Dynamic quantitative trait loci and copy number variation: the missing heritability of complex agronomic traits

Abstract

Genetic studies have identified thousands of loci controlling various agronomic traits, revealing important biological pathways and providing valuable insights into genetic basis of trait variation. However, genome-wide association studies (GWAS) have explained relatively small heritability of most complex traits, leading to the question of ‘missing’ heritability of complex traits. This study examined the ‘missing’ heritability and offered clues on the underlying genetic architecture of complex traits, such as biomass accumulation in maize. Twelve main effect and 6-pair of epistasis quantitative trait loci (QTL), displaying different patterns of expression at different developmental time points in 261 maize genotypes were used. Copy number variation (CNV) and presence absence variation (PAV) was used to study the genetic architecture in 34 maize genotypes. The identified QTL and CNV were mapped on maize B73 reference genome. A total of 182 genes were found harboured in the detected QTL regions. A complex CNV architecture, such as smaller CNV nested within larger CNV or overlapping CNV regions was detected throughout the maize genome, which may explain the extraordinary traits variation observed in maize. The complex CNV genetic architecture may partly explain the missing heritability. The differential gene expression and their interactions at different developmental time points may also explain the missing heritability. Consequently, the genetic model from final trait values cannot reflect the real gene action during the entire growth and development of a plant. It is necessary to understand the CNV and the dynamics of gene expression for complex traits at different developmental stages as a basis for quantitative trait manipulation.