Open Collections

UBC Graduate Research

Introgression underlies adaptively significant variation and range boundaries in forest trees Suarez-Gonzalez, Adriana; Hefer, Charles A.; Lexer, Christian; Cronk, Quentin C. B.; Douglas, Carl J. 2017

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

Item Metadata

Download

Media
42591-Suarez-Gonzalez_A_et_al_Introgression_underlies_adaptively_accepted_2017.pdf [ 1.43MB ]
42591-Suarez-Gonzalez_A_et_al_Introgression_Populus_balsamifera_P_trichocarpa_accepted_2017.pdf [ 1011.17kB ]
Metadata
JSON: 42591-1.0364053.json
JSON-LD: 42591-1.0364053-ld.json
RDF/XML (Pretty): 42591-1.0364053-rdf.xml
RDF/JSON: 42591-1.0364053-rdf.json
Turtle: 42591-1.0364053-turtle.txt
N-Triples: 42591-1.0364053-rdf-ntriples.txt
Original Record: 42591-1.0364053-source.json
Full Text
42591-1.0364053-fulltext.txt
Citation
42591-1.0364053.ris

Full Text

 1  Heading: Introgression underlies adaptively significant variation and range boundaries in forest 1 trees 2  3 Twitter for ASG: @AdriSuarezGonz 4  5 Introgression from Populus balsamifera underlies adaptively significant variation and range 6 boundaries in P. trichocarpa 7   8 Adriana Suarez-Gonzalez1, Charles A. Hefer1,2, Christian Lexer3, Carl J. Douglas†, and Quentin C. B. 9 Cronk1* 10 1Department of Botany, University of British Columbia, Vancouver, V6T 1Z4, Canada 11 2Biotechnology Platform, Agricultural Research Council, Private Bag X05, Onderstepoort, 0110, 12 South Africa 13 3 Department of Botany and Biodiversity Research, University of Vienna, 1030, Austria 14 † Deceased 25 July 2016 15 *Corresponding author: 16  17 Quentin Cronk 18 Email: quentin.cronk@ubc.ca 19 Total word count (excluding summary, references and legends):  4829 No. of figures:  4 (Figs 1–4 in colour)  Summary:  200 No. of Tables:  3 Introduction:  656 No of Supporting Information files:  7 (Fig. S1– S3; Table S1– S3, script)  Materials and Methods:  1420 Results:  1052 Discussion:  1587 Acknowledgements:  69  20 Keywords: adaptive introgression, admixture mapping, epistasis, genome architecture, latitudinal 21 cline, phenomics, Salicaceae, species range 22  23  2   24 INTROGRESSION FROM POPULUS BALSAMIFERA UNDERLIES ADAPTIVELY SIGNIFICANT 25 VARIATION AND RANGE BOUNDARIES IN P. TRICHOCARPA 26 SUMMARY 27 • Introgression can be an important source of adaptive phenotypes, although conversely it can 28 have deleterious effects. Evidence for adaptive introgression is accumulating but information on 29 the genetic architecture of introgressed traits lags behind.  30 • Here we determine trait architecture in Populus trichocarpa under introgression from P. 31 balsamifera using admixture mapping and phenotypic analyses.  32 • Our results reveal that admixture is a key driver of clinal adaptation and suggest that the 33 northern range extension of P. trichocarpa depends, at least in part, on introgression from P. 34 balsamifera. However, admixture with P. balsamifera can lead to potentially maladaptive early 35 phenology and a reduction in growth and disease resistance in P. trichocarpa. Strikingly, an 36 introgressed chromosome 9 haplotype block from P. balsamifera restores the late phenology 37 and high growth parental phenotype in admixed P. trichocarpa. This epistatic restorer block 38 may be strongly advantageous in maximizing carbon assimilation and disease resistance in the 39 southernmost populations where admixture has been detected. We also confirm a previously 40 demonstrated case of adaptive introgression in chromosome 15 and show that introgression 41 generates a transgressive chlorophyll–content phenotype.  42 • We provide strong support that introgression provides a reservoir of genetic variation 43 associated with adaptive characters that allows improved survival in new environments. 44  3  INTRODUCTION 45 Gene flow between species through hybridization can be a potent evolutionary force when 46 recombination increases standing variation and creates opportunities for adaptive evolution 47 (Harrison and Larson, 2014; Hedrick, 2013). Admixture can be seen as a collision of genomes, that 48 nevertheless may be followed by the stable integration of genomic regions of one parental species 49 into the genome of another (Buerkle and Lexer, 2008). This process, called introgression, occurs 50 through hybridization and repeated backcrossing (Rieseberg and Wendel, 1993). Although 51 introgression has the potential to detrimentally disrupt the recipient genomic background, it can 52 also provide beneficial variants that result in accelerated adaptation and improved survival in new 53 environments (Clarkson et al., 2014; Norris et al., 2015; Whitney et al., 2006; Whitney et al., 2015). 54 Simulations and empirical evidence show that gene flow between species results in increased 55 standing variation available for adaptive evolution (Barrett and Schluter, 2008; Jordan, 2016), which 56 may play a key role in species’ ability to respond to a changing climate (Hamilton and Miller, 2015). 57 Standing variation is often more likely to result in adaptation and evolutionary consequences under 58 rapidly changing conditions than de novo mutations (Orr and Unckless, 2014) and this is particularly 59 important for long-lived organisms such as trees (Petit and Hampe, 2006; Savolainen and Pyhäjärvi, 60 2007).  61 Populus trichocarpa (black cottonwood) is an ecologically and economically important forest tree 62 species distributed throughout the western US and Canada, from northern California to southern 63 Alaska, and adapted to relatively humid, moist, and mild conditions west of the Rocky Mountains. 64 Along this range, P. trichocarpa exhibits variation in several adaptive traits including phenology and 65 disease susceptibility, which suggests that local adaptation plays an important role shaping the 66 genetic and phenotypic distribution of this species (La Mantia et al., 2013; McKown et al., 2013). 67  4  These traits exhibit high heritability and are strongly correlated with geoclimatic variables such as 68 latitude, day length, and temperature (La Mantia et al., 2013; McKown et al., 2013).  69 In the interior and northern parts of its range, P. trichocarpa hybridizes freely with P. balsamifera 70 where the distributions of the two species overlap (Geraldes et al., 2014; Suarez-Gonzalez et al., 71 2016). Populus balsamifera is a boreal species distributed from Alaska to Newfoundland, with high 72 frost tolerance and able to tolerate a very large range in extreme temperatures (-62˚C to 44˚C) as 73 well as to moderate annual precipitation (Richardson et al. 2014). These two sibling poplar species 74 diverged in allopatry and despite their morphological similarity and recent divergence [~76 Ka 75 (Levsen et al., 2012); but see (Ismail et al., 2012)], they are ecologically divergent and adapted to 76 strongly contrasting environments (Geraldes et al., 2014; Richardson et al., 2014). Populus 77 balsamifera, like P. trichocarpa, has been an important subject for the study of variation and 78 adaptation in trees (Keller et al., 2011a; Keller et al., 2011b; Olson et al., 2013; Soolanayakanahally 79 et al., 2013). Introgression from the northern and continental species P. balsamifera could transfer 80 advantageous traits allowing admixed P. trichocarpa individuals to colonize colder environments in 81 northern and interior locations. A targeted analysis demonstrated adaptive introgression from P. 82 balsamifera into P. trichocarpa in a subtelomeric region of chromosome 15 (Suarez-Gonzalez et al., 83 2016), and a whole genome analysis identified additional candidate regions for adaptive 84 introgression (A Suarez-Gonzalez et al., unpublished). However, the targeted study did not include 85 phenotypic information from pure P. balsamifera and it is unknown if other introgressed regions, 86 identified in the whole genome analysis, are associated with phenotype in admixed P. trichocarpa.   87 Here we implemented admixture mapping to explore if introgressed regions with unusually high 88 levels of P. balsamifera ancestry are driving variation in locally adaptive traits. We used phenotypic 89 analyses to detect if trait variation along a latitudinal gradient was associated with variation in 90  5  geoclimatic variables in admixed and pure P. trichocarpa individuals. We also explore the specific 91 effects of one introgressed region on trait variation in admixed individuals. Finally, we revisit a case 92 of adaptive introgression in a subtelomeric region of chromosome 15, found in the targeted analysis 93 (Suarez-Gonzalez et al., 2016), and use phenotypic data from pure individuals to investigate the 94 effects of this genomic region in a P. trichocarpa and P. balsamifera background.  95 METHODS 96 Samples 97 We employed data from a whole genome local ancestry analysis and from association studies in P. 98 trichocarpa Torr. & Gray (La Mantia et al., 2013; McKown et al., 2013; McKown et al., 2014a) as well 99 as phenotypic data generated specifically for this study (i.e. chlorophyll content index). The 100 accessions used were from a collection of the British Columbia Ministry of Forests, Lands and 101 Natural Resource Operations, outplanted in a common garden at the University of British Columbia 102 (Xie et al., 2009). The local ancestry analysis and chlorophyll measurements also included P. 103 balsamifera L. accessions from the Agriculture and Agri-Food Canada AgCanBaP collection 104 (Soolanayakanahally et al., 2009; Suarez-Gonzalez et al., 2016). For admixture mapping, we used 105 118 admixed individuals that showed introgression from P. balsamifera ancestry in a whole genome 106 local ancestry analysis. We focused on accessions from northern and interior parts of the P. 107 trichocarpa range to target locations where admixture has been detected (Geraldes et al., 2014; 108 Suarez-Gonzalez et al., 2016). In addition, we used reference P. trichocarpa and P. balsamifera 109 individuals to detect SNPs (Single Nucleotide Polymorphism) with fixed differences. Pure P. 110 trichocarpa individuals were used in the phenotypic analyses while data from pure P. balsamifera 111 individuals was available for only one trait (chlorophyll content index) (Figure 1, Supporting 112 Information Table S1). 113  6  Ancestries and phenotypes 114 In the local ancestry analyses, SNPs were called in the whole genome of 168 individuals (25 115 reference individuals of each parental species and 118 admixed P. trichocarpa individuals with P. 116 balsamifera admixture) selected from the sympatric zone between P. trichocarpa and P. 117 balsamifera as well as from allopatric populations (Supporting Information Figure S1, (Suarez-118 Gonzalez et al., 2016)). Each of the genotypes was sequenced at an expected coverage ranging from 119 15X to 30X using the Illumina HiSeq2000 platform (SRA: PRJNA276056). Each SNP was annotated 120 using SNPeff (Cingolani et al., 2012) with version 3 of the P. trichocarpa genome assembly.  121 We estimated the probabilities for each of the possible ancestral configurations (P. balsamifera, P. 122 trichocarpa or mixed ancestry) in every SNP across the whole genome of 118 admixed individuals 123 using RASPberry, a software that implements a reliable Hidden Markov model (HMM) for admixture 124 (Wegmann et al., 2011), following the same pipeline and parameters as in our previous study on 125 adaptive introgression (Suarez-Gonzalez et al., 2016). Briefly, SNPs with missing data in the parental 126 genotypes were removed using Plink 1.07 (Purcell et al., 2007), the parental genotypes (50) were 127 then phased with fastphase (Scheet and Stephens, 2006) by creating the input files with FCGENE 128 (Roshyara and Scholz, 2014), and ancestries for each admixed individual were estimated in 129 ADMIXTURE (Alexander et al., 2009).  130 To determine the ancestral configurations of each SNP in one of the three categories, ancestries 131 were considered for probabilities >95%. The proportion of introgressed ancestry in admixed 132 individuals was calculated, for each SNP, by counting sites with introgressed ancestry as two and 133 sites with mixed ancestry as one. We detected regions with unusually high levels of introgression in 134 admixed individuals using a sliding window approach and a significance cut-off of three standard 135  7  deviations (SD) from the weighted mean across all the chromosomes based on SNP density per 136 window (100-kb, steps of 20-kb) (Suarez-Gonzalez et al., 2016).  137 For phenotypic data, we used 59 traits encompassing phenological events, biomass accumulation, 138 growth rates, disease susceptibility, as well as leaf, isotope and gas exchange-based ecophysiology 139 from previous studies (La Mantia et al., 2013; McKown et al., 2013; McKown et al., 2014a). This 140 dataset was collected throughout 2008 to 2012 from 461 P. trichocarpa accessions with 4 to 20 141 clonal replicates similar in age and condition, grown as stecklings under glasshouse conditions, and 142 then out-planted in a common garden at Totem Field, University of British Columbia, Canada. In 143 2015, we measured chlorophyll absorbance on leaves for the chlorophyll content index (CCI) using a 144 CCM-200 plus SPAD chlorophyll meter (Opti-Sciences Inc., Hudson, NH, USA) in P. trichocarpa and P. 145 balsamifera accessions in Totem Field.  146 Admixture mapping 147 To explore the effects of introgression on the genetic architecture of adaptation in P. trichocarpa, 148 we used a Bayesian method called BMIX (Shriner et al., 2011). BMIX empirically estimates the 149 testing burdens of admixture mapping by fitting an autoregressive model and estimating the 150 effective number of tests based on autocorrelation. Since a block of ancestry from one parental 151 population can be up to several megabases long, local ancestry estimates can be highly correlated 152 in a genome. First, we estimated the number of effectively independent tests for each chromosome 153 for each individual by fitting an autoregressive model to the local ancestries (Plummer et al., 2010). 154 The spectral density was estimated at frequency zero and the order of the fitted autoregressive 155 models was chosen by minimizing the Akaike information criterion. The effective number of tests 156 were then summed for the chromosomes of each individual and averaged across individuals. Next, 157 the phenotypes were regressed on local ancestry, adjusting for global ancestry using generalized 158  8  linear models (GLMs). Global ancestry was calculated for each individual as the local ancestry 159 averaged across all markers. Finally, the p-values from the regression models were converted into 160 posterior probabilities. The threshold was the value at which the hypothesis favored by the 161 posterior probability switches (i.e. 0.5) (Supporting Information Notes S1, (Shriner et al., 2011)).  An 162 R script was used to run the GLMs and convert the p-values to posterior probabilities (Supporting 163 Information, Shriner et al., 2011).    164 We focused on 19 regions with unusually high levels of P. balsamifera introgression identified by 165 the whole genome local ancestry analysis (Supporting Information Figure S2). Overall, 107 166 individuals with both local ancestry and phenotypic data were used in the admixture mapping 167 analysis. We focussed all downstream phenotypic analyses on ancestry-trait associations showing 168 very strong signals (i.e. posterior probabilities higher than 0.9), on SNPs fixed for different alleles in 169 the parentals, and on a region that showed signals of adaptive introgression in the genomic and 170 functional study (Suarez-Gonzalez et al., 2016). 171 Phenotypic analysis of associations from admixture mapping  172 To explore the effects of P. balsamifera alleles on the phenotype of admixed P. trichocarpa, we 173 selected SNPs that had both fixed differences in the pure species and displayed associations with 174 traits in the BMIX analysis. We also included all the fixed SNPs from the subtelomeric introgressed 175 region on chromosome 15, a candidate region for adaptive introgression associated with 176 chlorophyll levels (Suarez-Gonzalez et al., 2016). Then, haplotypes in the admixed individuals were 177 inferred with fastphase (Scheet and Stephens, 2006) by creating the input files of fixed SNPs with 178 FCGENE (Roshyara and Scholz, 2014). To identify P. balsamifera and P. trichocarpa haplotypes, we 179 performed neighbor-joining (NJ) analyses [1000 bootstrap replicates in MEGA (Tamura et al., 2007)] 180 including admixed and pure individuals (Supporting Information Figure S3). Each of the admixed P. 181  9  trichocarpa individuals was classified into one of three genotypic categories: homozygotes for P. 182 balsamifera haplotypes (bb), heterozygotes (bt) and homozygotes for P. trichocarpa haplotypes (tt), 183 based on phased genomic sequences. Finally, phenotypic traits showing associations with the 184 introgressed regions were compared among the three genotypic categories (bb, bt, and tt) in each 185 of the haplotypes using analysis of variance (ANOVA) with adjusted p-values with Bonferroni 186 correction. For the analysis of variance on the chlorophyll content index of 2009, we included 187 individuals that set bud after the summer solstice that year (prior to day 186 were removed), since 188 following bud set, trees have a greater amount of chlorophyll on average compared to trees still 189 within an active growing phase (McKown et al., 2016). For 2011 and 2015 data on bud set was not 190 recorded.  