Fernando S Aguilar, Carolina Saavedra-Diaz, Hugo Arley Jaimes and Claudia Echeverry
Centro de Investigación de la Caña de Azúcar de Colombia (Cenicaña), Cali, Colombia; fsilvaag@cenicana.org
Sugarcane is a major crop used to produce sugar, ethanol, and energy. Given its importance, plant-breeding programs worldwide focus on selecting varieties with higher biomass and sucrose yield, as well as resistance to major diseases. Selection is generally based on phenotypic information. To be more efficient, Cenicaña has implemented the use of plant genomics to assist the breeding program in the selection of the main agronomic traits (sucrose accumulation, enzymatic activity, plant height, stem diameter, spittlebug and stemborer resistance, and tolerance to waterlogging, among others). Cenicaña selected two populations: a discovery population with 220 genotypes, and a validation population with 150 genotypes from its breeding pool. Both populations have been phenotyped (under controlled or field conditions depending on the trait), and genotyped with GBS, RADSeq, and WGS technologies. In the first approach, a search for associated markers was performed using the GWASPoly package. In a second approach, genomic predictions (GP) were produced for the same traits. The association analysis found 14 SNPs for sucrose, 94 SNPs for enzymatic activity, 5 for plant height, 7 for stem diameter, 7 for spittlebug resistance, 7 for stemborer resistance, and 5 for waterlogging tolerance. SNPs associated with sucrose, spittlebug and waterlogging tolerance were further validated. The trained GP models had an accuracy of 0.64 for sucrose, 0.40 for height, 0.63 for diameter, and 0.59 for tolerance to waterlogging, suggesting a good prediction ability for the different traits. These findings underpin the implementation of a genomic prediction system.