QTL Identification, Mega-Environment Classification, and Strategy Development for Marker-Based Selection Using Biplots

Abstract

SUMMARY This paper describes a biplot approach to QTL identification based on phenotypic data from multiple environments, and demonstrates its use in the investigation of QTL-by-environment patterns. The effects of each marker on the target trait were estimated for each environment, leading to a marker-by-environment two-way table. This table was then visually investigated in a marker-by-environment biplot. In the biplot, markers with short vectors should have little or no associations with the trait and can be deleted. The remaining markers would fall into clusters, each suggesting the existence of one or more QTL with similar QTL-by-environment patterns. Within each cluster, the marker with the longest vector should be the one located closest to the QTL. When each QTL is represented by its closest marker, the marker-by-environment biplot is referred to as a QTL-by-environment (QQE) biplot. It can help visualize (1) groups of QTL with similar environmental responses; (2) major vs. minor QTL; (3) the average effect of a QTL and its stability across environments; (4) groups of environments with similar expressions of QTL effects, and (5) QTL allele combinations for maximizing/minimizing the expression of the trait for each mega-environment.