191 Climate of tree origin in admixed and pure individuals 192 To determine if trait variation was correlated with climate and admixture, we compiled eight 193 climate variables associated with moisture and 15 variables related to temperature from ClimateNA 194 (Wang et al., 2012) based on 1971–2000 (Supporting Information Table S2). We performed two 195 principal component analyses (PCA), one for moisture and one for temperature, using the function 196 prcomp in R. In addition, the maximum length of day (DAY; h) was also calculated at each location 197 as a proxy for photoperiodic regime using the package geosphere in R v. 3.3.2 (R Development Core 198 Team, http://www.r-project.org). DAY was not used in the PCA analysis.  199 Enrichment analysis  200 To detect overrepresented biological terms in an introgressed haplotype associated with trait 201 variation, we performed enrichment tests for various terms including Gene Ontology (GO) and 202 Protein Family (Pfam) using Popgenie (Sjödin et al., 2009). The list of introgressed genes was 203 compared with the list of all poplar genes (41,335) and the best-annotated orthologs in Arabidopsis 204  10  thaliana were identified based on Popgenie annotation using Fisher´s exact test and the default p-205 value threshold (0.05). 206 RESULTS 207 Introgressed regions from P. balsamifera are associated with trait variation in P. trichocarpa 208 Across the whole genome (1,168,955 SNPs examined) of 118 admixed P. trichocarpa individuals, we 209 detected 19 regions with unusually high levels of introgression (1107 genes) with P. balsamifera 210 ancestry (Supporting Information Figure S2). These regions included candidate genes for adaptive 211 introgression previously identified by our previous targeted study (Suarez-Gonzalez et al., 2016). 212 The 19 regions, occurring across 11 chromosomes, showed P. balsamifera ancestry peaks with a 213 height (i.e. percentage of P. balsamifera ancestry) ranging from 0.1778 to 0.2733 and width ranging 214 from 140 kb to 1.02 Mb.  215 The admixture mapping analysis revealed 86,150 SNPs associated with 57 traits (across multiple 216 years) with posterior probabilities above 0.5 (Table 1). In the 19 regions with unusually high levels 217 of P. balsamifera introgression identified in the local ancestry analysis, we detected 18,525 218 phenotype associations in 2,874 SNPs across seven chromosomes (5, 6, 9, 10, 14, 15 and 17). 219 Introgressed regions on chromosome 5, 6, 15 and 17 showed associations with only one to four 220 phenotypic traits, while regions on chromosomes 9, 10 and 14 showed associations with 15 to 20 221 different traits (Table 2). Introgressed regions on chromosomes 9, 10 and 14 also displayed the 222 strongest associations (i.e. posterior probabilities above 0.9), comprising 3,429 SNPs and 20 trait 223 associations (12 traits some measured in multiple years). In addition, the only introgressed region 224 showing associations with disease resistance was located on chromosome 9 (Table 3).  225 A total of 143 SNPs located on chromosomes 9 (119 SNPs), 10 (18 SNPs) and 14 (6 SNPs) displayed 226 both fixed differences in the pure species and associations with traits in BMIX. Neighbor-joining 227  11  analyses, based on haplotypes from these fixed differences, revealed clear clusters of P. balsamifera 228 and P. trichocarpa haplotypes and allowed us to classify admixed P. trichocarpa individuals as 229 heterozygous or homozygous for P. balsamifera and P. trichocarpa haplotypes (Supporting 230 Information Figure S1).  231 Admixture drives adaptively relevant trait variation across a latitudinal gradient  232 The traits showing the strongest associations with the introgressed regions were related to 233 phenology, biomass traits (e.g. height and bole volume), and disease resistance. These traits 234 showed strong correlations with maximum day length (DAY) and temperature (first principal 235 component based on climate variables associated with temperature explaining 60% of the 236 variance), which had significant interactions with the presence of admixture (i.e. significant 237 differences between the slopes of pure and admixed individuals, p<0.05; Figure 2). Admixed 238 individuals from colder environments had earlier phenology (i.e. bud set, leaf yellowing, and leaf 239 drop), were smaller, and had more damage by disease than those from southern regions. The same 240 types of correlations were found with maximum day length (DAY), where individuals from higher 241 latitudes (greater maximum daylength) had earlier phenology (i.e. bud set, leaf yellowing, and leaf 242 drop), and were smaller. An exception to this pattern, however, was damage by disease, which did 243 not increase with maximum day length. Curiously, in pure P. trichocarpa we did not detect a 244 relationship between phenotypic traits and DAY or temperature. Pure P. trichocarpa individuals 245 from northern regions had similar phenotypes to pure individuals from southern regions (Figures 2 246 and 3). Also, the phenotype of pure P. trichocarpa individuals was similar to that in admixed 247 individuals from lower latitudes (i.e. later phenology, greater size, less damage by disease) (Figure 248 2). The low statistical power in pure individuals, due to the small sample size at higher latitudes and 249 colder environments, could explain the lack of association between traits and geoclimatic variables. 250  12  Climate variables associated with moisture (first principal component explaining 63% of the 251 variance) did not show significant interactions with admixture. Furthermore, the whole genome 252 levels of admixture were generally not correlated with trait variation (Pearson's correlation, 253 p>0.05).  254 An introgressed haplotype on chromosome 9 restores parental phenotype in admixed P. trichocarpa   255 We focused on the introgressed ancestry block on chromosome 9 since this haplotype showed the 256 strongest and highest number of associations. ANOVAs based on a suite of phenology, biomass, and 257 disease resistance traits in admixed individuals with different haplotypes on chromosome 9 (Table 258 3) supported the results from BMIX (Tables 1 and 2) in most cases. All traits that showed significant 259 differences among haplotypes, after Bonferroni correction, revealed the same trend: admixed 260 individuals with two copies of the P. balsamifera haplotype in chromosome 9 were more similar to 261 pure P. trichocarpa, from northern and southern populations, than to other admixed individuals 262 without signals of introgression in this haplotype. Admixed individuals homozygous for the 263 introgressed haplotype in chromosome 9 had later phenology (bud set day mean: 222.5 Julian day, 264 2010), were taller (mean: 370.3 cm, 2010), and were more resistant to disease than admixed 265 individuals without P. balsamifera chromosome 9 haplotypes (mean height: 189.0 cm; bud set: 266 175.8 Julian day, 2010) (Figure 3). Phenotypes from heterozygous individuals were, for the most 267 part, midpoints of those from homozygous individuals. This result suggests that the introgressed 268 haplotype in chromosome 9 restores the parental P. trichocarpa phenotype.   269 Populus balsamifera haplotypes on chromosome 9 were geographically limited to the interior and 270 northwestern BC, with homozygotes for the P. balsamifera haplotypes occurring only in the Prince 271 George region, and heterozygotes also in Terrace and north of Juneau (Figure 1). Prince George was 272  13  the southernmost population, with the smallest value for maximum day length (“DAY”), from those 273 where admixture has been detected.  274 The introgressed haplotype on chromosome 9 was enriched for genes coding for two types of 275 protein families: PF07690, Major Facilitator Superfamily (Nitrate transporters NRT2: 276 Potri.009G008500 and Potri.009G008600) and PF00005, ABC transporter (ABC transporters: 277 Potri.009G007800, Potri.009G008200) as well as for three miRNAs [ptc-miR473 (Cleavage), ptc-278 miR6448 (Translation), ptc-miR172 (Cleavage)].  279 Introgression in the subtelomeric region of chromosome 15 confirms signals of adaptive introgression 280 We inferred haplotypes for the subtelomeric introgressed region on chromosome 15, already 281 known as a strong candidate region for adaptive introgression (Suarez-Gonzalez et al., 2016), using 282 132 fixed SNPs. Introgressed P. balsamifera haplotypes showed strong associations with the 283 chlorophyll content index across multiple years. The BMIX analysis showed associations in 284 measurements from 2015, but ANOVAs based on haplotypes revealed significant differences among 285 genotypes also in measurements from 2009 and 2011 (Figure 4). As in the targeted local ancestry 286 analysis (Suarez-Gonzalez et al., 2016), admixed individuals with P. balsamifera haplotypes in the 287 subtelomeric region of chromosome 15 exhibited higher values of the chlorophyll content index 288 compared to other admixed and pure individuals. The chlorophyll content index in pure P. 289 balsamifera individuals was similar to that in pure P. trichocarpa.  290 DISCUSSION 291 Admixture mapping connects phenotypic traits with genomic regions  292 The admixture mapping analysis in admixed P. trichocarpa individuals revealed numerous loci 293 introgressed from P. balsamifera that underlie adaptively-relevant variation in phenology, biomass, 294 disease resistance, and ecophysiology traits. These results give strong support that introgressive 295  14  hybridization in late generation backcrosses is providing a reservoir of new genetic variation 296 associated with adaptive characters in P. trichocarpa. Some of these traits are critical for survival in 297 northern environments, but it is possible that others are traits developmentally correlated with 298 these, and of less intrinsic importance. Many of the traits used here show strong correlations and 299 this has been discussed elsewhere (McKown et al., 2014). One corollary of this is when a particular 300 SNP is reported as having multiple trait associations, these may not all be independent but may be 301 the result of developmentally correlated traits. 302 In this study, seven of the 19 introgressed regions with unusually high levels of P. balsamifera 303 admixture, uncovered in the local ancestry analysis, were strongly associated with traits affecting 304 phenology, growth, response to disease, and ecophysiology in P. trichocarpa background. Two 305 introgressed regions, in particular, one on chromosome 9 and another on chromosome 15, showed 306 particularly strong and interesting associations with trait variation.  307 In discussing these analyses the issue of geographic structure should be noted (e.g. Geraldes et al., 308 2014) as this is a potential confounding factor. However, structure is unlikely to be a problem here 309 for the following reasons. Firstly, our trait mapping focused on samples from a limited latitudinal 310 range. Secondly, our mapping models were based on interspecific genetic ancestries rather than 311 raw genotype data (c.f. admixture mapping), which effectively mitigates the confounding effect of 312 within-species geographic structure. Thirdly, between species structure was robustly accounted for 313 because variation in genomic background at the between-species level was taken into account 314 during mapping. 315  316  15  Introgression extends the species range of P. trichocarpa by driving clinal adaptation of phenology traits 317 Numerous studies have shown that in trees from northern climates, bud set and growth cessation 318 are initiated earlier compared to their southern counterparts because the former are adapted to 319 the critical day length of northern regions (Evans et al., 2016; Hall et al., 2007; Rohde et al., 2011; 320 Vitasse et al., 2014). Earlier reports on trait variation in these P. trichocarpa individuals confirmed 321 these findings and revealed that most phenology events, including bud set, leaf yellowing, and leaf 322 drop were strongly associated with geoclimatic variables such as maximum day length (“DAY”) and 323 temperature (McKown et al., 2013). These previous results illustrate the adaptive importance of 324 phenology traits, and here we identified admixture as a key driver of this clinal adaptation.  325 Populus trichocarpa occurs from California to Alaska, and the northern range extension of P. 326 trichocarpa appears to be dependent, at least in a substantial part, on introgression of alleles from 327 P. balsamifera, which is a boreal species distributed from Alaska to Newfoundland. The whole 328 genome local ancestry analysis demonstrated a strong correlation between admixture and climate, 329 showing increased levels of P. balsamifera introgression in colder and drier environments (A Suarez-330 Gonzalez et al., unpublished). The phenotypic analyses support this finding and show that 331 admixture, as well as introgression of certain genomic regions (e.g. haplotypes on chromosome 9) 332 has strong pleiotropic effects on phenotypes that appear to play important roles in adaptation in 333 northern populations of P. trichocarpa.  334 An introgressed haplotype on chromosome 9 restores parental phenotype at certain latitudes 335 Although admixture with P. balsamifera clearly confers some adaptive traits on P. trichocarpa (i.e. 336 early fall phenology in the north), there is a trade-off since admixture has a tendency to reduce the 337 growth of the fast-growing P. trichocarpa and reduce disease resistance. Curiously, a particular P. 338 balsamifera haplotype block on chromosome 9, when introgressed, has the ability to restore the 339  16  parental P. trichocarpa phenotype (i.e. later phenology, higher growth and greater disease 340 resistance). This is not a latitudinal effect, where the haplotype block comes from trees at a 341 different latitude to the recipient trees, so mitigating traits. Rather, this is evident regardless of the 342 latitude of the haplotype donor or recipient. The introgressed “restorer haplotype block” on 343 chromosome 9 affected the most traits, with the highest number of associations with biomass (4), 344 phenology (5) and damage by disease (1) traits (in multiple years) and with the strongest 345 association signals (posterior probabilities > 0.9). Admixed individuals homozygous for introgressed 346 P. balsamifera loci in chromosome 9 were more similar to pure P. trichocarpa individuals (from 347 northern or southern regions) than to other admixed individuals heterozygous or homozygous for 348 the P. trichocarpa haplotype in chromosome 9. Homozygotes for the P. balsamifera haplotype in 349 chromosome 9 were found only in the southernmost location where admixture has been detected, 350 and had later phenology, were bigger and had less damage by disease compared to admixed 351 individuals homozygous for this P. trichocarpa haplotype.  352 Our results suggest that admixture contributes to adaptation at high latitudes, where cold is 353 combined with northern light regimes, but could be maladaptive in continental regions at lower 354 latitudes where cold is combined with southern light regimes. In Prince George, the population with 355 the shortest DAY from those where admixture has been detected, introgression of the haplotype on 356 chromosome 9 restores the parental phenotype. Therefore, an allele from a species with expected 357 early dormancy is conferring delayed dormancy when in the alternative genetic background. Later 358 phenology (i.e. delayed dormancy) in trees with these introgressed haplotypes could be an adaptive 359 strategy as it maximizes carbon assimilation with late dormant bud formation and delayed leaf 360 senescence, provided that it does not increase the risk of damage caused by early frosts on 361 vegetative organs (Chuine, 2010).  362  17  The physiological mechanisms underlying phenotype associations of this restorer block are of 363 potentially great interest. It is possible that certain introgressed alleles, via epistatic gene 364 interactions, allow admixed trees to respond to changes in the environment by tracking a 365 ‘phenological optimum’ more efficiently than other admixed individuals. This epistasis could include 366 interactions between P. trichocarpa alleles in chromosome 09 with introgressed P. balsamifera 367 alleles in other part(s) of the genome or interactions between introgressed P. balsamifera loci in the 368 absence of P. balsamifera introgression in the chromosome 9 locus. Future molecular studies and 369 common gardens including pure P. balsamifera individuals in the latitude of origin could be used to 370 further understand the role of this introgressed haplotype and epistatic interactions in clinal 371 adaptation in P. trichocarpa. 