DRD4 SLC6A3 DRD4 COMT;TXNRD2 SLC6A3 DRD4 BDNF SLC6A3 BDNF SLC6A3 SLC6A3 HTT BDNF SLC6A3 SLC6A3 BDNF;BDNFOS SLC6A3 BDNF;BDNFOS BDNF BDNF  Wilcoxon statistics 1396 1363 932 1467 774 1361 1006 995 1327 1364 902 951 915.5 963 933 894 1321 1310 1316 947 1079 884 1242 947 1284 1309 983 999 1262.5 929 1012 1119 1044 1054 917 983 989 1006 1321 1034 995 998 1056  Wilcoxon p-values 0.0625 0.1047 0.1303 0.0173 0.0077 0.1079 0.3304 0.2921 0.1739 0.1032 0.0833 0.1693 0.1024 0.1981 0.1322 0.0735 0.1881 0.2165 0.2007 0.1604 0.6591 0.0625 0.4593 0.1604 0.2954 0.2192 0.2538 0.3056 0.3737 0.1248 0.3526 0.8812 0.4862 0.5330 0.1047 0.2538 0.2725 0.3304 0.1881 0.4418 0.2921 0.3022 0.5426  100	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.8440 0.8267 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892  CpG Site cg09494176 cg20931042 cg00232679 cg09386376 cg02762115 cg01616529 cg26949694 cg07589972 cg01418645 cg12940669 cg20471798 cg24709001 cg09492354 cg08713711 cg06979684 cg12636882 cg06839148 cg09606766 cg22708635 cg02688422 cg13557594 cg20108357 cg21625301 cg06613263 cg06684850 cg14433983 cg27107893 cg24712501 cg00107488 cg07212818 cg01583131 cg16029939 cg10420578 cg05736642 cg16586394 cg09607276 cg11361387 cg24178621 cg07579946 cg25457956 cg20592995 cg12728623 cg15999077  	
    Tstatistics -1.0787 -1.0720 1.0683 1.0643 -1.0595 1.0524 -1.0363 1.0327 -1.0021 -0.9986 -0.9974 -0.9955 0.9775 -0.9776 -0.9702 0.9583 0.9557 -0.9354 0.9316 -0.9272 0.9275 -0.9238 -0.9220 0.9214 -0.9201 0.9194 -0.8879 -0.8820 -0.8742 0.8659 -0.8636 0.8638 0.8548 0.8540 0.8369 0.8332 0.8188 -0.8164 0.8055 -0.8018 0.7889 -0.7842 -0.7830  T statistic p-values 0.2840 0.2873 0.2891 0.2902 0.2929 0.2956 0.3037 0.3057 0.3192 0.3206 0.3217 0.3231 0.3310 0.3311 0.3348 0.3409 0.3428 0.3520 0.3548 0.3567 0.3568 0.3588 0.3594 0.3600 0.3600 0.3608 0.3772 0.3802 0.3846 0.3894 0.3901 0.3904 0.3951 0.3966 0.4052 0.4069 0.4157 0.4164 0.4234 0.4255 0.4324 0.4351 0.4358  Methylation difference 0.0024 0.0018 -0.0050 -0.0198 0.0237 -0.0224 0.0030 -0.0036 0.0026 0.0029 0.0061 0.0062 -0.0015 0.0074 0.0053 -0.0048 -0.0048 0.0023 -0.0041 0.0053 -0.0047 0.0106 0.0019 -0.0138 0.0033 -0.0092 0.0138 0.0050 0.0025 -0.0133 0.0018 -0.0146 -0.0026 -0.0028 -0.0026 -0.0025 -0.0158 0.0026 -0.0073 0.0022 -0.0024 0.0020 0.0043  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name HTT DRD4 SLC6A3 DRD4 DRD4 DRD4 BDNF NR3C1 BDNF;BDNFOS HTT SLC6A3 HTT BDNF SLC6A3 BDNF;BDNFOS HTT HTT BDNF SLC6A3 SLC6A3 SLC6A3 BDNF HTT NR3C1 BDNF DRD4 NR3C1 SLC6A3 COMT;TXNRD2 DRD4 BDNF DRD4 SLC6A3 HTT NR3C1 DRD4 COMT SLC6A3 COMT BDNF SLC6A4 COMT SLC6A3  Wilcoxon statistics 1006 1011 1231 1261 914 1314 1031 1250 1026 1002 1074 1086 1258 996 1009 1257 1244 1028 1290 1045 1333.5 934 974 1216 975 1244 1069 1001 1010.5 1238 1001 