Search

Buckler Lab team develops the HARE method to improve genomic prediction accuracy in maize

Genomic prediction uses DNA information to predict differences in important maize traits, like grain yield or flowering time. Typically, DNA information describes the variants at DNA sites in the maize genome. Although these single-site polymorphisms are convenient to capture genetic variability, they ignore the fact that genetic variability consists of more complex differences in the maize genome, which interact in intricate ways to impact biological mechanisms. As described in their new paper, "Haplotype Associated RNA Expression (HARE) improves prediction of complex traits in maize", Buckler lab postdoc Anju Giri, graduate student Merritt Khaipho-Burch, and former postdoc Guillaume Ramstein, addressed this limitation by representing genetic variability in maize using mRNA variation among gene sequences, rather than differences at DNA sites. They further investigated whether the effect of variation among gene haplotypes could be attributed to differences in the RNA expression of genes.


Giri and Ramstein used the Maize Practical Haplotype Graph (PHG), recently developed by Buckler Lab members Arcadio Valdes, Cinta Romay, and others, to infer gene haplotypes in a diverse sample of maize cultivars. They predicted RNA expression for each gene haplotype and then assessed whether the information contained in predicted RNA expression at each haplotype could accurately predict differences at agronomic traits. In genomic prediction, their method, called HARE (Haplotype Associated RNA Expression), was more accurate than haplotype information alone, or direct measurements of RNA expression.


Their results suggest that using gene haplotypes and functional information, like gene expression, is a useful alternative to the standard approach in which DNA information consists of single-site polymorphisms, without connection to any biological mechanism. Future research will investigate the usefulness of other sources of functional information to improve genomic prediction in maize, especially across diverse populations.