372 Adaptive introgression increases plant disease resistance 373 Admixed P. trichocarpa individuals with homozygous P. balsamifera haplotypes on chromosome 9 374 also showed less damage by disease than other admixed individuals, which could help explain their 375 higher height and bole volume (in addition to their longer growing period and the faster active 376 growth rate). Climate change, particularly at higher latitudes, can influence pathogen 377 aggressiveness. Pathogens may thus become increasingly important (Pachauri and Meyer, 2014; 378 Sturrock et al., 2011), especially for P. trichocarpa from northern populations. If adaptive 379 introgression of P. balsamifera on chromosome 9 indeed increases survival under rust attacks, 380 these admixed P. trichocarpa individuals could represent an important resource for forest 381 management and breeding programs. The conservation of natural hybridization represents an 382 underutilized option (Hamilton and Miller, 2015; Johnson et al., 2010) and our work contributes 383 empirical evidence to consider when assessing the conservation value of introgression.  384  18  Candidate genes for adaptive introgression 385 The introgressed region on chromosome 9 was enriched for genes from two protein families: Major 386 Facilitator Superfamily (PF07690) and ABC Transporter (PF00005). Two orthologs of the nitrate 387 transporter AtNRT2 (Potri.009G008500, Potri.009G008600), from the Major Facilitator Superfamily 388 (PF07690), were associated with a number of phenological (bud set, yellowing, leaf drop, growing 389 period, height growth cessation) and biomass traits (height, height gain) here and in a previous 390 association analysis (McKown et al., 2014b). One of the AtNRT2 orthologs (Potri.009G008600), 391 which was also associated with damage by disease in a different association study (La Mantia et al., 392 2013), showed exceptionally strong correlations with geoclimate variables including latitude and 393 temperature (Geraldes et al., 2014; Porth et al., 2015), and had SNPs that were Fst outliers in a 394 landscape genomic analysis (Geraldes et al., 2014). These previous reports used the same P. 395 trichocarpa populations but a different genotyping method and association mapping approach than 396 those used in this study. 397 Nitrate is essential for plant development and NRT2 genes are critical responders to both abiotic 398 and biotic stress (Gojon et al., 2011). In Brassica, expression of NRT2.1 was adversely affected under 399 abiotic stress conditions and in Arabidopsis loss of function mutants showed reduced disease 400 susceptibility to a bacterial pathogen (Pseudomonas syringae) (Camañes et al., 2012). In Poplar, 401 NRT2.1 genes are expressed more strongly in roots than in aerial tissues and are down-regulated in 402 early spring (Sjödin et al., 2009). These genes could function in nitrogen regulation during seasonal 403 remodeling of tree phenology related to the interphase between growth and dormancy (Footitt et 404 al., 2013; Forde, 2000; Larisch et al., 2012).  405  19  Introgressed alleles in a subtelomeric region of chromosome 15 create a transgressive phenotype in P. 406 trichocarpa background 407 This whole genome approach supports the findings from the targeted ancestry analysis showing 408 associations between a subtelomeric region on chromosome 15 and the chlorophyll content index 409 (Suarez-Gonzalez et al., 2016). Here, phenotypic analyses revealed that P. balsamifera alleles in a P. 410 trichocarpa background have an additive effect, with P. trichocarpa individuals homozygous for the 411 introgressed haplotypes showing the highest chlorophyll content index. In the P. balsamifera 412 background, these alleles seem to be acting in a different manner since the chlorophyll content 413 index was lower compared to those in P. trichocarpa admixed individuals with the same alleles. In 414 fact, the chlorophyll content index in pure P. balsamifera was similar to those in pure P. trichocarpa. 415 This suggests that introgressed genes interact with the genetic background of the receiving species 416 and result in a transgressive phenotype (Dittrich-Reed and Fitzpatrick, 2012). It is possible that P. 417 balsamifera transcription factors in chromosome 15 (e.g. PRR5, Potri.015G002300) interact with P. 418 trichocarpa (or P. trichocarpa and introgressed P. balsamifera) transcriptional enhancers and 419 repressors or paralogs in a different manner compared to P. trichocarpa alleles. Since phenology 420 traits were not associated with this subtelomeric region, we ruled out bud set timing as a potential 421 artifact explaining the greater chlorophyll content index in admixed P. trichocarpa individuals. If the 422 chlorophyll content index is associated with higher photosynthetic rates, faster growth, and higher 423 carbon acquisition, as is the case in P. balsamifera (Soolanayakanahally et al., 2009), this phenotype 424 could be of adaptive importance to counteract shorter growing seasons in P. trichocarpa 425 populations at higher latitudes.  426 Overall, our study provides strong support that introgressive hybridization from P. balsamifera 427 generates a reservoir of new genetic variation associated with adaptive characters that may allow 428  20  improved survival in northern regions of the P. trichocarpa range. More generally our results 429 support a growing consensus placing introgression as a central evolutionary process (Arnold and 430 Kunte, 2017). 431 ACKNOWLEDGEMENTS 432 This work was supported by the Genome Canada Large-Scale Applied Research Program (POPCAN, 433 project 168BIO), funds to QCBC and CJD, by a Natural Sciences and Engineering Research Council of 434 Canada (NSERC) Discovery Grant to CJD (RGPIN 36485-12), by NSERC Discovery Grant (RGPIN-2014-435 05820) to QCBC, and by grants from the Swiss National Science Foundation (SNF) to CL. We thank 436 Jaroslav Klápště, Armando Geraldes, Rob Guy and Sally Otto for assistance and discussions. We 437 would also like to thank two anonymous reviewers for their very valuable and detailed comments. 438  439 AUTHOR CONTRIBUTIONS 440 AS-G, CJD, QCBC, and CL designed the study. AS-G performed the research, analyzed data, and 441 wrote the manuscript. CH performed SNP calling. CJD, QCBQ, and CL provided funding. QCBC and CL 442 edited the manuscript and all the authors approved the final version. 443 REFERENCES 444 Alexander, D.H., J. Novembre and K. Lange. 2009. Fast model-based estimation of ancestry in unrelated 445 individuals. Genome Res. 19:1655-1664. 446 Barrett, R.D.H. and D. Schluter. 2008. Adaptation from standing genetic variation. Trends in Ecology & 447 Evolution 23:38-44. 448 Buerkle, C.A. and C. Lexer. 2008. Admixture as the basis for genetic mapping. Trends in Ecology & 449 Evolution 23:686-694. 450  21  Camañes, G., V. Pastor, M. Cerezo, J. Garcia-Andrade, B. Vicedo, P. Garcia-Agustin and V. Flors. 2012. A 451 deletion in NRT2.1 attenuates pseudomonas syringae-induced hormonal perturbation, resulting in 452 primed plant defenses. Plant Physiol. 158:1054-1066. 453 Chuine, I. 2010. Why does phenology drive species distribution? Philosophical Transactions of the Royal 454 Society B: Biological Sciences 365:3149-3160. 455 Cingolani, P., A. Platts, L.L. Wang, M. Coon, T. Nguyen, L. Wang, S.J. Land, X. Lu and D.M. Ruden. 2012. A 456 program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff. Fly 457 6:80-92. 458 Clarkson, C.S., D. Weetman, J. Essandoh, A.E. Yawson, G. Maslen, M. Manske, S.G. Field, M. Webster, T. 459 Antão, B. MacInnis, D. Kwiatkowski and M.J. Donnelly. 2014. Adaptive introgression between 460 anopheles sibling species eliminates a major genomic island but not reproductive isolation. Nature 461 Communications 5:4248. 462 Dittrich-Reed, D. and B.M. Fitzpatrick. 2012. Transgressive hybrids as hopeful monsters. Evolutionary 463 Biology 40:310-315. 464 Evans, L.M., S. Kaluthota, D.W. Pearce, G.J. Allan, K. Floate, S.B. Rood and T.G. Whitham. 2016. Bud 465 phenology and growth are subject to divergent selection across a latitudinal gradient in Populus 466 angustifolia and impact adaptation across the distributional range and associated arthropods. Ecol 467 Evol 6:4565-4581. 468 Footitt, S., Z. Huang, H.A. Clay, A. Mead and W. Finch-Savage. 2013. Temperature, light and nitrate 469 sensing coordinate Arabidopsis seed dormancy cycling, resulting in winter and summer annual 470 phenotypes. The Plant Journal 74:1003-1015. 471  22  Forde, B.G. 2000. Nitrate transporters in plants: Structure, function and regulation. Biochimica Et 472 Biophysica Acta (BBA) - Biomembranes 1465:219-235. 473 Geraldes, A., N. Farzaneh, C.J. Grassa, A.D. McKown, R.D. Guy, S.D. Mansfield, C.J. Douglas and Q.C.B. 474 Cronk. 2014. Landscape genomics of Populus trichocarpa: The role of hybridization, limited gene 475 flow, and natural selection in shaping patterns of population structure. Evolution 68:3260-3280. 476 Gojon, A., G. Krouk, F. Perrine-Walker and E. Laugier. 2011. Nitrate transceptor(s) in plants. J. Exp. Bot. 477 62:2299-2308. 478 Hall, D., V. Luquez, V.M. Garcia, K.R. St Onge, S. Jansson and P. Ingvarsson. 2007. Adaptive population 479 differentiation in phenology across a latitudinal gradient in european aspen (Populus tremula, L): A 480 comparison of neutral markers, candidate genes and phenotypic traits. Evolution 61:2849-2860. 481 Hamilton, J.A. and J.M. Miller. 2015. Adaptive introgression as a resource for management and genetic 482 conservation in a changing climate. Conservation Biology 30:33-41. 483 Harrison, R.G. and E.L. Larson. 2014. Hybridization, introgression, and the nature of species boundaries. 484 J. Hered. 105:795-809. 485 Hedrick, P.W. 2013. Adaptive introgression in animals: Examples and comparison to new mutation and 486 standing variation as sources of adaptive variation. Mol. Ecol. 22:4606-4618. 487 Pachauri R.K. and L.A. Meyer. 2014. IPCC Climate change 2014: Synthesis report. contribution of working 488 groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change. 489 IPCC, Geneva. 490  23  Ismail, M., R. Soolanayakanahally, P. Ingvarsson, R. Guy, S. Jansson, S. Silim and Y. El-Kassaby. 2012. 491 Comparative nucleotide diversity across north american and european Populus species. J. Mol. 492 Evol. 74:257-272. 493 Johnson, W.E., D.P. Onorato, M.E. Roelke, E.D. Land, M. Cunningham, R.C. Belden, R. McBride, D. Jansen, 494 M. Lotz, D. Shindle, J. Howard, D.E. Wildt, L.M. Penfold, J.A. Hostetler, M.K. Oli and S.J. O'Brien. 495 2010. Genetic restoration of the florida panther. Science 329:1641. 496 Jordan, I.K. 2016. The columbian exchange as a source of adaptive introgression in human populations. 497 Biology Direct 11:17. 498 Keller, S.R., N. Levsen, P. Ingvarsson, M.S. Olson and P. Tiffin. 2011a. Local selection across a latitudinal 499 gradient shapes nucleotide diversity in balsam poplar, Populus balsamifera L. Genetics 188:941. 500 Keller, S.R., R.Y. Soolanayakanahally, R.D. Guy, S.N. Silim, M.S. Olson and P. Tiffin. 2011b. Climate-driven 501 local adaptation of ecophysiology and phenology in balsam poplar, Populus balsamifera L. 502 (Salicaceae). Am. J. Bot. 98:99-108. 503 La Mantia, J., J. Klápště, Y. El-Kassaby, S. Azam, R. Guy, C. Douglas, S. Mansfield and R. Hamelin. 2013. 504 Association analysis identifies Melampsora × columbiana poplar leaf rust resistance SNPs 505 . PLoS ONE 8:e78423. 506 Larisch, C., M. Dittrich, H. Wildhagen, S. Lautner, J. Fromm, A. Polle, R. Hedrich, H. Rennenberg, T. Muller 507 and P. Ache. 2012. Poplar wood rays are involved in seasonal remodeling of tree physiology. Plant 508 Physiol. 160:1515-1529. 509  24  Levsen, N.D., P. Tiffin and M.S. Olson. 2012. Pleistocene speciation in the genus Populus (Salicaceae). 510 Syst. Biol. 61:401-412. 511 Little, E.L. 1971. Conifers and important hardwoods. p. 1146-1155. In U.S. Department of Agriculture 512 Miscellaneous Publication (ed.) Atlas of United States trees Volume 1. U.S. Department of 513 Agriculture Miscellaneous Publication, Washington D. C. 514 McKown, A.D., R.D. Guy and L.K. Quamme. 2016. Impacts of bud set and lammas phenology on root: 515 Shoot biomass partitioning and carbon gain physiology in poplar. Trees 30:2131-2141. 516 McKown, A.D., R.D. Guy, J. Klápště, A. Geraldes, M. Friedmann, Q.C.B. Cronk, Y. El-Kassaby, S.D. 517 Mansfield and C.J. Douglas. 2013. Geographical and environmental gradients shape phenotypic 518 trait variation and genetic structure in Populus trichocarpa. New Phytol. 201:1263-1276. 519 McKown, A.D., R.D. Guy, L. Quamme, J. Klápště, J. La Mantia, C.P. Constabel, Y. El-Kassaby, R.C. Hamelin, 520 M. Zifkin and M.S. Azam. 2014a. Association genetics, geography and ecophysiology link stomatal 521 patterning in Populus trichocarpa with carbon gain and disease resistance trade-offs. Mol. Ecol. 522 23:5771-5790. 523 McKown, A.D., J. Klápště, R.D. Guy, A. Geraldes, I. Porth, J. Hannemann, M. Friedmann, W. Muchero, 524 G.A. Tuskan, J. Ehlting, Q.C.B. Cronk, Y. El-Kassaby, S.D. Mansfield and C.J. Douglas. 2014b. 525 Genome-wide association implicates numerous genes underlying ecological trait variation in 526 natural populations of Populus trichocarpa. New Phytol. 203:535-553. 527 Norris, L.C., B.J. Main, Y. Lee, T.C. Collier, A. Fofana, A.J. Cornel and G.C. Lanzaro. 2015. Adaptive 528 introgression in an african malaria mosquito coincident with the increased usage of insecticide-529 treated bed nets. Proceedings of the National Academy of Sciences 112:815-820. 530  25  Olson, M.S., N. Levsen, R.Y. Soolanayakanahally, R.D. Guy, W.R. Schroeder, S.R. Keller and P. Tiffin. 2013. 531 The adaptive potential of Populus balsamifera L. to phenology requirements in a warmer global 532 climate. Mol. Ecol. 22:1214-1230. 533 Orr, H.A. and R.L. Unckless. 2014. The population genetics of evolutionary rescue. PLOS Genetics 534 10:e1004551. 535 Petit, R. and A. Hampe. 2006. Some evolutionary consequences of being a tree. Annu. Rev. Ecol. Evol. 536 Syst. 37:187-214. 537 Plummer, M., N. Best, K. Cowles and K. Vines. 2010. Coda, version 0.14-2. available: http://Cran.r-538 project.org.  539 Porth, I., J. Klápště, A.D. McKown, J. La Mantia, R.D. Guy, P. Ingvarsson, R. Hamelin, S.D. Mansfield, J. 540 Ehlting, C.J. Douglas and Y. El-Kassaby. 2015. Evolutionary quantitative genomics of Populus 541 trichocarpa. Plos One 10:e0142864. 542 Purcell, S., B. Neale, K. Todd-Brown, L. Thomas, M. Ferreira, D. Bender, J. Maller, P. Sklar, P. de Bakker, 543 M.J. Daly and P.C. Sham. 2007. PLINK: A toolset for whole-genome association and population-544 based linkage analysis. American Journal of Human Genetics 81:559-575. 545 Richardson, J., J.G. Isebrands and J.B. Ball. 2014. Ecology and physiology of poplars and willows. p. 92-546 123. In J.G. Isebrands and J. Richardson (eds.) Poplars and willows: Trees for society and the 547 environment. CABI, Oxfordshire, UK. 548 Rieseberg, L.H. and J.F. Wendel. 1993. Introgression and its consequences in plants. p. 70-109. In R.G. 549 Harrison (ed.)  In hybrid zones and the evolutionary process. Oxford Univ. Press, New York. 550  26  Rohde, A., V. Storme, V. Jorge, M. Gaudet, N. Vitacolonna, F. Fabbrini, T. Ruttink, G. Zaina, N. Marron, S. 551 Dillen, M. Steenackers, M. Sabatti, M. Morgante, W. Boerjan and C. Bastien. 2011. Bud set in poplar 552 genetic dissection of a complex trait in natural and hybrid populations. New Phytologist 189:106-553 121. 554 Roshyara, N.R. and M. Scholz. 2014. fcGENE: A versatile tool for processing and transforming SNP 555 datasets. PLoS ONE 9:e97589. 556 Savolainen, O. and T. Pyhäjärvi. 2007. Genomic diversity in forest trees. Curr. Opin. Plant Biol. 10:162-557 167. 558 Scheet, P. and M. Stephens. 2006. A fast and flexible statistical model for large-scale population 559 genotype data: Applications to inferring missing genotypes and haplotypic phase. Am. J. Hum. 560 Genet. 78:629-644. 561 Shriner, D., A. Adeyemo and C.N. Rotimi. 2011. Joint ancestry and association testing in admixed 562 individuals. PLOS Computational Biology 7:e1002325. 563 Sjödin, A., N.R. Street, G. Sandberg, P. Gustafsson and S. Jansson. 2009. PopGenIE: The Populus genome 564 integrative explorer. A new tool for exploring the populus genome. New Phytologist 182:1013-565 1025. 566 Soolanayakanahally, R.Y., R.D. Guy, S.N. Silim and M. Song. 2013. Timing of photoperiodic competency 567 causes phenological mismatch in balsam poplar (Populus balsamifera L.). Plant, Cell & Environment 568 36:116-127. 569  27  Soolanayakanahally, R.Y., R.D. Guy, S.N. Silim, E.C. Drewes and W.R. Schroeder. 2009. Enhanced 570 assimilation rate and water use efficiency with latitude through increased photosynthetic capacity 571 and internal conductance in balsam poplar (Populus balsamifera L.). Plant, Cell Environ. 32:1821-572 1832. 573 Sturrock, R.N., S.J. Frankel, A.V. Brown, P.E. Hennon, J.T. Kliejunas, K.J. Lewis, J.J. Worrall and A.J. 574 Woods. 2011. Climate change and forest diseases. Plant Pathol. 60:133-149. 575 Suarez-Gonzalez, A., C.A. Hefer, C. Christe, O. Corea, C. Lexer, Q.C.B. Cronk and C.J. Douglas. 2016. 576 Genomic and functional approaches reveal a case of adaptive introgression from Populus 577 balsamifera (balsam poplar) in P. trichocarpa (black cottonwood). Mol. Ecol. 25:2427-2442. 578 Tamura, K., J. Dudley, M. Nei and S. Kumar. 2007. MEGA4: Molecular evolutionary genetics analysis 579 (MEGA) software version 4.0. Mol. Biol. Evol. 24:1596-1599. 580 Vitasse, Y., A. Lenz and C. Korner. 2014. The interaction between freezing tolerance and phenology in 581 temperate deciduous trees. Frontiers in Plant Science 5:541. 582 Wang, T., A. Hamann, D.L. Spittlehouse and T.Q. Murdock. 2012. ClimateWNA – high-resolution spatial 583 climate data for western North America. J. Appl. Meteor. Climatol. 51:16-29. 