Variation within natural populations is fundamental to evolution, as it provides the raw material for natural selection. However, despite its central role, the study of variation and variability (the tendency of a trait to vary) has often been overlooked in evolutionary biology. Gene expression, a dynamic process that determines where, when, and how much of a gene product is formed, is a crucial intermediary between genotype and phenotype, making its variability a key yet understudied aspect of evolutionary processes. This thesis explores the evolutionary implications of gene expression variability in two main studies. Using bulk organ transcriptomic data from outbred individuals of three ray-finned fishes, I investigate (1) how variability reflects constraints on regulatory evolution, and (2) how variability reflects evolvability, based on patterns of duplication gene retention and functional divergence after whole-genome duplication (WGD).
The first study examines inter-individual gene expression variability measured in multiple organs in spotted gar (Lepisosteus oculatus), zebrafish (Danio rerio), and Northern pike (Esox lucius). I find that highly variable genes evolve under weaker purifying selection at the coding sequence, showing that intra-species variability in gene expression is linked to inter-species protein sequence divergence. The key finding is that the variability of a gene is largely determined by its primary organ of expression rather than being independently regulated in each organ. Notably, genes that are most expressed in the brain (brain-biased) are lowly variable across non-nervous organs, suggesting stabilizing selection on regulatory evolution. This suggests that organ-specific selective pressures shape gene regulatory evolution and thus variability.
The second study focuses on how gene expression variability influences the retention and functional divergence of WGD paralogs (ohnologs). Here, I use spotted gar and Northern pike as outgroups to the teleost and salmonid WGDs, respectively, to approximate the ancestral expression pattern pre-duplication. First, results suggest that genes retained in duplicate post-WGD tend to have higher variability preceding duplication, particularly organ-biased genes which evolve under weaker selective constraints relative to broadly expressed genes. Second, by characterizing the expression divergence patterns of teleost ohnolog pairs in zebrafish and relating it to variability in spotted gar, I show that variability is linked to functional divergence of lowly constrained ohnolog pairs. This suggests that variability can facilitate the evolution of gene function post-duplication. These results highlight the interplay between selective constraints and expression evolvability in shaping the evolutionary fates of ohnologs.
Lastly, I provide a perspective on approaches to classify the evolutionary outcomes of duplicate genes using transcriptomic data. This is a challenging task which involves translating conceptual models of post-duplication fates (e.g., subfunctionalization and neofunctionalization, among others) into appropriate empirical tests with suitable metrics and inclusion criteria. I discuss simplifying assumptions (heuristics) in using gene expression as a proxy for function, differences in the verbal and graphical descriptions of conceptual models (semantics), metrics, and other methodological considerations. I also highlight important factors to consider to improve comparability across studies and to advance our knowledge of gene function evolution.