A model and test for coordinated polygenic epistasis in complex traits

Abstract

Interactions between genetic variants—epistasis—is pervasive in model systems and can profoundly impact evolutionary adaption, population disease dynamics, genetic mapping, and precision medicine efforts. In this work, we develop a model for structured polygenic epistasis, called coordinated epistasis (CE), and prove that several recent theories of genetic architecture fall under the formal umbrella of CE. Unlike standard epistasis models that assume epistasis and main effects are independent, CE captures systematic correlations between epistasis and main effects that result from pathway-level epistasis, on balance skewing the penetrance of genetic effects. To test for the existence of CE, we propose the even-odd (EO) test and prove it is calibrated in a range of realistic biological models. Applying the EO test in the UK Biobank, we find evidence of CE in 18 of 26 traits spanning disease, anthropometric, and blood categories. Finally, we extend the EO test to tissue-specific enrichment and identify several plausible tissue–trait pairs. Overall, CE is a dimension of genetic architecture that can capture structured, systemic forms of epistasis in complex human traits.

A polygenic-score-based approach for identification of gene-drug interactions stratifying breast cancer risk

“…Conversely, if the SNP main effects have no correlation with the SNP interaction effects with the drug (''uncoordinated interactions''), then there is no power to detect a significant interaction. 37,38 Based on this, we can conclude that testing for a polygenic-drug interaction is an effective approach for assessing the relationship between main and interaction effects and thereby identifying drugs that modulate genetic risk. Furthermore and importantly, PGSs incorporate variation from multiple pathways.…”

“…Introduction to coordinated interactions through the lens of drug response As described previously by Sheppard et al 37 and Aschard, 38 statistical interactions between PGSs and other factors are driven by SNPs broadly interacting positively or negatively in a model named a coordinated interaction. Thus, a significant interaction between PGS and drug use (a ''polygenicdrug interaction'') with regard to a particular phenotype means that drug use serves to strengthen or dampen-on average-the marginal genetic effects of causal variants.…”

“…Second, the number of genetic variants to study can be enormous, the medication data available can be complex, and the interaction effects can be small and difficult to detect. 5 To overcome these challenges, we leveraged coordinated interactions, which aggregate antagonistic or synergistic interaction effects across SNPs 37,38 (Figure 3) and which have been applied across a range of phenotypes and exposures previously. [24][25][26][27][28][29][30][31] We utilize this PGS-based interaction testing approach, along with a mapping of 3,603 distinct medication groups in UKB to a final 99 ATC groups, to perform a simple analysis of the interaction between genetics and medication usage (Figure 2).…”

Discovering non-additive heritability using additive GWAS summary statistics

“…This is in part due to the longstanding and ongoing debate about the contribution of non-additive effects (e.g., epistasis and dominance effects) on the architecture of human complex traits [16][17][18][19][20][21][22][23][24][25][26] . However, despite these controversies, many association mapping studies in humans have identified candidates of epistasis that notably contribute to trait variation [27][28][29][30][31] , and some have recently shown that gene-by-gene interactions can drive heterogeneity of causal variant effect sizes across diverse human populations 32 . Epistasis is a well-known contributor to trait architecture in several model organisms [33][34][35][36][37][38][39][40][41][42][43][44] .…”

DNARecords: An extensible sparse format for petabyte scale genomics analysis

“…The copyright holder for this preprint this version posted August 15, 2022. ; https://doi.org/10.1101/2022.08.13.503863 doi: bioRxiv preprint 3 on datasets oriented in a sample-wise fashion. Lastly, there is significant evidence that there is a small, but significant non-additive component to the heritability of complex human traits [10], [11]. Modeling these nonlinearities is facilitated by data structures organized by sample.…”