584 Wegmann, D., D.E. Kessner, K.R. Veeramah, R.A. Mathias, D.L. Nicolae, L.R. Yanek, Y.V. Sun, D.G. 585 Torgerson, N. Rafaels, T. Mosley, L.C. Becker, I. Ruczinski, T.H. Beaty, S.L.R. Kardia, D.A. Meyers, K.C. 586 Barnes, D.M. Becker, N.B. Freimer and J. Novembre. 2011. Recombination rates in admixed 587 individuals identified by ancestry-based inference. Nat. Genet. 43:847-853. 588  28  Whitney, K., R.A. . Randell and L. Rieseberg. 2006. Adaptive introgression of herbivore resistance traits in 589 the weedy sunflower Helianthus annuus. Am. Nat. 167:794-807. 590 Whitney, K.D., K.W. Broman, N.C. Kane, S.M. Hovick, R.A. Randell and L.H. Rieseberg. 2015. Quantitative 591 trait locus mapping identifies candidate alleles involved in adaptive introgression and range 592 expansion in a wild sunflower. Mol. Ecol. 24:2194-2211. 593 Xie, C., C.C. Ying, A.D. Yanchuk and D.L. Holowachuk. 2009. Ecotypic mode of regional differentiation 594 caused by restricted gene migration: A case in black cottonwood (Populus trichocarpa) along the 595 pacific northwest coast. Can. J. for. Res. 39:519-525. 596   597 FIGURE LEGENDS 598 Figure 1. Geographic distribution of admixed P. trichocarpa individuals (circles) used in admixture 599 mapping. Additional phenotypic analyses also included pure P. trichocarpa individuals (red 600 triangles). Pure P. balsamifera individuals (blue triangles) were included in the phenotypic 601 analysis of one trait (chlorophyll content index). Ranges of P. trichocarpa and P. balsamifera are 602 shown in red and blue, respectively (Little, 1971). A square represents the location where P. 603 trichocarpa homozygotes for the P. balsamifera haplotype in the introgressed region in 604 chromosome 9 occur (Prince George). Diamonds represent the locations where heterozygous 605 individuals for the introgressed haplotype in chromosome 9 occur.  606  607 Figure 2 Relationship between trait variation and geoclimate variables in admixed and pure P. 608 trichocarpa. A: Correlation between trait variation and maximum length of day (DAY). B: 609 Correlation between trait variation and traits associated with temperature (first principal 610 component based on climate variables associated with temperature explaining 60% of the 611 variance). Asterisks represent significant interactions between the slopes of admixed and pure P. 612 trichocarpa (p < 0.05). Detailed information about the traits can be found in Tables 3; Supporting 613 Information Table S1. 614  615 Figure 3 Trait variation in admixed and pure P. trichocarpa across 16 traits associated with an 616 introgressed haplotype on chromosome 9. bb (blue): homozygotes for the P. balsamifera 617 haplotype, tb (purple): heterozygotes, tt (red): homozygotes for the P. trichocarpa haplotype. 618 Pure P. trichocarpa, all homozygotes for the P. trichocarpa haplotype, from northern (pink) and 619 southern (dark red) populations had similar phenotypes (p>0.05). Asterisks represent traits where 620 bb was significantly different than tt. All 16 traits showed significant differences between tt and 621 pure individuals, but there were no significant differences between bb and pure individuals, 622  29  except in Volumecm3_2011. Additional information about the ANOVAs can be found in 623 Supporting Information Table S3. Detailed information about the traits can be found in Tables 3 624 and Supporting Information Table S1. The box plot displays the distribution of data where the 625 central rectangle spans the first quartile to the third quartile (IQR), the segment inside the 626 rectangle shows the median and "whiskers" above and below the box show the locations of the 627 minimum and maximum. Outliers are 3×IQR or more above the third quartile or 3×IQR or more 628 below the first quartile. 629  630 Figure 4 Variation in the chlorophyll content index in admixed and pure P. trichocarpa (Pure Pt, 631 dark red) and P. balsamifera (Pure Pb, dark blue) individuals. The chlorophyll content index was 632 associated with an introgressed sutelomeric region on chromosome 15. The haplotype in the 633 introgressed subtelomeric region on chromosome 15 is shown inside parentheses [tt (red): 634 homozygotes for the P. trichocarpa haplotype, bt (purple): heterozygotes, bb(blue): homozygotes 635 for the P. balsamifera haplotype]. Data from pure P. balsamifera is only available for estimates of 636 chlorophyll content index from 2015. Shared letters above boxes indicate that the average 637 chlorophyll content indices were not significantly different between those genotypes (p<0.05). 638 See the legend of Figure 3 for details on the box plot.  639  640 TABLES 641 Table 1. Summary of the results from admixture mapping in Populus trichocarpa. The number of 642 traits, SNPs and chromosomes are shown for all associations with posterior probabilities higher 643 than 0.5 (all >0.5). Results are also displayed for association in introgressed regions [based on a 644 whole genome local ancestry study] with posterior probabilities higher than 0.5 (introgressed 645 >0.5) and 0.9 (introgressed >0.9). 646   Posterior probability category Trait category Dataset all >0.5 introgressed >0.5 introgressed >0.9 Biomass 26 22 14 11 Disease 2 2 1 0 Ecophysiology 60 42 3 2 Phenology 20 19 13 7 Total 108 85 31 20 Number of chromosomes 19 19 7 3 SNPs 1,168,955  86,150 2,874 1,275  Associations 223,161  18,525  3,429   647 Table 2. List of the Populus balsamifera introgressed regions found in a whole genome local 648 ancestry study and ancestry-trait associations. ‘Max. balsa %’ refers to the height of the 649 introgressed peaks in terms of the maximum percentage of P. balsamifera ancestry within the 650  30  peak. ‘SNP (fixed-association)’ are SNPs with fixed differences in the pure species and associated 651 with traits based on BMIX. ‘Association (BMIX)’ is the number of traits associated with SNPs 652 based on BMIX, and ‘Association (ANOVA)’ are the traits associated with haplotypes based on 653 ANOVAs.  654 Introgressed regions start end Size (kb) Max balsa % SNP (fixed- association) Association (BMIX) Association (ANOVA) ch01a 47,120,000 47,540,000 420 20.92% 0 0  ch01b 48,000,000 48,380,000 380 19.43% 0 0  ch03 21,040,000 21,620,000 580 20.85% 0 0  ch05a 140,000 380,000 240 19.97% 0 1  ch05b 11,580,000 11,820,000 240 18.52% 0 0  ch06 3,540,000 3,700,000 160 17.78% 0 4  ch07 14,680,000 15,700,000 1,020 27.33% 0 0  ch09 1,380,000 1,760,000 380 19.52% 119 19 16 ch10 22,180,000 22,660,000 480 19.93% 18 21 2 ch11a 1,940,000 2,120,000 180 18.36% 0 0  ch11b 17,580,000 17,780,000 200 20.10% 0 0  ch11c 17,800,000 18,600,000 800 21.81% 0 0  ch14a 12,720,000 13,300,000 580 22.67% 6 15 5 ch14b 13,340,000 13,480,000 140 19.25% 0 0  ch15a 0 580,000 580 20.29% 0 1 3* ch15b 13,480,000 13,920,000 440 19.31% 0 0  ch17a 3,080,000 3,900,000 820 21.50% 0 1  ch17b 9,720,000 10,420,000 700 26.00% 0 0  ch17c 12,720,000 13,380,000 660 22.93% 0 0  *Haplotypes on chromosome 15 were based on SNPs that had fixed differences in the pure species. On chromosome 9, 10 and 14, haplotypes were based on SNPs that had both fixed differences and associations with traits in the BMIX analysis. ANOVAs were conducted for two additional traits that did not show significant associations in BMIX (summer chlorophyll content index in 2009 and 2011) but were significantly associated with SNPs on chromosome 15 in the targeted study on adaptive introgression (Suarez-Gonzalez et al., 2016). SNP: Single Nucleotide Polymorphism BMIX: Bayesian method to explore admixture mapping ANOVA: Analysis of Variance   655  656  31   657 Table 3 List of traits that showed significant associations with introgressed haplotypes in both 658 BMIX and ANOVA analyses in Populus trichocarpa. Traits with posterior probabilities higher than 659 0.9 in introgressed regions [based on a whole genome local ancestry study] are also shown. 660 Detailed information about the traits can be found in Supporting Information Table S1.  661 Trait Chr Trait category Introgressed > 0.9 (BMIX) Activegrowthratecmday_2010 ch09 biomass  Heightcm_2010 ch09 biomass Yes Heightcm_2011 ch09 biomass Yes Heightgaincm_2009 ch09 biomass  Heightgaincm_2010 ch09 biomass Yes Volumecm3_2011 ch09 biomass  AUDPC_2010 ch09 disease  Budsetday_2009 ch09 phenology  Budsetday_2010 ch09 phenology Yes Growthperioddays_2010 ch09 phenology Yes Leafdropday_2009 ch09 phenology Yes Leaflifespandays_2010 ch09 phenology  Yellowing100_2010 ch09 phenology Yes Yellowing25_2010 ch09 phenology  Yellowing50_2010 ch09 phenology  Yellowing75_2010 ch09 phenology Yes Heightcm_2010 ch10 biomass Yes Heightcm_2011 ch10 biomass Yes Activegrowthratecmday_2010 ch14 biomass  Heightcm_2010 ch14 biomass Yes Heightcm_2011 ch14 biomass Yes Heightgaincm_2010 ch14 biomass Yes Yellowing100_2010 ch14 phenology  CCI_ju10 ch15 ecophysiology  Volumecm3: Bole volume calculated assuming a cone (McKown et al. 2013) AUDPC: Area under the disease progress curve (La Mantia et al. 2013) CCI_ju10: Chlorophyll concentration index June 10, 2015 BMIX: Bayesian method to explore admixture mapping ANOVA: Analysis of Variance  662   1  Heading: Introgression underlies adaptively significant variation and range boundaries in forest 1 trees 2  3 Twitter for ASG: @AdriSuarezGonz 4  5 Introgression from Populus balsamifera underlies adaptively significant variation and range 6 boundaries in P. trichocarpa 7   8 Adriana Suarez-Gonzalez1, Charles A. Hefer1,2, Christian Lexer3, Carl J. Douglas†, and Quentin C. B. 9 Cronk1* 10 1Department of Botany, University of British Columbia, Vancouver, V6T 1Z4, Canada 11 2Biotechnology Platform, Agricultural Research Council, Private Bag X05, Onderstepoort, 0110, 12 South Africa 13 3 Department of Botany and Biodiversity Research, University of Vienna, 1030, Austria 14 † Deceased 25 July 2016 15 *Corresponding author: 16  17 Quentin Cronk 18 Email: quentin.cronk@ubc.ca 19 Total word count (excluding summary, references and legends):  4829 No. of figures:  4 (Figs 1–4 in colour)  Summary:  200 No. of Tables:  3 Introduction:  656 No of Supporting Information files:  7 (Fig. S1– S3; Table S1– S3, script)  Materials and Methods:  1420 Results:  1052 Discussion:  1587 Acknowledgements:  69  20 Keywords: adaptive introgression, admixture mapping, epistasis, genome architecture, latitudinal 21 cline, phenomics, Salicaceae, species range 22  23  2   24 INTROGRESSION FROM POPULUS BALSAMIFERA UNDERLIES ADAPTIVELY SIGNIFICANT 25 VARIATION AND RANGE BOUNDARIES IN P. TRICHOCARPA 26 SUMMARY 27 • Introgression can be an important source of adaptive phenotypes, although conversely it can 28 have deleterious effects. Evidence for adaptive introgression is accumulating but information on 29 the genetic architecture of introgressed traits lags behind.  30 • Here we determine trait architecture in Populus trichocarpa under introgression from P. 31 balsamifera using admixture mapping and phenotypic analyses.  32 • Our results reveal that admixture is a key driver of clinal adaptation and suggest that the 33 northern range extension of P. trichocarpa depends, at least in part, on introgression from P. 34 balsamifera. However, admixture with P. balsamifera can lead to potentially maladaptive early 35 phenology and a reduction in growth and disease resistance in P. trichocarpa. Strikingly, an 36 introgressed chromosome 9 haplotype block from P. balsamifera restores the late phenology 37 and high growth parental phenotype in admixed P. trichocarpa. This epistatic restorer block 38 may be strongly advantageous in maximizing carbon assimilation and disease resistance in the 39 southernmost populations where admixture has been detected. We also confirm a previously 40 demonstrated case of adaptive introgression in chromosome 15 and show that introgression 41 generates a transgressive chlorophyll–content phenotype.  42 • We provide strong support that introgression provides a reservoir of genetic variation 43 associated with adaptive characters that allows improved survival in new environments. 44  3  INTRODUCTION 45 Gene flow between species through hybridization can be a potent evolutionary force when 46 recombination increases standing variation and creates opportunities for adaptive evolution 47 (Harrison and Larson, 2014; Hedrick, 2013). Admixture can be seen as a collision of genomes, that 48 nevertheless may be followed by the stable integration of genomic regions of one parental species 49 into the genome of another (Buerkle and Lexer, 2008). This process, called introgression, occurs 50 through hybridization and repeated backcrossing (Rieseberg and Wendel, 1993). Although 51 introgression has the potential to detrimentally disrupt the recipient genomic background, it can 52 also provide beneficial variants that result in accelerated adaptation and improved survival in new 53 environments (Clarkson et al., 2014; Norris et al., 2015; Whitney et al., 2006; Whitney et al., 2015). 54 Simulations and empirical evidence show that gene flow between species results in increased 55 standing variation available for adaptive evolution (Barrett and Schluter, 2008; Jordan, 2016), which 56 may play a key role in species’ ability to respond to a changing climate (Hamilton and Miller, 2015). 57 Standing variation is often more likely to result in adaptation and evolutionary consequences under 58 rapidly changing conditions than de novo mutations (Orr and Unckless, 2014) and this is particularly 59 important for long-lived organisms such as trees (Petit and Hampe, 2006; Savolainen and Pyhäjärvi, 60 2007).  61 Populus trichocarpa (black cottonwood) is an ecologically and economically important forest tree 62 species distributed throughout the western US and Canada, from northern California to southern 63 Alaska, and adapted to relatively humid, moist, and mild conditions west of the Rocky Mountains. 64 Along this range, P. trichocarpa exhibits variation in several adaptive traits including phenology and 65 disease susceptibility, which suggests that local adaptation plays an important role shaping the 66 genetic and phenotypic distribution of this species (La Mantia et al., 2013; McKown et al., 2013). 67  4  These traits exhibit high heritability and are strongly correlated with geoclimatic variables such as 68 latitude, day length, and temperature (La Mantia et al., 2013; McKown et al., 2013).  69 In the interior and northern parts of its range, P. trichocarpa hybridizes freely with P. balsamifera 70 where the distributions of the two species overlap (Geraldes et al., 2014; Suarez-Gonzalez et al., 71 2016). Populus balsamifera is a boreal species distributed from Alaska to Newfoundland, with high 72 frost tolerance and able to tolerate a very large range in extreme temperatures (-62˚C to 44˚C) as 73 well as to moderate annual precipitation (Richardson et al. 2014). These two sibling poplar species 74 diverged in allopatry and despite their morphological similarity and recent divergence [~76 Ka 75 (Levsen et al., 2012); but see (Ismail et al., 2012)], they are ecologically divergent and adapted to 76 strongly contrasting environments (Geraldes et al., 2014; Richardson et al., 2014). Populus 77 balsamifera, like P. trichocarpa, has been an important subject for the study of variation and 78 adaptation in trees (Keller et al., 2011a; Keller et al., 2011b; Olson et al., 2013; Soolanayakanahally 79 et al., 2013). Introgression from the northern and continental species P. balsamifera could transfer 80 advantageous traits allowing admixed P. trichocarpa individuals to colonize colder environments in 81 northern and interior locations. A targeted analysis demonstrated adaptive introgression from P. 82 balsamifera into P. trichocarpa in a subtelomeric region of chromosome 15 (Suarez-Gonzalez et al., 83 2016), and a whole genome analysis identified additional candidate regions for adaptive 84 introgression (A Suarez-Gonzalez et al., unpublished). However, the targeted study did not include 85 phenotypic information from pure P. balsamifera and it is unknown if other introgressed regions, 86 identified in the whole genome analysis, are associated with phenotype in admixed P. trichocarpa.   87 Here we implemented admixture mapping to explore if introgressed regions with unusually high 88 levels of P. balsamifera ancestry are driving variation in locally adaptive traits. We used phenotypic 89 analyses to detect if trait variation along a latitudinal gradient was associated with variation in 90  5  geoclimatic variables in admixed and pure P. trichocarpa individuals. We also explore the specific 91 effects of one introgressed region on trait variation in admixed individuals. Finally, we revisit a case 92 of adaptive introgression in a subtelomeric region of chromosome 15, found in the targeted analysis 93 (Suarez-Gonzalez et al., 2016), and use phenotypic data from pure individuals to investigate the 94 effects of this genomic region in a P. trichocarpa and P. balsamifera background.  