Integrating transcriptomic network reconstruction and QTL analyses reveals mechanistic connections between genomic architecture andBrassica rapadevelopment

Abstract

63Plant developmental dynamics can be heritable, genetically correlated with fitness and 64 yield, and undergo selection. Therefore, characterizing the mechanistic connections between 65 the genetic architecture governing plant development and the resulting ontogenetic dynamics 66 of plants in field settings is critically important for agricultural production and evolutionary 67 ecology. We use a hierarchical Bayesian Function-Valued Trait (FVT) approach to estimate 68 Brassica rapa growth curves throughout ontogeny, across two treatments and in two growing 69 seasons. We find that the shape of growth curves is relatively plastic across environments 70 compared to final height, and that there are trade-offs between growth rate and duration. We 71 determined that combining FVT Quantitative Trait Loci (QTL) and genes/eigengene expression 72 identified via transcriptomic co-expression network reconstructions best characterized 73 phenotypic variation. Further, targeted eQTL analyses identified regulatory hotspots that 74 colocalized with FVT QTL and co-expression network identified genes and mechanistically link 75 FVT QTL with structural trait variation throughout development in agroecologically relevant 76 field settings. 77 78 al. Li et al. 2018;Luo et al. 2018;Schaefer et al. 2018). However, co-expression network 92 analyses can also provide insight into the mechanistic connections between QTL genotypes and 93 phenotypes. Here, we ask whether QTL, co-expression analyses, or a combination thereof best 94 predict phenotypic variation. In combination with a targeted eQTL analyses in agroecologically 95 relevant field settings, we characterize the mechanistic connections between the genomic 96 architecture, transcriptomic expression networks, and phenotypic variation throughout plant 97 development. 98