Screening of triploid banana population under natural and controlled black sigatoka disease for genomic selection.

The black sigatoka disease (BSD) is the most important foliar threat in banana production and breeding efforts against it should take advantage of genomic selection (GS) that has become one of the most explored tools to increase genetic gain, save time and reduce selection costs. In order to evaluate the potential of GS in banana for BSD, 210 triploid accessions were obtained from the African Banana and Plantain Research Center (CARBAP) to constitute a training population (TP). The variability in the population was assessed at the phenotypic level using BSD- and agronomic-related traits and at the molecular level using Single Nucleotide Polymorphism (SNPs). The analysis of variance showed a significant difference between accessions for almost all traits measured, while at the genomic group level; there was no significant difference for BSD-related traits. The Index of Non spotted leave among accessions ranged from 0.11 to 0.8. The accessions screening in controlled conditions confirmed the susceptibility of all genomic groups to BSD. The principal components analysis with phenotypic data revealed no clear diversity partition of the population. However, the structure analysis and the hierarchical clustering analysis with SNPs grouped the population into four (4) clusters and two (2) sub-populations respectively. The field and laboratory screening of the banana genomic selection TP confirmed that all genomic groups are susceptible to BSD but did not reveal any genetic structure while SNPs markers exhibited clear genetic structure and provided useful information in the perspective of applying genomic selection.