95 METHODS 96 Samples 97 We employed data from a whole genome local ancestry analysis and from association studies in P. 98 trichocarpa Torr. & Gray (La Mantia et al., 2013; McKown et al., 2013; McKown et al., 2014a) as well 99 as phenotypic data generated specifically for this study (i.e. chlorophyll content index). The 100 accessions used were from a collection of the British Columbia Ministry of Forests, Lands and 101 Natural Resource Operations, outplanted in a common garden at the University of British Columbia 102 (Xie et al., 2009). The local ancestry analysis and chlorophyll measurements also included P. 103 balsamifera L. accessions from the Agriculture and Agri-Food Canada AgCanBaP collection 104 (Soolanayakanahally et al., 2009; Suarez-Gonzalez et al., 2016). For admixture mapping, we used 105 118 admixed individuals that showed introgression from P. balsamifera ancestry in a whole genome 106 local ancestry analysis. We focused on accessions from northern and interior parts of the P. 107 trichocarpa range to target locations where admixture has been detected (Geraldes et al., 2014; 108 Suarez-Gonzalez et al., 2016). In addition, we used reference P. trichocarpa and P. balsamifera 109 individuals to detect SNPs (Single Nucleotide Polymorphism) with fixed differences. Pure P. 110 trichocarpa individuals were used in the phenotypic analyses while data from pure P. balsamifera 111 individuals was available for only one trait (chlorophyll content index) (Figure 1, Supporting 112 Information Table S1). 113  6  Ancestries and phenotypes 114 In the local ancestry analyses, SNPs were called in the whole genome of 168 individuals (25 115 reference individuals of each parental species and 118 admixed P. trichocarpa individuals with P. 116 balsamifera admixture) selected from the sympatric zone between P. trichocarpa and P. 117 balsamifera as well as from allopatric populations (Supporting Information Figure S1, (Suarez-118 Gonzalez et al., 2016)). Each of the genotypes was sequenced at an expected coverage ranging from 119 15X to 30X using the Illumina HiSeq2000 platform (SRA: PRJNA276056). Each SNP was annotated 120 using SNPeff (Cingolani et al., 2012) with version 3 of the P. trichocarpa genome assembly.  121 We estimated the probabilities for each of the possible ancestral configurations (P. balsamifera, P. 122 trichocarpa or mixed ancestry) in every SNP across the whole genome of 118 admixed individuals 123 using RASPberry, a software that implements a reliable Hidden Markov model (HMM) for admixture 124 (Wegmann et al., 2011), following the same pipeline and parameters as in our previous study on 125 adaptive introgression (Suarez-Gonzalez et al., 2016). Briefly, SNPs with missing data in the parental 126 genotypes were removed using Plink 1.07 (Purcell et al., 2007), the parental genotypes (50) were 127 then phased with fastphase (Scheet and Stephens, 2006) by creating the input files with FCGENE 128 (Roshyara and Scholz, 2014), and ancestries for each admixed individual were estimated in 129 ADMIXTURE (Alexander et al., 2009).  130 To determine the ancestral configurations of each SNP in one of the three categories, ancestries 131 were considered for probabilities >95%. The proportion of introgressed ancestry in admixed 132 individuals was calculated, for each SNP, by counting sites with introgressed ancestry as two and 133 sites with mixed ancestry as one. We detected regions with unusually high levels of introgression in 134 admixed individuals using a sliding window approach and a significance cut-off of three standard 135  7  deviations (SD) from the weighted mean across all the chromosomes based on SNP density per 136 window (100-kb, steps of 20-kb) (Suarez-Gonzalez et al., 2016).  137 For phenotypic data, we used 59 traits encompassing phenological events, biomass accumulation, 138 growth rates, disease susceptibility, as well as leaf, isotope and gas exchange-based ecophysiology 139 from previous studies (La Mantia et al., 2013; McKown et al., 2013; McKown et al., 2014a). This 140 dataset was collected throughout 2008 to 2012 from 461 P. trichocarpa accessions with 4 to 20 141 clonal replicates similar in age and condition, grown as stecklings under glasshouse conditions, and 142 then out-planted in a common garden at Totem Field, University of British Columbia, Canada. In 143 2015, we measured chlorophyll absorbance on leaves for the chlorophyll content index (CCI) using a 144 CCM-200 plus SPAD chlorophyll meter (Opti-Sciences Inc., Hudson, NH, USA) in P. trichocarpa and P. 145 balsamifera accessions in Totem Field.  146 Admixture mapping 147 To explore the effects of introgression on the genetic architecture of adaptation in P. trichocarpa, 148 we used a Bayesian method called BMIX (Shriner et al., 2011). BMIX empirically estimates the 149 testing burdens of admixture mapping by fitting an autoregressive model and estimating the 150 effective number of tests based on autocorrelation. Since a block of ancestry from one parental 151 population can be up to several megabases long, local ancestry estimates can be highly correlated 152 in a genome. First, we estimated the number of effectively independent tests for each chromosome 153 for each individual by fitting an autoregressive model to the local ancestries (Plummer et al., 2010). 154 The spectral density was estimated at frequency zero and the order of the fitted autoregressive 155 models was chosen by minimizing the Akaike information criterion. The effective number of tests 156 were then summed for the chromosomes of each individual and averaged across individuals. Next, 157 the phenotypes were regressed on local ancestry, adjusting for global ancestry using generalized 158  8  linear models (GLMs). Global ancestry was calculated for each individual as the local ancestry 159 averaged across all markers. Finally, the p-values from the regression models were converted into 160 posterior probabilities. The threshold was the value at which the hypothesis favored by the 161 posterior probability switches (i.e. 0.5) (Supporting Information Notes S1, (Shriner et al., 2011)).  An 162 R script was used to run the GLMs and convert the p-values to posterior probabilities (Supporting 163 Information, Shriner et al., 2011).    164 We focused on 19 regions with unusually high levels of P. balsamifera introgression identified by 165 the whole genome local ancestry analysis (Supporting Information Figure S2). Overall, 107 166 individuals with both local ancestry and phenotypic data were used in the admixture mapping 167 analysis. We focussed all downstream phenotypic analyses on ancestry-trait associations showing 168 very strong signals (i.e. posterior probabilities higher than 0.9), on SNPs fixed for different alleles in 169 the parentals, and on a region that showed signals of adaptive introgression in the genomic and 170 functional study (Suarez-Gonzalez et al., 2016). 171 Phenotypic analysis of associations from admixture mapping  172 To explore the effects of P. balsamifera alleles on the phenotype of admixed P. trichocarpa, we 173 selected SNPs that had both fixed differences in the pure species and displayed associations with 174 traits in the BMIX analysis. We also included all the fixed SNPs from the subtelomeric introgressed 175 region on chromosome 15, a candidate region for adaptive introgression associated with 176 chlorophyll levels (Suarez-Gonzalez et al., 2016). Then, haplotypes in the admixed individuals were 177 inferred with fastphase (Scheet and Stephens, 2006) by creating the input files of fixed SNPs with 178 FCGENE (Roshyara and Scholz, 2014). To identify P. balsamifera and P. trichocarpa haplotypes, we 179 performed neighbor-joining (NJ) analyses [1000 bootstrap replicates in MEGA (Tamura et al., 2007)] 180 including admixed and pure individuals (Supporting Information Figure S3). Each of the admixed P. 181  9  trichocarpa individuals was classified into one of three genotypic categories: homozygotes for P. 182 balsamifera haplotypes (bb), heterozygotes (bt) and homozygotes for P. trichocarpa haplotypes (tt), 183 based on phased genomic sequences. Finally, phenotypic traits showing associations with the 184 introgressed regions were compared among the three genotypic categories (bb, bt, and tt) in each 185 of the haplotypes using analysis of variance (ANOVA) with adjusted p-values with Bonferroni 186 correction. For the analysis of variance on the chlorophyll content index of 2009, we included 187 individuals that set bud after the summer solstice that year (prior to day 186 were removed), since 188 following bud set, trees have a greater amount of chlorophyll on average compared to trees still 189 within an active growing phase (McKown et al., 2016). For 2011 and 2015 data on bud set was not 190 recorded.  191 Climate of tree origin in admixed and pure individuals 192 To determine if trait variation was correlated with climate and admixture, we compiled eight 193 climate variables associated with moisture and 15 variables related to temperature from ClimateNA 194 (Wang et al., 2012) based on 1971–2000 (Supporting Information Table S2). We performed two 195 principal component analyses (PCA), one for moisture and one for temperature, using the function 196 prcomp in R. In addition, the maximum length of day (DAY; h) was also calculated at each location 197 as a proxy for photoperiodic regime using the package geosphere in R v. 3.3.2 (R Development Core 198 Team, http://www.r-project.org). DAY was not used in the PCA analysis.  199 Enrichment analysis  200 To detect overrepresented biological terms in an introgressed haplotype associated with trait 201 variation, we performed enrichment tests for various terms including Gene Ontology (GO) and 202 Protein Family (Pfam) using Popgenie (Sjödin et al., 2009). The list of introgressed genes was 203 compared with the list of all poplar genes (41,335) and the best-annotated orthologs in Arabidopsis 204  10  thaliana were identified based on Popgenie annotation using Fisher´s exact test and the default p-205 value threshold (0.05). 206 RESULTS 207 Introgressed regions from P. balsamifera are associated with trait variation in P. trichocarpa 208 Across the whole genome (1,168,955 SNPs examined) of 118 admixed P. trichocarpa individuals, we 209 detected 19 regions with unusually high levels of introgression (1107 genes) with P. balsamifera 210 ancestry (Supporting Information Figure S2). These regions included candidate genes for adaptive 211 introgression previously identified by our previous targeted study (Suarez-Gonzalez et al., 2016). 212 The 19 regions, occurring across 11 chromosomes, showed P. balsamifera ancestry peaks with a 213 height (i.e. percentage of P. balsamifera ancestry) ranging from 0.1778 to 0.2733 and width ranging 214 from 140 kb to 1.02 Mb.  215 The admixture mapping analysis revealed 86,150 SNPs associated with 57 traits (across multiple 216 years) with posterior probabilities above 0.5 (Table 1). In the 19 regions with unusually high levels 217 of P. balsamifera introgression identified in the local ancestry analysis, we detected 18,525 218 phenotype associations in 2,874 SNPs across seven chromosomes (5, 6, 9, 10, 14, 15 and 17). 219 Introgressed regions on chromosome 5, 6, 15 and 17 showed associations with only one to four 220 phenotypic traits, while regions on chromosomes 9, 10 and 14 showed associations with 15 to 20 221 different traits (Table 2). Introgressed regions on chromosomes 9, 10 and 14 also displayed the 222 strongest associations (i.e. posterior probabilities above 0.9), comprising 3,429 SNPs and 20 trait 223 associations (12 traits some measured in multiple years). In addition, the only introgressed region 224 showing associations with disease resistance was located on chromosome 9 (Table 3).  225 A total of 143 SNPs located on chromosomes 9 (119 SNPs), 10 (18 SNPs) and 14 (6 SNPs) displayed 226 both fixed differences in the pure species and associations with traits in BMIX. Neighbor-joining 227  11  analyses, based on haplotypes from these fixed differences, revealed clear clusters of P. balsamifera 228 and P. trichocarpa haplotypes and allowed us to classify admixed P. trichocarpa individuals as 229 heterozygous or homozygous for P. balsamifera and P. trichocarpa haplotypes (Supporting 230 Information Figure S1).  231 Admixture drives adaptively relevant trait variation across a latitudinal gradient  232 The traits showing the strongest associations with the introgressed regions were related to 233 phenology, biomass traits (e.g. height and bole volume), and disease resistance. These traits 234 showed strong correlations with maximum day length (DAY) and temperature (first principal 235 component based on climate variables associated with temperature explaining 60% of the 236 variance), which had significant interactions with the presence of admixture (i.e. significant 237 differences between the slopes of pure and admixed individuals, p<0.05; Figure 2). Admixed 238 individuals from colder environments had earlier phenology (i.e. bud set, leaf yellowing, and leaf 239 drop), were smaller, and had more damage by disease than those from southern regions. The same 240 types of correlations were found with maximum day length (DAY), where individuals from higher 241 latitudes (greater maximum daylength) had earlier phenology (i.e. bud set, leaf yellowing, and leaf 242 drop), and were smaller. An exception to this pattern, however, was damage by disease, which did 243 not increase with maximum day length. Curiously, in pure P. trichocarpa we did not detect a 244 relationship between phenotypic traits and DAY or temperature. Pure P. trichocarpa individuals 245 from northern regions had similar phenotypes to pure individuals from southern regions (Figures 2 246 and 3). Also, the phenotype of pure P. trichocarpa individuals was similar to that in admixed 247 individuals from lower latitudes (i.e. later phenology, greater size, less damage by disease) (Figure 248 2). The low statistical power in pure individuals, due to the small sample size at higher latitudes and 249 colder environments, could explain the lack of association between traits and geoclimatic variables. 250  12  Climate variables associated with moisture (first principal component explaining 63% of the 251 variance) did not show significant interactions with admixture. Furthermore, the whole genome 252 levels of admixture were generally not correlated with trait variation (Pearson's correlation, 253 p>0.05).  254 An introgressed haplotype on chromosome 9 restores parental phenotype in admixed P. trichocarpa   255 We focused on the introgressed ancestry block on chromosome 9 since this haplotype showed the 256 strongest and highest number of associations. ANOVAs based on a suite of phenology, biomass, and 257 disease resistance traits in admixed individuals with different haplotypes on chromosome 9 (Table 258 3) supported the results from BMIX (Tables 1 and 2) in most cases. All traits that showed significant 259 differences among haplotypes, after Bonferroni correction, revealed the same trend: admixed 260 individuals with two copies of the P. balsamifera haplotype in chromosome 9 were more similar to 261 pure P. trichocarpa, from northern and southern populations, than to other admixed individuals 262 without signals of introgression in this haplotype. Admixed individuals homozygous for the 263 introgressed haplotype in chromosome 9 had later phenology (bud set day mean: 222.5 Julian day, 264 2010), were taller (mean: 370.3 cm, 2010), and were more resistant to disease than admixed 265 individuals without P. balsamifera chromosome 9 haplotypes (mean height: 189.0 cm; bud set: 266 175.8 Julian day, 2010) (Figure 3). Phenotypes from heterozygous individuals were, for the most 267 part, midpoints of those from homozygous individuals. This result suggests that the introgressed 268 haplotype in chromosome 9 restores the parental P. trichocarpa phenotype.   269 Populus balsamifera haplotypes on chromosome 9 were geographically limited to the interior and 270 northwestern BC, with homozygotes for the P. balsamifera haplotypes occurring only in the Prince 271 George region, and heterozygotes also in Terrace and north of Juneau (Figure 1). Prince George was 272  13  the southernmost population, with the smallest value for maximum day length (“DAY”), from those 273 where admixture has been detected.  274 The introgressed haplotype on chromosome 9 was enriched for genes coding for two types of 275 protein families: PF07690, Major Facilitator Superfamily (Nitrate transporters NRT2: 276 Potri.009G008500 and Potri.009G008600) and PF00005, ABC transporter (ABC transporters: 277 Potri.009G007800, Potri.009G008200) as well as for three miRNAs [ptc-miR473 (Cleavage), ptc-278 miR6448 (Translation), ptc-miR172 (Cleavage)].  279 Introgression in the subtelomeric region of chromosome 15 confirms signals of adaptive introgression 280 We inferred haplotypes for the subtelomeric introgressed region on chromosome 15, already 281 known as a strong candidate region for adaptive introgression (Suarez-Gonzalez et al., 2016), using 282 132 fixed SNPs. Introgressed P. balsamifera haplotypes showed strong associations with the 283 chlorophyll content index across multiple years. The BMIX analysis showed associations in 284 measurements from 2015, but ANOVAs based on haplotypes revealed significant differences among 285 genotypes also in measurements from 2009 and 2011 (Figure 4). As in the targeted local ancestry 286 analysis (Suarez-Gonzalez et al., 2016), admixed individuals with P. balsamifera haplotypes in the 287 subtelomeric region of chromosome 15 exhibited higher values of the chlorophyll content index 288 compared to other admixed and pure individuals. The chlorophyll content index in pure P. 289 balsamifera individuals was similar to that in pure P. trichocarpa.  