Mango seedling genotyping reveals potential self-incompatibility and pollinator behavior

Abstract

Mango (Mangifera indica) is tropical fruit widely cultivated and appreciated for its sweet and aromatic fruit. There is great diversity among mango cultivars for important fruit quality and agronomic traits that could be leveraged as part of a mango breeding program. Historically, most genetic improvement of this species has been the result of chance seedlings. The recent publication of multiple mango genomes and availability of genomic resources is facilitating the efficient breeding of mango, but controlled pollination of mango is challenging with only a few out of thousands of panicle-borne flowers developing into fruit. Therefore, genotyping open-pollinated seedlings could be a labor-saving method to select favorable hybrids in a breeding program. An array of 384 single-nucleotide-polymorphism markers was used to genotype 140 mango cultivars and 575 mango seedlings from eleven open-pollinated populations in a mango diversity collection. Analysis of mango cultivars confirmed hybrid parentage in many instances and identified potential paternal parents for cultivars with unknown parentage. For the seedlings analyzed, the Dixon test of Gower's genetic dissimilarity (GD) values among identified triads revealed 161 significant triads (P≤0.05), indicating high confidence for paternal identification. The remaining triads had GD values less than 0.15 (n = 323), between 0.15 and 0.2 (n = 10), or between 0.2 and 0.3 (n = 5). In total, 70% of the hybrids were paired to a single paternal parent. Analysis of selfing rates indicated ‘Tommy Atkins’ had the highest selfing rate (40.4%) and ‘Glenn’ had the lowest (1.1%). Analysis of population structure and phylogeny divided the mango germplasm into three distinct clusters. The identified paternal pollen donors were mapped within a mango diversity collection and indicated mango pollinators can affect pollination beyond adjoining trees. This strategy can be used to aid mango breeding research and has identified intriguing hybrids for future evaluation.