1252 1228 1215 1266.5 1231 1303 1029.5 1236 1068 1218 1005 1043  Wilcoxon p-values 0.3304 0.3488 0.5093 0.3796 0.1001 0.2059 0.4289 0.4246 0.4079 0.3161 0.6329 0.6965 0.3916 0.2954 0.3413 0.3956 0.4505 0.4162 0.2757 0.4908 0.1594 0.1340 0.2275 0.5820 0.2304 0.4505 0.6072 0.3126 0.3469 0.4771 0.3126 0.4162 0.5235 0.5870 0.3582 0.5093 0.2361 0.4225 0.4862 0.6021 0.5720 0.3267 0.4817  101	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892  CpG Site cg17999358 cg23947039 cg12074493 cg06952416 cg23497217 cg06860277 cg13974632 cg06968181 cg27351358 cg06816235 cg12928379 cg22982173 cg16180821 cg09133032 cg06825142 cg15861585 cg23728864 cg20709110 cg19728345 cg00465975 cg14765933 cg08388004 cg22597340 cg12882697 cg25725890 cg08072487 cg03654618 cg02527472 cg10122187 cg24547396 cg04481212 cg01204634 cg08845721 cg26205131 cg24065044 cg07392358 cg16524332 cg04856117 cg22043168 cg21010859 cg11718030 cg19440506 cg10847032  	
    Tstatistics 0.7772 -0.7751 -0.7629 -0.7589 -0.7516 0.7509 -0.7470 -0.7444 -0.7351 -0.7284 -0.7228 -0.7158 -0.7130 -0.7107 0.7076 0.7062 -0.7062 0.7041 -0.6963 -0.6844 -0.6766 -0.6726 -0.6717 -0.6678 0.6674 0.6637 0.6512 -0.6470 0.6384 -0.6384 0.6362 0.6282 0.6284 0.6208 -0.6196 0.6189 0.6090 0.6060 -0.5951 0.5904 0.5778 0.5708 0.5660  T statistic p-values 0.4394 0.4408 0.4478 0.4500 0.4549 0.4552 0.4573 0.4585 0.4652 0.4689 0.4718 0.4763 0.4777 0.4790 0.4814 0.4818 0.4822 0.4835 0.4881 0.4966 0.5006 0.5032 0.5040 0.5061 0.5064 0.5091 0.5170 0.5197 0.5251 0.5251 0.5265 0.5315 0.5318 0.5363 0.5372 0.5383 0.5440 0.5463 0.5535 0.5563 0.5651 0.5703 0.5729  Methylation difference -0.0040 0.0015 0.0023 0.0115 0.0027 -0.0019 0.0011 0.0017 0.0035 0.0019 0.0014 0.0010 0.0030 0.0194 -0.0013 -0.0018 0.0028 -0.0110 0.0134 0.0034 0.0044 0.0017 0.0019 0.0009 -0.0010 -0.0018 -0.0015 0.0033 -0.0021 0.0050 -0.0013 -0.0022 -0.0174 -0.0008 0.0017 -0.0033 -0.0014 -0.0063 0.0009 -0.0011 -0.0017 -0.0018 -0.0018  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name SLC6A3 BDNF SLC6A4 NR3C1 BDNF COMT;TXNRD2 BDNF NR3C1 BDNF BDNF DRD4 HTT SLC6A3 DRD4 DRD4 DRD4 HTT COMT SLC6A3 COMT;TXNRD2 SLC6A3 BDNF;BDNFOS HTT SLC6A3 SLC6A4 HTT HTT BDNF COMT COMT;TXNRD2 BDNF SLC6A3 NR3C1 SLC6A3 BDNF SLC6A3 HTT COMT BDNF BDNF BDNF SLC6A3 NR3C1  Wilcoxon statistics 1269 1090 990 986 1077 1277 1080.5 1010 1090 1069 1017 1088 1010 1147 1252 1203 1079 1249 1069 1158 1083 1099 991 1013 1179 1289 1204 1061 1233 1067 1122 1224 1260 1173 1054 1215 1144 1221 1090 1277 1223 1155 1234  Wilcoxon p-values 0.3488 0.7182 0.2757 0.2630 0.6486 0.3196 0.6671 0.3451 0.7182 0.6072 0.3718 0.7073 0.3451 0.9622 0.4162 0.6486 0.6591 0.4289 0.6072 0.8985 0.6804 0.7678 0.2789 0.3563 0.7789 0.2789 0.6434 0.5671 0.5000 0.5971 0.8985 0.5426 0.3836 0.8127 0.5330 0.5870 0.9796 0.5572 0.7182 