RNA-seq meta-analysis and machine learning identify stress-responsive genes and improve genomic prediction in common bean (Phaseolus vulgaris L.) with cross-species application in cowpea (Vigna unguiculata L.)

Common bean (Phaseolus vulgaris L.) is a crop stressed by numerous abiotic and biotic factors, limiting its productivity. The molecular mechanisms behind stress tolerance are not well understood due to conflicting and incomplete research. A team of scientists performed a systematic meta-analysis of publicly available RNA-sequencing data across various stress conditions and tissues in common bean.

By combining statistical meta-analysis with machine-learning techniques, the researchers identified a set of robust stress-responsive genes validated through independent dataset validation.

The machine-learning framework uncovered stress-responsive genes that were not detected in earlier studies due to their failure to meet traditional significance thresholds. These genes were analyzed for their functional connections using co-expression and protein-protein interaction network analyses, leading to the identification of gene clusters associated with specific stress-response pathways.

Ethylene-responsive transcription factors were found to be hub genes in three of the four stress-tissue groups, while NAC domain transcription factors emerged as additional hub genes in response to biotic stress.

The identified gene sets were validated for their biological significance, showing that marker panels targeting these regions improved the accuracy of genomic predictions for disease resistance traits in common bean and abiotic stress tolerance traits in cowpea. This cross-species application of the machine-learning framework provided a reliable and evidence-based approach for prioritizing candidate genes and developing genomic selection tools, ultimately aiding in the breeding of stress-resilient legumes.