290 DISCUSSION 291 Admixture mapping connects phenotypic traits with genomic regions  292 The admixture mapping analysis in admixed P. trichocarpa individuals revealed numerous loci 293 introgressed from P. balsamifera that underlie adaptively-relevant variation in phenology, biomass, 294 disease resistance, and ecophysiology traits. These results give strong support that introgressive 295  14  hybridization in late generation backcrosses is providing a reservoir of new genetic variation 296 associated with adaptive characters in P. trichocarpa. Some of these traits are critical for survival in 297 northern environments, but it is possible that others are traits developmentally correlated with 298 these, and of less intrinsic importance. Many of the traits used here show strong correlations and 299 this has been discussed elsewhere (McKown et al., 2014). One corollary of this is when a particular 300 SNP is reported as having multiple trait associations, these may not all be independent but may be 301 the result of developmentally correlated traits. 302 In this study, seven of the 19 introgressed regions with unusually high levels of P. balsamifera 303 admixture, uncovered in the local ancestry analysis, were strongly associated with traits affecting 304 phenology, growth, response to disease, and ecophysiology in P. trichocarpa background. Two 305 introgressed regions, in particular, one on chromosome 9 and another on chromosome 15, showed 306 particularly strong and interesting associations with trait variation.  307 In discussing these analyses the issue of geographic structure should be noted (e.g. Geraldes et al., 308 2014) as this is a potential confounding factor. However, structure is unlikely to be a problem here 309 for the following reasons. Firstly, our trait mapping focused on samples from a limited latitudinal 310 range. Secondly, our mapping models were based on interspecific genetic ancestries rather than 311 raw genotype data (c.f. admixture mapping), which effectively mitigates the confounding effect of 312 within-species geographic structure. Thirdly, between species structure was robustly accounted for 313 because variation in genomic background at the between-species level was taken into account 314 during mapping. 315  316  15  Introgression extends the species range of P. trichocarpa by driving clinal adaptation of phenology traits 317 Numerous studies have shown that in trees from northern climates, bud set and growth cessation 318 are initiated earlier compared to their southern counterparts because the former are adapted to 319 the critical day length of northern regions (Evans et al., 2016; Hall et al., 2007; Rohde et al., 2011; 320 Vitasse et al., 2014). Earlier reports on trait variation in these P. trichocarpa individuals confirmed 321 these findings and revealed that most phenology events, including bud set, leaf yellowing, and leaf 322 drop were strongly associated with geoclimatic variables such as maximum day length (“DAY”) and 323 temperature (McKown et al., 2013). These previous results illustrate the adaptive importance of 324 phenology traits, and here we identified admixture as a key driver of this clinal adaptation.  325 Populus trichocarpa occurs from California to Alaska, and the northern range extension of P. 326 trichocarpa appears to be dependent, at least in a substantial part, on introgression of alleles from 327 P. balsamifera, which is a boreal species distributed from Alaska to Newfoundland. The whole 328 genome local ancestry analysis demonstrated a strong correlation between admixture and climate, 329 showing increased levels of P. balsamifera introgression in colder and drier environments (A Suarez-330 Gonzalez et al., unpublished). The phenotypic analyses support this finding and show that 331 admixture, as well as introgression of certain genomic regions (e.g. haplotypes on chromosome 9) 332 has strong pleiotropic effects on phenotypes that appear to play important roles in adaptation in 333 northern populations of P. trichocarpa.  334 An introgressed haplotype on chromosome 9 restores parental phenotype at certain latitudes 335 Although admixture with P. balsamifera clearly confers some adaptive traits on P. trichocarpa (i.e. 336 early fall phenology in the north), there is a trade-off since admixture has a tendency to reduce the 337 growth of the fast-growing P. trichocarpa and reduce disease resistance. Curiously, a particular P. 338 balsamifera haplotype block on chromosome 9, when introgressed, has the ability to restore the 339  16  parental P. trichocarpa phenotype (i.e. later phenology, higher growth and greater disease 340 resistance). This is not a latitudinal effect, where the haplotype block comes from trees at a 341 different latitude to the recipient trees, so mitigating traits. Rather, this is evident regardless of the 342 latitude of the haplotype donor or recipient. The introgressed “restorer haplotype block” on 343 chromosome 9 affected the most traits, with the highest number of associations with biomass (4), 344 phenology (5) and damage by disease (1) traits (in multiple years) and with the strongest 345 association signals (posterior probabilities > 0.9). Admixed individuals homozygous for introgressed 346 P. balsamifera loci in chromosome 9 were more similar to pure P. trichocarpa individuals (from 347 northern or southern regions) than to other admixed individuals heterozygous or homozygous for 348 the P. trichocarpa haplotype in chromosome 9. Homozygotes for the P. balsamifera haplotype in 349 chromosome 9 were found only in the southernmost location where admixture has been detected, 350 and had later phenology, were bigger and had less damage by disease compared to admixed 351 individuals homozygous for this P. trichocarpa haplotype.  352 Our results suggest that admixture contributes to adaptation at high latitudes, where cold is 353 combined with northern light regimes, but could be maladaptive in continental regions at lower 354 latitudes where cold is combined with southern light regimes. In Prince George, the population with 355 the shortest DAY from those where admixture has been detected, introgression of the haplotype on 356 chromosome 9 restores the parental phenotype. Therefore, an allele from a species with expected 357 early dormancy is conferring delayed dormancy when in the alternative genetic background. Later 358 phenology (i.e. delayed dormancy) in trees with these introgressed haplotypes could be an adaptive 359 strategy as it maximizes carbon assimilation with late dormant bud formation and delayed leaf 360 senescence, provided that it does not increase the risk of damage caused by early frosts on 361 vegetative organs (Chuine, 2010).  362  17  The physiological mechanisms underlying phenotype associations of this restorer block are of 363 potentially great interest. It is possible that certain introgressed alleles, via epistatic gene 364 interactions, allow admixed trees to respond to changes in the environment by tracking a 365 ‘phenological optimum’ more efficiently than other admixed individuals. This epistasis could include 366 interactions between P. trichocarpa alleles in chromosome 09 with introgressed P. balsamifera 367 alleles in other part(s) of the genome or interactions between introgressed P. balsamifera loci in the 368 absence of P. balsamifera introgression in the chromosome 9 locus. Future molecular studies and 369 common gardens including pure P. balsamifera individuals in the latitude of origin could be used to 370 further understand the role of this introgressed haplotype and epistatic interactions in clinal 371 adaptation in P. trichocarpa. 372 Adaptive introgression increases plant disease resistance 373 Admixed P. trichocarpa individuals with homozygous P. balsamifera haplotypes on chromosome 9 374 also showed less damage by disease than other admixed individuals, which could help explain their 375 higher height and bole volume (in addition to their longer growing period and the faster active 376 growth rate). Climate change, particularly at higher latitudes, can influence pathogen 377 aggressiveness. Pathogens may thus become increasingly important (Pachauri and Meyer, 2014; 378 Sturrock et al., 2011), especially for P. trichocarpa from northern populations. If adaptive 379 introgression of P. balsamifera on chromosome 9 indeed increases survival under rust attacks, 380 these admixed P. trichocarpa individuals could represent an important resource for forest 381 management and breeding programs. The conservation of natural hybridization represents an 382 underutilized option (Hamilton and Miller, 2015; Johnson et al., 2010) and our work contributes 383 empirical evidence to consider when assessing the conservation value of introgression.  384  18  Candidate genes for adaptive introgression 385 The introgressed region on chromosome 9 was enriched for genes from two protein families: Major 386 Facilitator Superfamily (PF07690) and ABC Transporter (PF00005). Two orthologs of the nitrate 387 transporter AtNRT2 (Potri.009G008500, Potri.009G008600), from the Major Facilitator Superfamily 388 (PF07690), were associated with a number of phenological (bud set, yellowing, leaf drop, growing 389 period, height growth cessation) and biomass traits (height, height gain) here and in a previous 390 association analysis (McKown et al., 2014b). One of the AtNRT2 orthologs (Potri.009G008600), 391 which was also associated with damage by disease in a different association study (La Mantia et al., 392 2013), showed exceptionally strong correlations with geoclimate variables including latitude and 393 temperature (Geraldes et al., 2014; Porth et al., 2015), and had SNPs that were Fst outliers in a 394 landscape genomic analysis (Geraldes et al., 2014). These previous reports used the same P. 395 trichocarpa populations but a different genotyping method and association mapping approach than 396 those used in this study. 397 Nitrate is essential for plant development and NRT2 genes are critical responders to both abiotic 398 and biotic stress (Gojon et al., 2011). In Brassica, expression of NRT2.1 was adversely affected under 399 abiotic stress conditions and in Arabidopsis loss of function mutants showed reduced disease 400 susceptibility to a bacterial pathogen (Pseudomonas syringae) (Camañes et al., 2012). In Poplar, 401 NRT2.1 genes are expressed more strongly in roots than in aerial tissues and are down-regulated in 402 early spring (Sjödin et al., 2009). These genes could function in nitrogen regulation during seasonal 403 remodeling of tree phenology related to the interphase between growth and dormancy (Footitt et 404 al., 2013; Forde, 2000; Larisch et al., 2012).  405  19  Introgressed alleles in a subtelomeric region of chromosome 15 create a transgressive phenotype in P. 406 trichocarpa background 407 This whole genome approach supports the findings from the targeted ancestry analysis showing 408 associations between a subtelomeric region on chromosome 15 and the chlorophyll content index 409 (Suarez-Gonzalez et al., 2016). Here, phenotypic analyses revealed that P. balsamifera alleles in a P. 410 trichocarpa background have an additive effect, with P. trichocarpa individuals homozygous for the 411 introgressed haplotypes showing the highest chlorophyll content index. In the P. balsamifera 412 background, these alleles seem to be acting in a different manner since the chlorophyll content 413 index was lower compared to those in P. trichocarpa admixed individuals with the same alleles. In 414 fact, the chlorophyll content index in pure P. balsamifera was similar to those in pure P. trichocarpa. 415 This suggests that introgressed genes interact with the genetic background of the receiving species 416 and result in a transgressive phenotype (Dittrich-Reed and Fitzpatrick, 2012). It is possible that P. 417 balsamifera transcription factors in chromosome 15 (e.g. PRR5, Potri.015G002300) interact with P. 418 trichocarpa (or P. trichocarpa and introgressed P. balsamifera) transcriptional enhancers and 419 repressors or paralogs in a different manner compared to P. trichocarpa alleles. Since phenology 420 traits were not associated with this subtelomeric region, we ruled out bud set timing as a potential 421 artifact explaining the greater chlorophyll content index in admixed P. trichocarpa individuals. If the 422 chlorophyll content index is associated with higher photosynthetic rates, faster growth, and higher 423 carbon acquisition, as is the case in P. balsamifera (Soolanayakanahally et al., 2009), this phenotype 424 could be of adaptive importance to counteract shorter growing seasons in P. trichocarpa 425 populations at higher latitudes.  426 Overall, our study provides strong support that introgressive hybridization from P. balsamifera 427 generates a reservoir of new genetic variation associated with adaptive characters that may allow 428  20  improved survival in northern regions of the P. trichocarpa range. More generally our results 429 support a growing consensus placing introgression as a central evolutionary process (Arnold and 430 Kunte, 2017). 431 ACKNOWLEDGEMENTS 432 This work was supported by the Genome Canada Large-Scale Applied Research Program (POPCAN, 433 project 168BIO), funds to QCBC and CJD, by a Natural Sciences and Engineering Research Council of 434 Canada (NSERC) Discovery Grant to CJD (RGPIN 36485-12), by NSERC Discovery Grant (RGPIN-2014-435 05820) to QCBC, and by grants from the Swiss National Science Foundation (SNF) to CL. We thank 436 Jaroslav Klápště, Armando Geraldes, Rob Guy and Sally Otto for assistance and discussions. We 437 would also like to thank two anonymous reviewers for their very valuable and detailed comments. 438  439 AUTHOR CONTRIBUTIONS 440 AS-G, CJD, QCBC, and CL designed the study. AS-G performed the research, analyzed data, and 441 wrote the manuscript. CH performed SNP calling. CJD, QCBQ, and CL provided funding. QCBC and CL 442 edited the manuscript and all the authors approved the final version. 443 REFERENCES 444 Alexander, D.H., J. Novembre and K. Lange. 2009. Fast model-based estimation of ancestry in unrelated 445 individuals. Genome Res. 19:1655-1664. 446 Barrett, R.D.H. and D. Schluter. 2008. Adaptation from standing genetic variation. Trends in Ecology & 447 Evolution 23:38-44. 448 Buerkle, C.A. and C. Lexer. 2008. Admixture as the basis for genetic mapping. Trends in Ecology & 449 Evolution 23:686-694. 450  21  Camañes, G., V. Pastor, M. Cerezo, J. Garcia-Andrade, B. Vicedo, P. Garcia-Agustin and V. Flors. 2012. A 451 deletion in NRT2.1 attenuates pseudomonas syringae-induced hormonal perturbation, resulting in 452 primed plant defenses. Plant Physiol. 158:1054-1066. 453 Chuine, I. 2010. Why does phenology drive species distribution? Philosophical Transactions of the Royal 454 Society B: Biological Sciences 365:3149-3160. 455 Cingolani, P., A. Platts, L.L. Wang, M. Coon, T. Nguyen, L. Wang, S.J. Land, X. Lu and D.M. Ruden. 2012. A 456 program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff. Fly 457 6:80-92. 458 Clarkson, C.S., D. Weetman, J. Essandoh, A.E. Yawson, G. Maslen, M. Manske, S.G. Field, M. Webster, T. 459 Antão, B. MacInnis, D. Kwiatkowski and M.J. Donnelly. 2014. Adaptive introgression between 460 anopheles sibling species eliminates a major genomic island but not reproductive isolation. Nature 461 Communications 5:4248. 462 Dittrich-Reed, D. and B.M. Fitzpatrick. 2012. Transgressive hybrids as hopeful monsters. Evolutionary 463 Biology 40:310-315. 464 Evans, L.M., S. Kaluthota, D.W. Pearce, G.J. Allan, K. Floate, S.B. Rood and T.G. Whitham. 2016. Bud 465 phenology and growth are subject to divergent selection across a latitudinal gradient in Populus 466 angustifolia and impact adaptation across the distributional range and associated arthropods. Ecol 467 Evol 6:4565-4581. 468 Footitt, S., Z. Huang, H.A. Clay, A. Mead and W. Finch-Savage. 2013. Temperature, light and nitrate 469 sensing coordinate Arabidopsis seed dormancy cycling, resulting in winter and summer annual 470 phenotypes. The Plant Journal 74:1003-1015. 471  22  Forde, B.G. 2000. Nitrate transporters in plants: Structure, function and regulation. Biochimica Et 472 Biophysica Acta (BBA) - Biomembranes 1465:219-235. 473 Geraldes, A., N. Farzaneh, C.J. Grassa, A.D. McKown, R.D. Guy, S.D. Mansfield, C.J. Douglas and Q.C.B. 474 Cronk. 2014. Landscape genomics of Populus trichocarpa: The role of hybridization, limited gene 475 flow, and natural selection in shaping patterns of population structure. Evolution 68:3260-3280. 476 Gojon, A., G. Krouk, F. Perrine-Walker and E. Laugier. 2011. Nitrate transceptor(s) in plants. J. Exp. Bot. 477 62:2299-2308. 478 Hall, D., V. Luquez, V.M. Garcia, K.R. St Onge, S. Jansson and P. Ingvarsson. 2007. Adaptive population 479 differentiation in phenology across a latitudinal gradient in european aspen (Populus tremula, L): A 480 comparison of neutral markers, candidate genes and phenotypic traits. Evolution 61:2849-2860. 481 Hamilton, J.A. and J.M. Miller. 2015. Adaptive introgression as a resource for management and genetic 482 conservation in a changing climate. Conservation Biology 30:33-41. 483 Harrison, R.G. and E.L. Larson. 2014. Hybridization, introgression, and the nature of species boundaries. 