Identification of CYP89A6 as a positive regulator of glucosinolate biosynthesis in Brassica rapa via machine learning analysis of large-scale RNA-seq data

Glucosinolates (GLS) are critical for the defense, flavor, and sulfur storage of cruciferous plants; however, the regulatory mechanisms underlying their biosynthesis and metabolism remain largely elusive. This study aimed to systematically identify the key genes regulating GLS. To this end, a systematic screening and validation pipeline was established by integrating machine learning predictions (leveraging over 1000 RNA-seq samples from Brassica rapa and Arabidopsis thaliana), multidimensional bioinformatics validation (functional annotation, protein interaction networks), and molecular experimental validation. With this pipeline, 12 high-confidence candidate regulatory genes were identified, as potential targets for subsequently analyzing the multifaceted regulatory network governing GLS biosynthesis and metabolism. Further functional validation of the key candidate gene CYP89A6 through virus-induced gene silencing assays confirmed its role as a positive GLS biosynthesis regulator. Silencing CYP89A6 significantly reduced the total GLS content in B. rapa, with markedly reduced levels of key components including aliphatic GLS (PRO, GRA, GBN), aromatic GLS (NAS), and indole GLS (4 ME, GBC). Overall, this study validated the robustness of the machine learning and multidimensional bioinformatics screening pipeline, and advanced our understanding underlying the regulatory mechanisms of secondary metabolism in B. rapa.