Machine learning using genetic and clinical data identifies a signature that robustly predicts methotrexate response in rheumatoid arthritis

Abstract

OBJECTIVE: To develop a hypothesis-free model that best predicts response to methotrexate (MTX) drug in rheumatoid arthritis (RA) patients utilizing biologically meaningful genetic feature selection of potentially functional single nucleotide polymorphisms (pfSNPs) through robust machine learning (ML) feature selection methods. METHODS: MTX-treated RA patients with known response were divided in a 4:1 ratio into training and test sets. From the patients' exomes, potential features for classifier prediction were identified from pfSNPs and non-genetic factors through ML using recursive feature elimination with cross-validation incorporating Random Forest Classifier. Feature selection was repeated on random subsets of the training cohort, and consensus features were assembled into the final feature set. This feature set was evaluated for predictive potential using six ML classifiers, first by cross-validation within the training set, and finally by analyzing its performance with the unseen test set. RESULTS: The final feature set contains 56 pfSNPs and five non-genetic factors. The majority of these pfSNPs are located in pathways related to RA pathogenesis or methotrexate action and are predicted to modulate gene expression. When used for training in six ML classifiers, performance was good in both the training set (AUC : 0·855-0·916, sensitivity : 0·715-0·892 and specificity : 0·733-0·862) in the unseen test set (AUC : 0·751-0·826, sensitivity: 0·581-0·839 and specificity: 0·641-0·923). CONCLUSION: Sensitive and specific predictors of MTX response in RA patients were identified in this study through a novel strategy combining biologically meaningful and machine learning feature selection and training. These predictors may facilitate better treatment decision-making in RA management.