In conventional multi-trait GWAS approaches the covariance between phenotypes are leveraged to improve the detection of genetic signal, however with these approaches we don’t know how the QTL affects the phenotype(s). Medhi used a Bayesian network approach to eludicate the relationship between biomass (shoot and root), water use, and water use efficiency. Structural equation modeling integrated with GWAS to decompose QTL effects and elucidate how a QTL acts on mutiple traits.
