The relevance of dominance and functional annotations to predict agronomic traits in hybrid maize

Abstract

25Heterosis has been key to the development of maize breeding but describing its genetic basis has 26 been challenging. Previous studies of heterosis have shown the contribution of within-locus 27 complementation effects (dominance) and their differential importance across genomic regions. 28 However, they have generally considered panels of limited genetic diversity and have shown 29 little benefit to including dominance effects for predicting genotypic value in breeding 30 populations. This study examined within-locus complementation and enrichment of genetic 31 effects by functional classes in maize. We based our analyses on a diverse panel of inbred lines 32 crossed with two testers representative of the major heterotic groups in the United States (1,106 33 hybrids), as well as a collection of 24 biparental populations crossed with a single tester (1,640 34 hybrids). We assayed three agronomic traits: days to silking (DTS), plant height (PH) and grain 35 yield (GY). Our results point to the presence of dominance for all traits, but also among-locus 36 complementation (epistasis) for DTS and genotype-by-environment interactions for GY. 37Consistently, dominance improved genomic prediction for PH only. In addition, we assessed 38 enrichment of genetic effects in classes defined by genic regions (gene annotation), structural 39 features (recombination rate and chromatin openness), and evolutionary features (minor allele 40 frequency and evolutionary constraint). We found support for enrichment in genic regions and 41 subsequent improvement of genomic prediction for all traits. Our results point to mechanisms by 42 which heterosis arises through local complementation in proximal gene regions and suggest the 43

Prediction of evolutionary constraint by genomic annotations improves prioritization of causal variants in maize

“…The functional relevance of predicted PNC was also supported by its positive association with chromatin accessibility (Fig. 4b), which is correlated with phenotypic effects in maize [14,17]. However, there was a significant increase in expression QTL (eQTL) effect only for observed PNC (P = 0.003 and P = 0.034 in shoot and root tissues respectively, compared to P = 0.120 and P = 0.485 for predicted PNC; Fig.…”

“…The positive result for GY was not observed when training a genomic prediction model in Ames-H. In this panel representative of maize diversity, variation at SNPs -and the information available to learn their effect -was negatively correlated with species-wide MAF [14].…”

“…Accordingly, enrichment of rare SNPs (MAF < 0.01 in Hapmap 3.2.1) for genomic effects on GY was large and significant in NAM-H (> 16-fold enrichment) but not in Ames-H [14].…”

“…PICNC did not rely on within-species variability, so we could include nonsynonymous mutations at monomorphic sites in maize for model training. This helped us avoid survivorship bias at SNP sites and provide many more instances of PNC to learn about the genomic characteristics of fitness effects: 20,136,310 monomorphic sites instead of 483,448 SNPs in Hapmap 3.2.1 (the reference panel in maize) [13] or 103,905 SNPs in elite maize panels (hybrid panels) [14] (Fig. 1, Supplementary Fig.…”

“…Expectation (i) was not met for any of the agronomic traits (Supplementary Fig. S7), probably because of the large LD extent in the hybrid panels (average squared correlation above 0.1 within 100-kb distance), such that causal variants were adequately tagged even by randomly weighted SNPs [14]. Expectation (ii) was met for GY, our trait most related to fitness; a gradual increase in prediction accuracy was observed as prioritization of SNPs was more stringent, with a trend similar to that for lower MAF (Supplementary Fig.…”