Machine Learning-Based Heading Date QTL Detection in Rice

Machine Learning-Based Heading Date QTL Detection in Rice

Abstract Quantitative trait locus (QTL) analysis is a powerful approach for identifying variantsassociated with the phenotypic variation of complex traits. However, selecting optimalmethods and pre-processing steps require considerable time and effort. In this study,we demonstrated applicability and replicability of machine learning (ML) models in QTLanalysis by evaluating their performance in comparison with conventional QTL analysismethods using 142 recombinant inbred lines derived from two japonica rice cultivars,Koshihikari and Baegilmi. Random forest and gradient boosting models showed the highestpredictive accuracy, and consistently identified three QTLs associated with headingdate: qDTH3, qDTH6, and qDTH7. Moreover, ML-based QTL analysis detected minor-effectqDTH10, where Koshihikari allele promoted heading date when combined withKoshihikari alleles of qDTH6 and qDTH7. These results demonstrate the applicability of MLmodels in QTL analysis on bi-parental mapping population in rice.