484 J. Hered. 105:795-809. 485 Hedrick, P.W. 2013. Adaptive introgression in animals: Examples and comparison to new mutation and 486 standing variation as sources of adaptive variation. Mol. Ecol. 22:4606-4618. 487 Pachauri R.K. and L.A. Meyer. 2014. IPCC Climate change 2014: Synthesis report. contribution of working 488 groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change. 489 IPCC, Geneva. 490  23  Ismail, M., R. Soolanayakanahally, P. Ingvarsson, R. Guy, S. Jansson, S. Silim and Y. El-Kassaby. 2012. 491 Comparative nucleotide diversity across north american and european Populus species. J. Mol. 492 Evol. 74:257-272. 493 Johnson, W.E., D.P. Onorato, M.E. Roelke, E.D. Land, M. Cunningham, R.C. Belden, R. McBride, D. Jansen, 494 M. Lotz, D. Shindle, J. Howard, D.E. Wildt, L.M. Penfold, J.A. Hostetler, M.K. Oli and S.J. O'Brien. 495 2010. Genetic restoration of the florida panther. Science 329:1641. 496 Jordan, I.K. 2016. The columbian exchange as a source of adaptive introgression in human populations. 497 Biology Direct 11:17. 498 Keller, S.R., N. Levsen, P. Ingvarsson, M.S. Olson and P. Tiffin. 2011a. Local selection across a latitudinal 499 gradient shapes nucleotide diversity in balsam poplar, Populus balsamifera L. Genetics 188:941. 500 Keller, S.R., R.Y. Soolanayakanahally, R.D. Guy, S.N. Silim, M.S. Olson and P. Tiffin. 2011b. Climate-driven 501 local adaptation of ecophysiology and phenology in balsam poplar, Populus balsamifera L. 502 (Salicaceae). Am. J. Bot. 98:99-108. 503 La Mantia, J., J. Klápště, Y. El-Kassaby, S. Azam, R. Guy, C. Douglas, S. Mansfield and R. Hamelin. 2013. 504 Association analysis identifies Melampsora × columbiana poplar leaf rust resistance SNPs 505 . PLoS ONE 8:e78423. 506 Larisch, C., M. Dittrich, H. Wildhagen, S. Lautner, J. Fromm, A. Polle, R. Hedrich, H. Rennenberg, T. Muller 507 and P. Ache. 2012. Poplar wood rays are involved in seasonal remodeling of tree physiology. Plant 508 Physiol. 160:1515-1529. 509  24  Levsen, N.D., P. Tiffin and M.S. Olson. 2012. Pleistocene speciation in the genus Populus (Salicaceae). 510 Syst. Biol. 61:401-412. 511 Little, E.L. 1971. Conifers and important hardwoods. p. 1146-1155. In U.S. Department of Agriculture 512 Miscellaneous Publication (ed.) Atlas of United States trees Volume 1. U.S. Department of 513 Agriculture Miscellaneous Publication, Washington D. C. 514 McKown, A.D., R.D. Guy and L.K. Quamme. 2016. Impacts of bud set and lammas phenology on root: 515 Shoot biomass partitioning and carbon gain physiology in poplar. Trees 30:2131-2141. 516 McKown, A.D., R.D. Guy, J. Klápště, A. Geraldes, M. Friedmann, Q.C.B. Cronk, Y. El-Kassaby, S.D. 517 Mansfield and C.J. Douglas. 2013. Geographical and environmental gradients shape phenotypic 518 trait variation and genetic structure in Populus trichocarpa. New Phytol. 201:1263-1276. 519 McKown, A.D., R.D. Guy, L. Quamme, J. Klápště, J. La Mantia, C.P. Constabel, Y. El-Kassaby, R.C. Hamelin, 520 M. Zifkin and M.S. Azam. 2014a. Association genetics, geography and ecophysiology link stomatal 521 patterning in Populus trichocarpa with carbon gain and disease resistance trade-offs. Mol. Ecol. 522 23:5771-5790. 523 McKown, A.D., J. Klápště, R.D. Guy, A. Geraldes, I. Porth, J. Hannemann, M. Friedmann, W. Muchero, 524 G.A. Tuskan, J. Ehlting, Q.C.B. Cronk, Y. El-Kassaby, S.D. Mansfield and C.J. Douglas. 2014b. 525 Genome-wide association implicates numerous genes underlying ecological trait variation in 526 natural populations of Populus trichocarpa. New Phytol. 203:535-553. 527 Norris, L.C., B.J. Main, Y. Lee, T.C. Collier, A. Fofana, A.J. Cornel and G.C. Lanzaro. 2015. Adaptive 528 introgression in an african malaria mosquito coincident with the increased usage of insecticide-529 treated bed nets. Proceedings of the National Academy of Sciences 112:815-820. 530  25  Olson, M.S., N. Levsen, R.Y. Soolanayakanahally, R.D. Guy, W.R. Schroeder, S.R. Keller and P. Tiffin. 2013. 531 The adaptive potential of Populus balsamifera L. to phenology requirements in a warmer global 532 climate. Mol. Ecol. 22:1214-1230. 533 Orr, H.A. and R.L. Unckless. 2014. The population genetics of evolutionary rescue. PLOS Genetics 534 10:e1004551. 535 Petit, R. and A. Hampe. 2006. Some evolutionary consequences of being a tree. Annu. Rev. Ecol. Evol. 536 Syst. 37:187-214. 537 Plummer, M., N. Best, K. Cowles and K. Vines. 2010. Coda, version 0.14-2. available: http://Cran.r-538 project.org.  539 Porth, I., J. Klápště, A.D. McKown, J. La Mantia, R.D. Guy, P. Ingvarsson, R. Hamelin, S.D. Mansfield, J. 540 Ehlting, C.J. Douglas and Y. El-Kassaby. 2015. Evolutionary quantitative genomics of Populus 541 trichocarpa. Plos One 10:e0142864. 542 Purcell, S., B. Neale, K. Todd-Brown, L. Thomas, M. Ferreira, D. Bender, J. Maller, P. Sklar, P. de Bakker, 543 M.J. Daly and P.C. Sham. 2007. PLINK: A toolset for whole-genome association and population-544 based linkage analysis. American Journal of Human Genetics 81:559-575. 545 Richardson, J., J.G. Isebrands and J.B. Ball. 2014. Ecology and physiology of poplars and willows. p. 92-546 123. In J.G. Isebrands and J. Richardson (eds.) Poplars and willows: Trees for society and the 547 environment. CABI, Oxfordshire, UK. 548 Rieseberg, L.H. and J.F. Wendel. 1993. Introgression and its consequences in plants. p. 70-109. In R.G. 549 Harrison (ed.)  In hybrid zones and the evolutionary process. Oxford Univ. Press, New York. 550  26  Rohde, A., V. Storme, V. Jorge, M. Gaudet, N. Vitacolonna, F. Fabbrini, T. Ruttink, G. Zaina, N. Marron, S. 551 Dillen, M. Steenackers, M. Sabatti, M. Morgante, W. Boerjan and C. Bastien. 2011. Bud set in poplar 552 genetic dissection of a complex trait in natural and hybrid populations. New Phytologist 189:106-553 121. 554 Roshyara, N.R. and M. Scholz. 2014. fcGENE: A versatile tool for processing and transforming SNP 555 datasets. PLoS ONE 9:e97589. 556 Savolainen, O. and T. Pyhäjärvi. 2007. Genomic diversity in forest trees. Curr. Opin. Plant Biol. 10:162-557 167. 558 Scheet, P. and M. Stephens. 2006. A fast and flexible statistical model for large-scale population 559 genotype data: Applications to inferring missing genotypes and haplotypic phase. Am. J. Hum. 560 Genet. 78:629-644. 561 Shriner, D., A. Adeyemo and C.N. Rotimi. 2011. Joint ancestry and association testing in admixed 562 individuals. PLOS Computational Biology 7:e1002325. 563 Sjödin, A., N.R. Street, G. Sandberg, P. Gustafsson and S. Jansson. 2009. PopGenIE: The Populus genome 564 integrative explorer. A new tool for exploring the populus genome. New Phytologist 182:1013-565 1025. 566 Soolanayakanahally, R.Y., R.D. Guy, S.N. Silim and M. Song. 2013. Timing of photoperiodic competency 567 causes phenological mismatch in balsam poplar (Populus balsamifera L.). Plant, Cell & Environment 568 36:116-127. 569  27  Soolanayakanahally, R.Y., R.D. Guy, S.N. Silim, E.C. Drewes and W.R. Schroeder. 2009. Enhanced 570 assimilation rate and water use efficiency with latitude through increased photosynthetic capacity 571 and internal conductance in balsam poplar (Populus balsamifera L.). Plant, Cell Environ. 32:1821-572 1832. 573 Sturrock, R.N., S.J. Frankel, A.V. Brown, P.E. Hennon, J.T. Kliejunas, K.J. Lewis, J.J. Worrall and A.J. 574 Woods. 2011. Climate change and forest diseases. Plant Pathol. 60:133-149. 575 Suarez-Gonzalez, A., C.A. Hefer, C. Christe, O. Corea, C. Lexer, Q.C.B. Cronk and C.J. Douglas. 2016. 576 Genomic and functional approaches reveal a case of adaptive introgression from Populus 577 balsamifera (balsam poplar) in P. trichocarpa (black cottonwood). Mol. Ecol. 25:2427-2442. 578 Tamura, K., J. Dudley, M. Nei and S. Kumar. 2007. MEGA4: Molecular evolutionary genetics analysis 579 (MEGA) software version 4.0. Mol. Biol. Evol. 24:1596-1599. 580 Vitasse, Y., A. Lenz and C. Korner. 2014. The interaction between freezing tolerance and phenology in 581 temperate deciduous trees. Frontiers in Plant Science 5:541. 582 Wang, T., A. Hamann, D.L. Spittlehouse and T.Q. Murdock. 2012. ClimateWNA – high-resolution spatial 583 climate data for western North America. J. Appl. Meteor. Climatol. 51:16-29. 584 Wegmann, D., D.E. Kessner, K.R. Veeramah, R.A. Mathias, D.L. Nicolae, L.R. Yanek, Y.V. Sun, D.G. 585 Torgerson, N. Rafaels, T. Mosley, L.C. Becker, I. Ruczinski, T.H. Beaty, S.L.R. Kardia, D.A. Meyers, K.C. 586 Barnes, D.M. Becker, N.B. Freimer and J. Novembre. 2011. Recombination rates in admixed 587 individuals identified by ancestry-based inference. Nat. Genet. 43:847-853. 588  28  Whitney, K., R.A. . Randell and L. Rieseberg. 2006. Adaptive introgression of herbivore resistance traits in 589 the weedy sunflower Helianthus annuus. Am. Nat. 167:794-807. 590 Whitney, K.D., K.W. Broman, N.C. Kane, S.M. Hovick, R.A. Randell and L.H. Rieseberg. 2015. Quantitative 591 trait locus mapping identifies candidate alleles involved in adaptive introgression and range 592 expansion in a wild sunflower. Mol. Ecol. 24:2194-2211. 593 Xie, C., C.C. Ying, A.D. Yanchuk and D.L. Holowachuk. 2009. Ecotypic mode of regional differentiation 594 caused by restricted gene migration: A case in black cottonwood (Populus trichocarpa) along the 595 pacific northwest coast. Can. J. for. Res. 39:519-525. 596   597 FIGURE LEGENDS 598 Figure 1. Geographic distribution of admixed P. trichocarpa individuals (circles) used in admixture 599 mapping. Additional phenotypic analyses also included pure P. trichocarpa individuals (red 600 triangles). Pure P. balsamifera individuals (blue triangles) were included in the phenotypic 601 analysis of one trait (chlorophyll content index). Ranges of P. trichocarpa and P. balsamifera are 602 shown in red and blue, respectively (Little, 1971). A square represents the location where P. 603 trichocarpa homozygotes for the P. balsamifera haplotype in the introgressed region in 604 chromosome 9 occur (Prince George). Diamonds represent the locations where heterozygous 605 individuals for the introgressed haplotype in chromosome 9 occur.  606  607 Figure 2 Relationship between trait variation and geoclimate variables in admixed and pure P. 608 trichocarpa. A: Correlation between trait variation and maximum length of day (DAY). B: 609 Correlation between trait variation and traits associated with temperature (first principal 610 component based on climate variables associated with temperature explaining 60% of the 611 variance). Asterisks represent significant interactions between the slopes of admixed and pure P. 612 trichocarpa (p < 0.05). Detailed information about the traits can be found in Tables 3; Supporting 613 Information Table S1. 614  615 Figure 3 Trait variation in admixed and pure P. trichocarpa across 16 traits associated with an 616 introgressed haplotype on chromosome 9. bb (blue): homozygotes for the P. balsamifera 617 haplotype, tb (purple): heterozygotes, tt (red): homozygotes for the P. trichocarpa haplotype. 618 Pure P. trichocarpa, all homozygotes for the P. trichocarpa haplotype, from northern (pink) and 619 southern (dark red) populations had similar phenotypes (p>0.05). Asterisks represent traits where 620 bb was significantly different than tt. All 16 traits showed significant differences between tt and 621 pure individuals, but there were no significant differences between bb and pure individuals, 622  29  except in Volumecm3_2011. Additional information about the ANOVAs can be found in 623 Supporting Information Table S3. Detailed information about the traits can be found in Tables 3 624 and Supporting Information Table S1. The box plot displays the distribution of data where the 625 central rectangle spans the first quartile to the third quartile (IQR), the segment inside the 626 rectangle shows the median and "whiskers" above and below the box show the locations of the 627 minimum and maximum. Outliers are 3×IQR or more above the third quartile or 3×IQR or more 628 below the first quartile. 629  630 Figure 4 Variation in the chlorophyll content index in admixed and pure P. trichocarpa (Pure Pt, 631 dark red) and P. balsamifera (Pure Pb, dark blue) individuals. The chlorophyll content index was 632 associated with an introgressed sutelomeric region on chromosome 15. The haplotype in the 633 introgressed subtelomeric region on chromosome 15 is shown inside parentheses [tt (red): 634 homozygotes for the P. trichocarpa haplotype, bt (purple): heterozygotes, bb(blue): homozygotes 635 for the P. balsamifera haplotype]. Data from pure P. balsamifera is only available for estimates of 636 chlorophyll content index from 2015. Shared letters above boxes indicate that the average 637 chlorophyll content indices were not significantly different between those genotypes (p<0.05). 638 See the legend of Figure 3 for details on the box plot.  639  640 TABLES 641 Table 1. Summary of the results from admixture mapping in Populus trichocarpa. The number of 642 traits, SNPs and chromosomes are shown for all associations with posterior probabilities higher 643 than 0.5 (all >0.5). Results are also displayed for association in introgressed regions [based on a 644 whole genome local ancestry study] with posterior probabilities higher than 0.5 (introgressed 645 >0.5) and 0.9 (introgressed >0.9). 646   Posterior probability category Trait category Dataset all >0.5 introgressed >0.5 introgressed >0.9 Biomass 26 22 14 11 Disease 2 2 1 0 Ecophysiology 60 42 3 2 Phenology 20 19 13 7 Total 108 85 31 20 Number of chromosomes 19 19 7 3 SNPs 1,168,955  86,150 2,874 1,275  Associations 223,161  18,525  3,429   647 Table 2. List of the Populus balsamifera introgressed regions found in a whole genome local 648 ancestry study and ancestry-trait associations. ‘Max. balsa %’ refers to the height of the 649 introgressed peaks in terms of the maximum percentage of P. balsamifera ancestry within the 650  30  peak. ‘SNP (fixed-association)’ are SNPs with fixed differences in the pure species and associated 651 with traits based on BMIX. ‘Association (BMIX)’ is the number of traits associated with SNPs 652 based on BMIX, and ‘Association (ANOVA)’ are the traits associated with haplotypes based on 653 ANOVAs.  654 Introgressed regions start end Size (kb) Max balsa % SNP (fixed- association) Association (BMIX) Association (ANOVA) ch01a 47,120,000 47,540,000 420 20.92% 0 0  ch01b 48,000,000 48,380,000 380 19.43% 0 0  ch03 21,040,000 21,620,000 580 20.85% 0 0  ch05a 140,000 380,000 240 19.97% 0 1  ch05b 11,580,000 11,820,000 240 18.52% 0 0  ch06 3,540,000 3,700,000 160 17.78% 0 4  ch07 14,680,000 15,700,000 1,020 27.33% 0 0  ch09 1,380,000 1,760,000 380 19.52% 119 19 16 ch10 22,180,000 22,660,000 480 19.93% 18 21 2 ch11a 1,940,000 2,120,000 180 18.36% 0 0  ch11b 17,580,000 17,780,000 200 20.10% 0 0  ch11c 17,800,000 18,600,000 800 21.81% 0 0  ch14a 12,720,000 13,300,000 580 22.67% 6 15 5 ch14b 13,340,000 13,480,000 140 19.25% 0 0  ch15a 0 580,000 580 20.29% 0 1 3* ch15b 13,480,000 13,920,000 440 19.31% 0 0  ch17a 3,080,000 3,900,000 820 21.50% 0 1  ch17b 9,720,000 10,420,000 700 26.00% 0 0  ch17c 12,720,000 13,380,000 660 22.93% 0 0  *Haplotypes on chromosome 15 were based on SNPs that had fixed differences in the pure species. On chromosome 9, 10 and 14, haplotypes were based on SNPs that had both fixed differences and associations with traits in the BMIX analysis. ANOVAs were conducted for two additional traits that did not show significant associations in BMIX (summer chlorophyll content index in 2009 and 2011) but were significantly associated with SNPs on chromosome 15 in the targeted study on adaptive introgression (Suarez-Gonzalez et al., 2016). SNP: Single Nucleotide Polymorphism BMIX: Bayesian method to explore admixture mapping ANOVA: Analysis of Variance   655  656  31   657 Table 3 List of traits that showed significant associations with introgressed haplotypes in both 658 BMIX and ANOVA analyses in Populus trichocarpa. Traits with posterior probabilities higher than 659 0.9 in introgressed regions [based on a whole genome local ancestry study] are also shown. 660 Detailed information about the traits can be found in Supporting Information Table S1.  661 Trait Chr Trait category Introgressed > 0.9 (BMIX) Activegrowthratecmday_2010 ch09 biomass  Heightcm_2010 ch09 biomass Yes Heightcm_2011 ch09 biomass Yes Heightgaincm_2009 ch09 biomass  Heightgaincm_2010 ch09 biomass Yes Volumecm3_2011 ch09 biomass  AUDPC_2010 ch09 disease  Budsetday_2009 ch09 phenology  Budsetday_2010 ch09 phenology Yes Growthperioddays_2010 ch09 phenology Yes Leafdropday_2009 ch09 phenology Yes Leaflifespandays_2010 ch09 phenology  Yellowing100_2010 ch09 phenology Yes Yellowing25_2010 ch09 phenology  Yellowing50_2010 ch09 phenology  Yellowing75_2010 ch09 phenology Yes Heightcm_2010 ch10 biomass Yes Heightcm_2011 ch10 biomass Yes Activegrowthratecmday_2010 ch14 biomass  Heightcm_2010 ch14 biomass Yes Heightcm_2011 ch14 biomass Yes Heightgaincm_2010 ch14 biomass Yes Yellowing100_2010 ch14 phenology  CCI_ju10 ch15 ecophysiology  Volumecm3: Bole volume calculated assuming a cone (McKown et al. 2013) AUDPC: Area under the disease progress curve (La Mantia et al. 2013) CCI_ju10: Chlorophyll concentration index June 10, 2015 BMIX: Bayesian method to explore admixture mapping ANOVA: Analysis of Variance  662    New Phytologist Supporting Information  Article title: Introgression from Populus balsamifera underlies adaptation and range boundaries in P. trichocarpa  Authors: Adriana Suarez-Gonzalez, Charles A. Hefer, Christian Lexer, Quentin C. B. Cronk and Carl J. Douglas Article acceptance date: 03 August 2017  The following Supporting Information is available for this article: Fig. S1 Geographic distribution of admixed and reference individuals used in local ancestry analysis   Fig. S2 Proportion of P. balsamifera admixed ancestry in admixed P. trichocarpa individuals across 19 chromosomes showing unusually high levels of introgression Fig. S3 Haplotype blocks based on SNPs that had both fixed differences and displayed associations with traits Table S1  List of traits used in the admixture mapping analysis in P. trichocarpa.  