Post-hoc permutation GWAS refinement reveals robust marker-trait associations in faba bean

Introduction Faba bean ( Vicia faba L.) is a high-yielding, protein-rich grain legume with considerable potential for sustainable cropping systems. However, its genetic improvement has been constrained by a large, complex genome and historically limited genomic resources. We aimed to identify agronomically relevant QTLs in a diverse faba bean panel using genome-wide associations studies (GWAS) with rigorous control of false positives due to model violations. Methods The NORFAB faba bean diversity panel was evaluated for agronomic traits over four years (2018–2021) at two locations in Denmark. GWAS were conducted using three complementary methods, incorporating a standard predictor-permutation test, novel in a GWAS context, for the top associations post-hoc to rigorously control false positives due to model violations. Results This approach identified 13 QTLs associated with nine agronomic traits. Several QTLs colocalized with genes previously associated with similar traits, such as the flowering time regulator FLOWERING LOCUS T . Incorporating a permutation step consistently discovered more internally consistent QTLs compared to using only parametric p -values, with each QTL detected by more GWAS methods and associated with a greater number of SNPs, traits and trials. Discussion Our results add interesting QTLs and biologically plausible candidate genes for multiple agronomic traits for incorporation into the faba bean molecular breeding toolkit and suggest that SNP-level post-hoc permutation testing can improve GWAS reliability.