A functionally, structurally, and evolutionarily distinct set of genes are linked to phenome wide variation in maize

Naturally occurring functionally variable alleles in specific genes within a population allows the identification of which genes are involved in the determination of which phenotypes. The omnigenetic model proposes that essentially all genes which are expressed in relevant contexts likely play some role in determining phenotypic outcomes. Here, we develop an approach to identify genes where natural functional variation plays a role in shaping many phenotypic traits simultaneously. We demonstrate that this approach identifies a distinct set of genes relative to conventional genome wide association using data for 260 traits scored a maize diversity panel, and the genes identified using this approach are more likely to be independently validated than genes identified by convetional genome wide association. Genes identified by the new approach share a number of features with a gold standard set of genes characterized through forward genetics which separate them from both genes identified by conventional genome wide association and the overall population of annotated gene models. These features include evidence of significantly stronger purifying selection, positional conservation across the genomes of related species, greater length, and a scarcity of presence absence variation for these loci in natural populations. Genes identified by phenome-wide analyses also showed much stronger signals of GO enrichment and purification than genes identified by conventional genome wide association. Overall these findings are consistent with large subset of annotated gene models - despite support from transcriptional and homology evidence - being unlikely to play any role in determining organismal phenotypes.