0.3196 0.5475 0.9158 0.4954  102	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9922 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892  CpG Site cg05717871 cg09926649 cg08362738 cg23619332 cg17342132 cg17306747 cg22659953 cg27449975 cg24249411 cg14694952 cg01330016 cg26840770 cg25354617 cg13149573 cg11335335 cg23501406 cg22891413 cg01135950 cg26126367 cg16257091 cg05030481 cg11432275 cg10558494 cg09362796 ch.4.132387F cg01225698 cg12448003 cg07194846 cg15462887 cg01636003 cg20954537 cg14558428 cg04607521 cg26741280 cg13764248 cg06521673 cg04598517 cg19280196 cg18867480 cg21163347 cg15710245 cg08763102 cg25855286  	
    Tstatistics 0.5410 -0.5377 0.5293 0.5259 0.5137 0.5128 -0.5077 -0.5068 -0.5019 0.4957 0.4853 0.4805 -0.4765 -0.4750 0.4724 0.4695 0.4681 -0.4580 -0.4498 -0.4489 0.4394 0.4354 -0.4301 -0.4288 -0.4227 -0.4201 -0.4152 0.4099 -0.3991 -0.3949 0.3934 -0.3886 0.3830 -0.3739 -0.3655 0.3655 -0.3641 0.3608 -0.3587 0.3527 -0.3478 -0.3447 0.3440  T statistic p-values 0.5900 0.5921 0.5978 0.6004 0.6092 0.6093 0.6130 0.6137 0.6174 0.6213 0.6288 0.6320 0.6354 0.6360 0.6379 0.6399 0.6411 0.6481 0.6542 0.6548 0.6616 0.6646 0.6686 0.6693 0.6737 0.6755 0.6790 0.6830 0.6910 0.6939 0.6949 0.6988 0.7029 0.7095 0.7156 0.7157 0.7169 0.7192 0.7207 0.7253 0.7291 0.7312 0.7320  Methylation difference -0.0101 0.0019 -0.0006 -0.0009 -0.0111 -0.0014 0.0055 0.0014 0.0016 -0.0008 -0.0020 -0.0011 0.0015 0.0008 -0.0100 -0.0017 -0.0013 0.0048 0.0029 0.0029 -0.0007 -0.0035 0.0016 0.0025 0.0017 0.0013 0.0006 -0.0033 0.0030 0.0008 -0.0006 0.0029 -0.0013 0.0026 0.0010 -0.0006 0.0011 -0.0019 0.0009 -0.0043 0.0012 0.0020 -0.0014  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name DRD4 COMT BDNF BDNF NR3C1 SLC6A3 SLC6A3 HTT BDNF HTT SLC6A4 BDNF HTT HTT DRD4 DRD4 SLC6A3 SLC6A3 SLC6A4 BDNF SLC6A3 HTT BDNF HTT HTT BDNF BDNF COMT;TXNRD2 BDNF BDNF BDNF NR3C1 HTT SLC6A4 HTT NR3C1 SLC6A3 HTT BDNF SLC6A3 BDNF HTT HTT  Wilcoxon statistics 1231 1099 1184 1171 1234 1114.5 1088 1090 1153 1111 1205 1173 1107 1115 1191 1115 1171 1148.5 1084 1084 1233 1213 1161 986 1048 1104 1077 1213 1117 1070 1195 1304 1222 1103 1086 1210 1157 1205 1059 1227 1147.5 1038 1183  Wilcoxon p-values 0.5093 0.7678 0.7511 0.8240 0.4954 0.8554 0.7073 0.7182 0.9274 0.8354 0.6381 0.8127 0.8127 0.8582 0.7127 0.8582 0.8240 0.9535 0.6857 0.6857 0.5000 0.5971 0.8812 0.2630 0.5047 0.7958 0.6486 0.5971 0.8697 0.6123 0.6911 0.2332 0.5523 0.7901 0.6965 0.6123 0.9042 0.6381 0.5572 0.5282 0.9593 0.4593 0.7567  103	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892  CpG Site cg21702128 cg05218375 cg13175282 cg15910486 cg20340655 cg07733851 cg02929434 cg18146873 cg25962210 cg03988700 cg06045576 cg13202751 cg27404921 cg17898621 cg27580375 cg13877631 cg00556112 cg25412831 cg03857453 cg24650785 cg11806762 cg21905167 cg16834011 cg23231692 cg00298481 cg14692377 cg24756227 cg26720913 cg18849621 cg03167496 cg27122725 cg15600751 cg01642653 cg03909863 cg01335087 cg11616777 cg25060525 cg06299284 cg25836061 cg12466613 cg14879948 cg25880358 cg19930203  	
