Supervised machine learning and heterotic classification of maize (Zea mays L.) using molecular marker data

Abstract

The development of molecular techniques for genetic analysis has enabled great advances in cereal breeding. However, their usefulness in hybrid breeding, particularly in assigning new lines to heterotic groups previously established, still remains unsolved. In this work we evaluate the performance of several state-of-art multiclass classifiers onto three molecular marker datasets representing a broad spectrum of maize heterotic patterns. Even though results are variable, they suggest supervised learning algorithms as a valuable complement to traditional breeding programs.