Characterization of Genetic Diversity and Genomic Prediction of Secondary Metabolites in Pea Genetic Resources

This study aimed to characterize the variation and genetic architecture of traits with nutritional and health relevance in 156 pea (Pisum sativum L.) accessions representing diverse geographic origins. The traits included total phenolic compounds (TPC), two saponins (Ssβg, Ss1), sucrose, three raffinose-family oligosaccharides (RFOs) and in vitro antioxidant activity (AA). Analysis of variance revealed significant effects of regional germplasm pools for all traits. Accessions from West Asia showed the highest TPC and AA levels, while those from the East Balkans and the UK displayed the lowest values. High saponin and RFO concentrations characterized accessions from Germany and the UK. Correlation and PCA analyses highlighted strong associations within compound classes and an overall negative relationship between TPC/AA and saponins/RFOs. Hierarchical clustering separated accessions into seven metabolically distinct groups partially reflecting their geographic origin. Linkage disequilibrium decayed rapidly (average 4.7 kb). GWAS with FarmCPU and BLINK identified 37 significant SNPs, 35 within annotated genes, associated with the metabolites. The polygenic genetic architecture supported the development of genomic prediction models, which showed moderately high predictive ability (> 0.40) for all traits except raffinose content. Our findings can support line selection and the identification of genetic resources with a desired level of secondary metabolites.