Using an interpretable amino acid-based machine learning method to enhance the diagnosis of major depressive disorder

Background: Major depressive disorder (MDD) is a leading cause of disability worldwide. It is, however, a condition that is frequently overlooked and inadequately managed, given that its diagnosis relies heavily on subjective methods. At present, there are no established biomarkers that have been validated for the purposes of diagnosing and treating MDD. Objective: This study sought to assess the diagnostic and predictive potential of the differences in serum amino acid concentration levels between MDD patients and healthy controls (HCs), integrating them into interpretable machine learning models. Methods: 70 MDD patients and 70 HCs matched in age, gender and ethnicity were recruited for the study. Serum amino acid profiling was conducted by means of chromatography-mass spectrometry (LC-MS). A total of 21 metabolites were analysed, with 17 from a preset amino acid panel and the remaining 4 from a preset kynurenine panel. Logistic regression (with and without feature selection and hyperparameter optimization) was applied to differentiate MDD patients from HCs. Findings: The best-performing model utilized both feature selection and hyperparameter optimization, and it yielded a moderate area under the receiver operating curve (AUC) classification value of 0.76 on testing data. The top five metabolites identified by this model as potential biomarkers for MDD were 3-hydroxy-kynurenine, valine, kynurenine, glutamic acid, and xanthurenic acid. Conclusions: and Clinical Implications Our study highlights the potential of using an interpretable machine learning analysis model based on amino acids to aid and increase the diagnostic accuracy of MDD in clinical practice.