Development of kompetitive allele-specific polymerase chain reaction markers and genomic prediction for winter hardiness in African bermudagrass

Identification of winter hardy turf-type bermudagrass solely relies on phenotypic selection. Although effective, this method takes multiple years of field evaluation before we can reliably select superior genotypes. More efficient and precise selection methods are needed by breeding programs. African bermudagrass (Cynodon transvaalensis Burtt-Davy) has shown variations in winter hardiness and is routinely used for the development of turf-type interspecific hybrid cultivars (C. dactylon × C. transvaalensis). The objectives of this study were to develop kompetitive-allele specific polymerase chain reaction (KASP) markers linked to winter hardiness quantitative trait loci (QTL) and to implement genomic prediction for winter hardiness in African bermudagrass. On the basis of previous genomic information associated with winter hardiness, seven KASP markers linked to winter hardiness traits were developed and validated from three previously identified QTL regions. In addition, through the use of traditional genomic best linear unbiased prediction and machine learning-based random forest selection models, the highest prediction accuracy was 0.52 for winterkill, 0.50 for spring green-up, and 0.40 for spring green-up percent green cover. This study is the first report of polymerase chain reaction (PCR)-based, closely trait-linked molecular marker development and genomic selection in African bermudagrass, providing a foundation for the implementation of advanced methods for improving the selection accuracy and efficiency in winter hardiness improvement of African bermudagrass.