Unravelling the Genetic Architecture of Field Traits through Multi-Omics Platform Data integration

Abstract

BackgroundIdentifying the genes and regulatory regions underlying a complex trait is a long-standing challenge. GWAS is generally used, but it suffers from insufficient power, lack of resolution and inefficiency to systematically screen for epistasis. We propose a systems genetics approach integrating multi-omics to overcome these limits. It was applied to a panel of maize with the objective to analyze the genetic determinism of yield in a multi-environment trials by combining genomics with transcriptomics measured on a platform. ResultsDespite the contrasted conditions between the platform and the fields, transcriptomics could be used to identify candidate genes. A presence-absence variation was in particular detected, and the transcripts allowed the identification of causal genes, increasing resolution in comparison to GWAS. In total, 47 genes were identified along the genome, and we could characterize their contrasted effect on yield according to environmental covariates. We demonstrated that the cis- and also the trans-eQTLs of these genes had an important contribution to genetic variance, suggesting a key role of epistatic interactions. In terms of predictive ability, the cis-eQTL resulted in an increase of 39 to 52% on average across the environments, in comparison to random SNP sets. ConclusionsBy efficiently combining multi-omics, it is possible to considerably increase our understanding of genetic architecture in comparison to standard GWAS. We demonstrated that omics data even measured on a phenotyping platform can be used for the analysis of field traits, opening the way for their routine use in plant breeding both for marker-assisted selection and bio-informed predictions.