Table S2 List of environmental variables used in a principal component analysis.  Table S3  P-values of ANOVAs of 16 traits in pure and admixed P. trichocarpa.   Notes S1: Script for BMIX                       Fig. S1 Geographic distribution of admixed and reference individuals used in local ancestry analysis across the contact zones between P. trichocarpa and P. balsamifera.  Ranges of P. trichocarpa and P. balsamifera are shown in red and blue, respectively (Little 1971).                 Fig. S2 Proportion of P. balsamifera admixed ancestry in admixed P. trichocarpa individuals across 19 chromosomes showing unusually high levels of introgression (sliding window analysis: size: 100-kb, step: 20-kb). Regions with unusually high levels of introgression – peaks above broken line - have ancestry higher than 3 standard deviations from the weighted mean across the whole genome based on SNP density per window (broken line). Chromosomes in gray did not show unusually high levels of introgression. Putative centromeres are shown in gray (Pinosio et al 2016).    Fig. S3 NJ trees based on SNPs that had both fixed differences in the pure species and displayed associations with traits on the BMIX analysis revealing well-defined clusters. The cluster of P. balsamifera haplotypes (cluster with filled blue circles) and cluster of P. trichocarpa haplotypes (cluster with filled red circles) had strong support in chromosome 9 and 15 (bootstrap values: 90 and 80 for chromosome 9 and 15 respectively). The NJ trees were used to determined individuals that were homozygote for P. balsamifera haplotypes (individuals with both haplotypes in the P. balsamifera cluster), homozygote for P. trichocarpa haplotypes (individuals with both haplotypes in the P. trichocarpa cluster) or heterozygote. These trees include admixed P. trichocarpa individuals (red open circles) as well as admixed P. balsamifera individuals (blue open circles). The latter were not used in the admixture mapping analysis but were included in the whole genome local ancestry analysis.  The NJ analyses were conducted in MEGA6.    Table S1 List of traits used in the admixture mapping analysis in P. trichocarpa. Trait code Trait details Trait category Source AB_density Abaxial density (# mm-2) ecophysiology McKown et al 2014 AB_pore_length Abaxial pore length (µm) ecophysiology McKown et al 2014 AB_SPI Abaxial SPI - stomatal pore index. ecophysiology McKown et al 2014 Activegrowthratecmday_2009 Active growth rate cm day biomass McKown et al 2013 Activegrowthratecmday_2010 Active growth rate cm day biomass McKown et al 2013 AD_density Adaxial density (# mm-2) ecophysiology McKown et al 2014 AD_pore_length Adaxial pore length (µm) ecophysiology McKown et al 2014 AD_SPI Adaxial SPI - stomatal pore index. ecophysiology McKown et al 2014 Ad_STM_distribution Adaxial stomata distribution ecophysiology McKown et al 2014 Ad_STM_presence Adaxial stomata presence ecophysiology McKown et al 2014 Ad_StomataNUM1 Adaxial stomata numbers ecophysiology McKown et al 2014 ADAB AD:AB density ratio-AD:AB pore ratio ecophysiology McKown et al 2014 ADAB_PL AD:AB density ratio-AD:AB pore ratio ecophysiology McKown et al 2014 Amax_2009 Maximum photosynthetic rate ecophysiology McKown et al 2013 Amax_2010 Maximum photosynthetic rate ecophysiology McKown et al 2013 Amax_mass_2009 Photosynthetic rate per unit dry mass ecophysiology McKown et al 2013 Amax_mass_2010 Photosynthetic rate per unit dry mass ecophysiology McKown et al 2013 AUDPC_2010 Area under the disease progress curve disease McKown et al 2014 AUDPC_2011 Area under the disease progress curve disease McKown et al 2014 Boledensitykgm_3_2012 Bole density kgm3 biomass McKown et al 2013 Bolemasskg_2012 Bole mass kg biomass McKown et al 2013 Branch_2009 Branch biomass McKown et al 2013 Budbreakday_2010 Bud break day phenology McKown et al 2013 Budbreakday_2011 Bud break day phenology McKown et al 2013 Budsetday_2008 Bud set day  phenology McKown et al 2013 Budsetday_2009 Bud set day  phenology McKown et al 2013 Budsetday_2010 Bud set day  phenology McKown et al 2013 C_N_2009 Carbon nitrogen ecophysiology McKown et al 2013 C_N_2010 Carbon nitrogen ecophysiology McKown et al 2013 Canopydurationdays_2009 Canopy duration days phenology McKown et al 2013 Canopydurationdays_2010 Canopy duration days phenology McKown et al 2013 CCI2015_ap29 Chlorophyll concentration index summer ecophysiology This study CCI2015_au7 Chlorophyll concentration index summer ecophysiology This study CCI2015_jl1 Chlorophyll concentration index summer ecophysiology This study CCI2015_jl15 Chlorophyll concentration index summer ecophysiology This study   Trait code Trait details Trait category Source CCI2015_ju10 Chlorophyll concentration index summer ecophysiology This study CCI2015_ju16 Chlorophyll concentration index summer ecophysiology This study CCI2015_ju24 Chlorophyll concentration index summer ecophysiology This study CCI2015_ma15 Chlorophyll concentration index summer ecophysiology This study CCI2015_ma22 Chlorophyll concentration index summer ecophysiology This study CCI2015_ma7 Chlorophyll concentration index summer ecophysiology This study Chlpost_budsetCCI_2009 Chlorophyll concentration index post bud set ecophysiology McKown et al 2013 Chlpost_budsetCCI_2011 Chlorophyll concentration index post bud set ecophysiology McKown et al 2013 ChlspringCCI_2009 Chlorophyll concentration index spring ecophysiology McKown et al 2013 ChlsummerCCI_2009 Chlorophyll concentration index summer ecophysiology McKown et al 2013 ChlsummerCCI_2011 Chlorophyll concentration index summer ecophysiology McKown et al 2013 d13Cwood_2012 Stable carbon isotope ratio wood ecophysiology McKown et al 2013 d15N_2009 Stable nitrogen isotope ratio ecophysiology McKown et al 2013 d15N_2010 Stable nitrogen isotope ratio ecophysiology McKown et al 2013 DENS Total density (# mm-2) ecophysiology McKown et al 2014 Dleaf_2009 Net discrimination leaf ecophysiology McKown et al 2013 Dleaf_2010 Net discrimination leaf ecophysiology McKown et al 2013 Growthperioddays_2009 Growth period days biomass McKown et al 2013 Growthperioddays_2010 Growth period days biomass McKown et al 2013 gsmol_2009 Stomatal conductance ecophysiology McKown et al 2013 gsmol_2010 Stomatal conductance ecophysiology McKown et al 2013 H_Dcm_cm_2009 Height diameter cm biomass McKown et al 2013 H_Dcm_cm_2010 Height diameter cm biomass McKown et al 2013 H_Dcm_cm_2011 Height diameter cm biomass McKown et al 2013 Heightcm_2008 Height cm biomass McKown et al 2013 Heightcm_2009 Height cm biomass McKown et al 2013 Heightcm_2010 Height cm biomass McKown et al 2013 Heightcm_2011 Height cm biomass McKown et al 2013 Heightgaincm_2009 Height gain cm biomass McKown et al 2013 Heightgaincm_2010 Height gain cm biomass McKown et al 2013 Heightgaincm_2011 Height gain cm biomass McKown et al 2013 Heightgrowthcessationday_2009 Height growth cessation day biomass McKown et al 2013 Leafdropday_2008 Leaf drop day phenology McKown et al 2013 Leafdropday_2009 Leaf drop day phenology McKown et al 2013 Leafdropday_2010 Leaf drop day phenology McKown et al 2013 Leafflushday_2010 Leaf flush day phenology McKown et al 2013 Leafflushday_2011 Leaf flush day phenology McKown et al 2013 Leafflushday_2012 Leaf flush day phenology McKown et al 2013 Leaflifespandays_2010 Leaf life span days phenology McKown et al 2013 Leafshapelength_width_2009 Leaf shape length width ecophysiology McKown et al 2013 Leavesperbud_2011 Leaves per bud ecophysiology McKown et al 2013 Leavesperbud_2012 Leaves per bud ecophysiology McKown et al 2013 LMApost_budset_2010 Leaf mass per unit area post budset ecophysiology McKown et al 2013 LMApost_budset_2011 Leaf mass per unit area post budset ecophysiology McKown et al 2013 LMAspring_2010 Leaf mass per unit area spring ecophysiology McKown et al 2013 LMAspring_2011 Leaf mass per unit area spring ecophysiology McKown et al 2013 LMAsummer_2009.1 Leaf mass per unit area summer ecophysiology McKown et al 2013 LMAsummer_2010.1 Leaf mass per unit area summer ecophysiology McKown et al 2013 LMAsummer_2011.1 Leaf mass per unit area summer ecophysiology McKown et al 2013 Logheightgrowthlogcmday_2009 Log height growth log cm day biomass McKown et al 2013 Logvolumegrowthlogcm3day_2009 Log volume growth log cm3 day biomass McKown et al 2013 Narea_2009 Nitrogen area ecophysiology McKown et al 2013 Narea_2010 Nitrogen area ecophysiology McKown et al 2013 Nmass_2009 Nitrogen mass ecophysiology McKown et al 2013 Nmass_2010 Nitrogen mass ecophysiology McKown et al 2013 NUE_2009 Photosynthetic nitrogen-use efficiency ecophysiology McKown et al 2013 NUE_2010 Photosynthetic nitrogen-use efficiency ecophysiology McKown et al 2013   Trait code Trait details Trait category Source Post_budsetperioddays_2009 Post bud set period days phenology McKown et al 2013 Post_budsetperioddays_2010 Post bud set period days phenology McKown et al 2013 SPI Stomatal pore index. ecophysiology McKown et al 2014 Tannins Tannins (µg mg DW-1) ecophysiology McKown et al 2014 Volumecm3_2009 Volume cm3 biomass McKown et al 2013 Volumecm3_2010 Volume cm3 biomass McKown et al 2013 Volumecm3_2011 Volume cm3 biomass McKown et al 2013 Volumegaincm3_2010 Volume gain cm3 biomass McKown et al 2013 Volumegaincm3_2011 Volume gain cm3 biomass McKown et al 2013 Wholetreemasskg_2012 Whole tree mass kg biomass McKown et al 2013 WUE_2009 Instantaneous water-use efficiency ecophysiology McKown et al 2013 WUE_2010 Instantaneous water-use efficiency ecophysiology McKown et al 2013 Yellowing100_2010 Yellowing 100 phenology McKown et al 2013 Yellowing25_2010 Yellowing 25 phenology McKown et al 2013 Yellowing50_2010 Yellowing 50 phenology McKown et al 2013 Yellowing75_2010 Yellowing 75 phenology McKown et al 2013   Table S2 List of environmental variables used in a principal component analysis. Twenty-three climate variables were compiled from ClimateNA (Wang et al. 2012) based on 1971–2000. Type of variable Abbreviation Variable temperature MAT mean annual temperature (°C) temperature MWMT  mean warmest month temperature (°C) temperature MCMT  mean coldest month temperature (°C) temperature TD temperature difference between MWMT and MCMT or continentality (°C) temperature DD<0 degree-days below 0°C chilling degree-days temperature DD>5 degree-days above 5°C growing degree-days temperature DD<18 degree-days below 18°C heating degree-days temperature DD>18 degree-days above 18°C cooling degree-days temperature NFFD the number of frost-free days temperature FFP frost-free period temperature bFFP the day of the year on which FFP begins temperature eFFP the day of the year on which FFP ends temperature EMT extreme minimum temperature over 30 years temperature EXT extreme maximum temperature over 30 years temperature MAR mean annual solar radiation (MJ m-2 d-1) humidity MAP mean annual precipitation (mm) humidity MSP May to September precipitation (mm) humidity AHM  annual heat-moisture index (MAT+10)/(MAP/1000)) humidity SHM  summer heat-moisture index ((MWMT)/(MSP/1000)) humidity PAS precipitation as snow (mm) between August in previous year and July in current year humidity Eref Hargreaves reference evaporation (mm) humidity CMD Hargreaves climatic moisture deficit (mm) humidity RH mean annual relative humidity (%)                 Table S3 P-values of ANOVAs of 16 traits in pure and admixed P. trichocarpa Comparisons Yellowing100_2010 Heightcm_2010 Heightgaincm_2009 Heightcm_2011 ch09_tb-Ch09_bb 0.4086 0.2839 0.6642 0.1928 Ch09_tt-Ch09_bb 0.0074 0.0148 0.0181 0.0142 Purenorth-Ch09_bb 0.8026 0.8933 0.7922 0.8791 Puresouth-Ch09_bb 0.8955 0.9978 1.0000 0.9847 Ch09_tt-ch09_tb 0.3503 0.7591 0.2562 0.9060 Purenorth-ch09_tb 0.0195 0.0193 0.0606 0.0088 Puresouth-ch09_tb 0.0323 0.1014 0.5749 0.0305 Purenorth-Ch09_tt 0.0000 0.0001 0.0000 0.0001 Puresouth-Ch09_tt 0.0000 0.0012 0.0079 0.0004 Puresouth-Purenorth 0.9993 0.9701 0.8193 0.9913  Growthperioddays_2010 Leaflifespandays_2010 Yellowing75_2010 Heightgaincm_2010 ch09_tb-Ch09_bb 0.2790 0.5959 0.4603 0.3796 Ch09_tt-Ch09_bb 0.0131 0.0253 0.0213 0.0320 Purenorth-Ch09_bb 1.0000 0.9803 0.7427 0.9560 Puresouth-Ch09_bb 0.9870 1.0000 0.9993 0.9990 Ch09_tt-ch09_tb 0.7376 0.3909 0.5527 0.8003 Purenorth-ch09_tb 0.2148 0.2024 0.0174 0.0626 Puresouth-ch09_tb 0.0577 0.5050 0.2489 0.1791 Purenorth-Ch09_tt 0.0060 0.0012 0.0000 0.0008 Puresouth-Ch09_tt 0.0004 0.0089 0.0029 0.0044 Puresouth-Purenorth 0.9913 0.9815 0.8375 0.9897  Budsetday_2010 Yellowing25_2010 AUDPC_2010 Yellowing50_2010 ch09_tb-Ch09_bb 0.2706 0.7051 0.3129 0.4867 Ch09_tt-Ch09_bb 0.0210 0.0351 0.0108 0.0208 Purenorth-Ch09_bb 0.9029 0.9443 1.0000 0.8652 Puresouth-Ch09_bb 0.9205 0.9251 0.9998 0.9815 Ch09_tt-ch09_tb 0.8616 0.3252 0.6139 0.5013 Purenorth-ch09_tb 0.0194 0.1870 0.3034 0.0434 Puresouth-ch09_tb 0.0188 0.9943 0.1608 0.8164 Purenorth-Ch09_tt 0.0002 0.0007 0.0101 0.0001 Puresouth-Ch09_tt 0.0001 0.2649 0.0014 0.0776 Puresouth-Purenorth 1.0000 0.4817 0.9998 0.5027  Volumecm3_2011 Activegrowthratecmday_2010 Leafdropday_2009 Budsetday_2009   ch09_tb-Ch09_bb 0.6619 0.5519 0.4029 0.8804 Ch09_tt-Ch09_bb 0.8374 0.0611 0.2255 0.2710 Purenorth-Ch09_bb 0.0063 0.9796 0.2144 0.6552 Puresouth-Ch09_bb 0.6837 1.0000 0.4192 0.6958 Ch09_tt-ch09_tb 0.9466 0.7359 1.0000 0.7409 Purenorth-ch09_tb 0.0000 0.1734 0.0003 0.0841 Puresouth-ch09_tb 0.0276 0.4331 0.0013 0.0912 Purenorth-Ch09_tt 0.0000 0.0039 0.0000 0.0010 Puresouth-Ch09_tt 0.0263 0.0227 0.0000 0.0009 Puresouth-Purenorth 0.1548 0.9851 0.9909 1.0000  Notes S1: Script for BMIX (Shriner et al., 2011) #script to run IBMIX Shriner et al 2011 - published script only for one SNP one trait #here SNPs and traits are looped and each chromosome is run independently #input files: #data1 (phen=traits) #data2 (globalanc= global ancestries average of local ancestries from RASPberry across chromosomes) #data3 (localanc = local ancestries from RASPberry for each chromosome) #data4 (geno = genotype data using .raw files from plink) #admixture_burden and association_burden should be changed accordingly, based on AR model   for(i in 3:ncol(data1)){   y=data.matrix(data1[,i])    globalanc<-data.matrix(data2[,2])   resF<-NULL   for(j in 2:ncol(data3)){ tryCatch({     localanc<-data.matrix(data3[,j])     geno<-data.matrix(data4[,j])          posterior <- function(x,prior,lambda) {(dchisq(x,1,lambda)*prior)/((dchisq(x,1,lambda)*prior)+(dchisq(x,1,0)*(1-prior)))}     admixture_burden <- 474.59372 #order Max     association_burden <- 149467.2837 #order Max          result <- summary(glm(y~localanc+globalanc,family=gaussian))     admixture_p <- result$coefficients[2,4]     admixture_lambda <- (qnorm(1-0.05/admixture_burden/2)+qnorm(0.8))^2     admixture_prior <- 1/admixture_burden     admixture_test <- qchisq(admixture_p,1,0,lower.tail=FALSE)     admixture_posterior <- posterior(x=admixture_test,prior=admixture_prior,lambda=admixture_lambda)         res<-paste((j-1),admixture_p,admixture_test,admixture_prior,admixture_posterior,admixture_lambda,sep=" ")   resF<-rbind(resF,res) }, error=function(e){cat("ERROR :",conditionMessage(e), "\n")}) } write.table(resF,file=paste("/home/adriana/IBMX/results_admi_oc2015/ch01_trait",(i),"-results.txt",sep=""),row.names=F,col.names=T,quote=F) }  Reference   McKown, A.D., R.D. Guy, J. Klápště, A. Geraldes, M. Friedmann, Q.C.B. Cronk, Y. El-Kassaby, S.D. Mansfield and C.J. Douglas. 2013. Geographical and environmental gradients shape phenotypic trait variation and genetic structure in Populus trichocarpa. New Phytol. 201:1263-1276. McKown, A.D., J. Klápště, R.D. Guy, A. Geraldes, I. Porth, J. Hannemann, M. Friedmann, W. Muchero, G.A. Tuskan, J. Ehlting, Q.C.B. Cronk, Y. El-Kassaby, S.D. Mansfield and C.J. Douglas. 2014. Genome-wide association implicates numerous genes underlying ecological trait variation in natural populations of Populus trichocarpa. New Phytol. 203:535-553. Little, E.L. 1971. Conifers and important hardwoods. p. 1146-1155. In U.S. Department of Agriculture Miscellaneous Publication (ed.) Atlas of United States trees Volume 1. U.S. Department of Agriculture Miscellaneous Publication, Washington D. C. Pinosio, S., S. Giacomello, P. Faivre-Rampant, G. Taylor, V. Jorge, M. Le Paslier, G. Zaina, C. Bastien, F. Cattonaro, F. Marroni and M. Morgante. 2016. Mol. Biol. Evol. 33: 2706-2719. Shriner, D., A. Adeyemo and C.N. Rotimi. 2011. Joint ancestry and association testing in admixed individuals. PLOS Computational Biology 7:e1002325. Wang, T., A. Hamann, D.L. Spittlehouse and T.Q. Murdock. 2012. ClimateWNA – high-resolution spatial climate data for western North America . J. Appl. Meteor. Climatol. 51:16-29.   

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.42591.1-0364053/manifest

Comment

Related Items