Selective serotonin reuptake inhibitors (SSRIs) like sertraline are crucial in treating depression and anxiety disorders, and studies indicate their potential as chemosensitizers in cancer therapy. This research develops a machine-learning predictive model to identify novel compounds with sertraline-like antidepressant activity. We constructed and validated a customized machine-learning model to predict SSRI activity in new compounds. By applying feature engineering to the chemical structures and bioactivity data of sertraline and its analogs, we trained multiple machine-learning algorithms. Through extensive comparative analysis, we found that the support vector machine (SVM) model demonstrated exceptional performance, achieving an accuracy rate of up to 93%. By further optimizing and integrating the SVM model, we successfully enhanced its accuracy, reaching an impressive 95% capability in predicting more active SSRI compounds. This study successfully developed a targeted, rapid, and efficient machine learning model capable of accurately predicting SSRI activity. The model serves as a valuable tool for rapidly screening novel SSRI drug candidates with superior activity, bringing immense value to the field of drug development.
Keywords: SSRI activity prediction; drug discovery; escitalopram analogs; feature engineering; machine learning; predictive model.