Open Collections

The interpretation of gene coexpression in systems biology

University of British Columbia

One of the key features of transcriptomic data is the similarity of expression patterns among groups of genes, referred to as coexpression. It has been shown that coexpressed genes tend to share similar functions. Based on this, a common assumption is that gene coexpression is a result of transcriptional regulation and therefore, regulatory relationships could be inferred from coexpression. However, success in inferring such relationships has been limited and there are questions about the source and interpretation of coexpression. Here I explore coexpression as an observed signal from the data, examine its source and assess its relevance for inferring regulatory relationships. In chapter 2 I studied differential coexpression, which refers to the alteration of gene coexpression between biological conditions. It is commonly assumed that differential coexpression can reveal rewiring of transcription regulatory networks, specifically among the genes that maintain their average expression level between the conditions. However, I show that to a large extent and in contrast to this common assumption, differential coexpression is more parsimoniously explained by changes in average expression levels. This finding demonstrates limitations for inference of regulatory rewiring from coexpression and poses questions for the underlying causes of the observed coexpression. In Chapter 3, I studied cellular composition variation among bulk tissue samples as a source of variance and the observed coexpression. I found that for most genes, differences in expression levels across cell types account for a large fraction of their variance and as a result genes with similar cell-type expression profiles appear to be coexpressed. Finally, I showed that this coexpression dominates the underlying intra-cell-type coexpression and also has the two prominent features of coexpression in the bulk tissue: reproducibility and biological relevance. Through my studies, I was able to provide an explanation for much of the observed coexpression in the bulk tissue and shed light on its resolution and limitation for inference of regulatory relationships. I also studied coexpression in single-nucleus data and show that some of the observed coexpression in it is likely to be attributed to the transcriptional regulation, which could be a subject for future studies.

Text; Dataset

2020-01-06

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/73156

Bioinformatics

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/