Machine learning-driven GWAS uncovers novel candidate genes for resistance to frosty pod rot and witches' broom disease in cacao.

Cacao (Theobroma cacao), the source of chocolate, is threatened by devastating diseases like frosty pod rot (FPR) and witches' broom disease (WBD), impacting global production and farmer livelihoods. Here, we employ a machine learning-driven genome-wide association study to dissect the genetic architecture of disease resistance and productivity in cacao. Upon analyzing phenotypic data for healthy pod rate along with FPR and WBD resistance across 102 diverse accessions, coupled with single nucleotide polymorphism (SNP) markers mapped to the Criollo and Matina reference genomes, we identified numerous novel candidate genes. These genes are implicated in various biological processes, including cell wall modification, stress response signaling, and defense-related mechanisms. Notably, associations varied between the reference genomes, highlighting the genomic complexity of these traits. Our analyses, using Bootstrap Forest and Boosted Tree models, uncovered loci not previously reported, demonstrating the power of machine learning in uncovering complex genetic interactions. This study offers important insights into the polygenic nature of disease resistance in cacao and presents a genomic roadmap for developing disease-resistant varieties.