Machine Learning Approaches for Genotype to Phenotype Prediction in Plant Science

Advancements in genome sequencing have facilitated a deeper understanding of genetic variation in plants. However, predicting phenotypes from genomic data remains challenging. Machine learning (ML) techniques offer promising solutions in this context, particularly in deciphering genotype-phenotype relationships. Our study compares various ML methods to investigate this association using almond cultivar data. We preprocess the dataset, conduct feature selection, and employ regression models to predict almond yields. Results demonstrate strong genotype-phenotype associations, as indicated by evaluation metrics. Importantly, we employ explainable artificial intelligence algorithms to enhance model interpretability, identifying crucial genetic variations impacting yield. This underscores ML’s efficacy in predicting phenotypic traits from genomic data and highlights its significance in optimizing crop production for sustainable agriculture.