Tuesday, March 20, 2018 at 12:20pm to 1:10pm
The sub-genomes of an allopolyploid will each contain complete, yet evolutionarily divergent, sets of genes. With the availability of affordable genome-wide markers, breeders of allopolyploids now have the opportunity to manipulate individual sub-genomes and investigate interactions of homeoalleles across sub-genomes. We present theory and a statistical framework for partitioning genetic variance and predicting breeding values for each sub-genome and their inter-genomic interactions. Using an allohexaploid wheat breeding population for demonstration, sub-genome main effects and interactions were fit using multi-kernel mixed models for variance component estimation and genomic prediction. Strictly modeling inter-genomic interactions resulted in equivalent increases in genomic prediction accuracy as modeling all pairwise marker interactions. Using the IWGSC RefSeq v1.0 wheat genome sequence, 18,184 triplicate and 5,612 duplicate homeoallelic gene sets were identified and anchored to the nearest GBS marker, forming 10,172 unique sets of homeologous markers. Functional homeologous marker interactions for each homeoallelic marker set were used to predict whole genome breeding values, as well as estimate homeologous main and interaction effects. Using gain in genomic prediction accuracy as a proxy for importance of marker interactions, we show that homeologous marker interactions can explain up to 66% of the additional genetic signal from the additive model. Negative relationships observed between homeologous marker main effects and interaction effects point to a pattern indicative of homeoallelic subfunctionalization. Thus, we provide new tools for breeders of allopolyploid crops to characterize the genetic architecture of existing populations, determine breeding goals, and develop strategies for selection of sub-genome additive effects and inter-genomic epistasis.