    Tstatistics 0.3305 -0.3295 0.3208 -0.3205 0.3201 -0.3189 0.3139 -0.3111 0.3102 -0.3088 0.3089 0.2994 0.2982 -0.2931 0.2915 -0.2891 -0.2888 0.2874 -0.2817 -0.2813 -0.2708 -0.2709 -0.2706 -0.2592 0.2584 0.2563 0.2563 -0.2492 -0.2423 -0.2318 -0.2283 0.2274 -0.2235 0.2201 0.2171 -0.2170 -0.2115 -0.2097 0.2091 -0.2071 -0.2069 -0.2044 -0.2040  T statistic p-values 0.7419 0.7426 0.7492 0.7495 0.7499 0.7506 0.7544 0.7566 0.7573 0.7582 0.7583 0.7654 0.7665 0.7701 0.7715 0.7734 0.7739 0.7747 0.7789 0.7793 0.7872 0.7873 0.7874 0.7962 0.7968 0.7983 0.7985 0.8039 0.8093 0.8173 0.8200 0.8206 0.8237 0.8264 0.8287 0.8288 0.8330 0.8344 0.8350 0.8364 0.8367 0.8385 0.8388  Methylation difference -0.0009 0.0010 -0.0015 0.0010 -0.0007 0.0063 -0.0018 0.0010 -0.0007 0.0046 -0.0011 -0.0018 -0.0007 0.0010 -0.0009 0.0045 0.0007 -0.0007 0.0022 0.0008 0.0013 0.0009 0.0015 0.0007 -0.0004 -0.0003 -0.0017 0.0057 0.0007 0.0004 0.0018 -0.0010 0.0007 -0.0036 -0.0034 0.0006 0.0006 0.0050 -0.0007 0.0008 0.0009 0.0008 0.0012  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934  Gene name NR3C1 BDNF COMT NR3C1 BDNF NR3C1 SLC6A3 NR3C1 BDNF SLC6A3 COMT SLC6A3 HTT HTT SLC6A3 HTT DRD4 BDNF NR3C1 BDNF BDNF COMT COMT DRD4 BDNF SLC6A4 SLC6A3 NR3C1 NR3C1 BDNF NR3C1 SLC6A3 BDNF DRD4 COMT SLC6A3 HTT DRD4 COMT NR3C1 SLC6A3 HTT COMT  Wilcoxon statistics 1224 1029 1169 1097 1283 1076 1182 1126 1198 1127 1255 1109 1189 1088 1237 978 1188 1207 1053 1129 1102 1089 1143 1094 1181 1121 1171 1078 1219 1205 1110 1183 1082 1191 1160 1047 1131 1105 1151 1144 1124 1129 1080  Wilcoxon p-values 0.5426 0.4204 0.8354 0.7567 0.2988 0.6434 0.7622 0.9216 0.6750 0.9274 0.4038 0.8240 0.7236 0.7073 0.4817 0.2390 0.7291 0.6278 0.5282 0.9390 0.7845 0.7127 0.9855 0.7401 0.7678 0.8927 0.8240 0.6539 0.5671 0.6381 0.8297 0.7567 0.6750 0.7127 0.8869 0.5000 0.9506 0.8014 0.9390 0.9796 0.9100 0.9390 0.6644  104	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9922 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9922 0.9892 0.9892 0.9892  CpG Site cg16614020 cg11712482 cg03738331 cg07528216 cg10253022 cg25535999 cg16338944 cg22546130 cg14416248 cg10838500 cg03363743 cg18319433 cg18998365 cg08818984 cg15914769 cg23601416 cg19664229 cg01542143 cg26924408 cg02197303 cg04106006 cg04210284 cg05951817 cg26057780 cg16335926 cg18731680 cg24984698 cg11152298 cg07664579 cg06346307 cg18354203 cg05016953 cg03747251 cg18584905 cg08288806 cg15926585 cg13648501 cg12457376 cg11858516 cg11234429 cg22584138 cg26464411 cg27345592  	
    Tstatistics -0.1929 0.1817 -0.1779 -0.1771 -0.1668 0.1609 -0.1594 0.1587 0.1526 0.1459 -0.1433 -0.1423 -0.1379 0.1344 0.1308 0.1277 0.1217 -0.1211 0.1117 0.1112 0.1096 -0.1077 -0.1074 -0.1068 0.1047 -0.1028 0.1023 -0.0964 -0.0917 0.0876 0.0809 0.0753 0.0710 0.0663 -0.0540 0.0517 0.0501 0.0457 0.0420 -0.0360 -0.0294 0.0287 0.0210  T statistic p-values 0.8475 0.8563 0.8593 0.8599 0.8680 0.8726 0.8737 0.8743 0.8791 0.8843 0.8864 0.8872 0.8907 0.8934 0.8962 0.8987 0.9034 0.9039 0.9114 0.9118 0.9131 0.9144 0.9148 0.9152 0.9169 0.9184 0.9188 0.9235 0.9272 0.9304 0.9358 0.9402 0.9436 0.9473 0.9571 0.9589 0.9602 0.9636 0.9666 0.9714 0.9767 0.9772 0.9833  Methylation difference 0.0024 -0.0014 0.0003 0.0005 0.0004 -0.0006 0.0006 -0.0018 -0.0024 -0.0007 0.0006 0.0005 0.0027 -0.0021 -0.0003 -0.0022 -0.0010 0.0007 -0.0011 -0.0004 -0.0013 0.0001 0.0016 0.0005 -0.0003 0.0017 -0.0004 0.0001 0.0041 -0.0012 -0.0014 -0.0001 -0.0003 -0.0001 0.0001 -0.0002 -0.0003 -0.0001 -0.0002 0.0002 0.0003 -0.0001 -0.0001  T-test adjusted p-value 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9934 0.9945 0.9945 0.9948 0.9953 0.9953 0.9953 0.9953 0.9953 0.9953 0.9953 0.9953 0.9953  Gene name SLC6A3 COMT;TXNRD2 HTT NR3C1 COMT;TXNRD2 NR3C1 SLC6A3 COMT HTT SLC6A3 SLC6A4 HTT NR3C1 NR3C1 BDNF COMT SLC6A3 HTT SLC6A3 HTT BDNF SLC6A3 SLC6A4 BDNF NR3C1 COMT SLC6A4 NR3C1 SLC6A3 COMT BDNF;BDNFOS SLC6A4 BDNF SLC6A4 HTT COMT;TXNRD2 NR3C1 COMT;TXNRD2 HTT SLC6A3 SLC6A4 NR3C1 NR3C1  Wilcoxon statistics 1098 1071.5 1138 1086 1147 1175 1153 1148 1224 1197 1165 1065 1139 1103 1087 1186 1086 1153 1132 1114 1160.5 1088 1137 1149 1128 1007 1104 1212 1176 1172 1200 1213 1112 1160 1152 1157 1211 1198 1243 1136 1151 1172 1123  Wilcoxon p-values 0.7622 0.6200 0.9913 0.6965 0.9622 0.8014 0.9274 0.9564 0.5426 0.6804 0.8582 0.5870 0.9971 0.7901 0.7019 0.7401 0.6965 0.9274 0.9564 0.8525 0.8841 0.7073 0.9855 0.9506 0.9332 0.3340 0.7958 0.6021 0.7958 0.8183 0.6644 0.5971 0.8411 0.8869 0.9332 0.9042 0.6072 0.6750 0.4548 0.9796 0.9390 0.8183 0.9042  105	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9947 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9971 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9922 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9892 0.9922 0.9892 0.9892 0.9892  CpG Site cg12287055 cg26269038 cg06787004 cg21919834  	
    Tstatistics -0.0173 0.0112 0.0101 -0.0036  T statistic p-values 0.9862 0.9911 0.9919 0.9972  Methylation difference 0.0001 0.0000 0.0000 0.0000  T-test adjusted p-value 0.9953 0.9953 0.9953 0.9972  Gene name HTT SLC6A3 COMT COMT  Wilcoxon statistics 1149 1174 1109 1135  Wilcoxon p-values 0.9506 0.8070 0.8240 0.9738  106	
    Wilcoxon adjusted p-values 0.9892 0.9892 0.9892 0.9922  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0105169/manifest

Comment

Related Items