Predicting drug responsiveness with deep learning from the effects on gene expression of Obsessive-Compulsive Disorder affected cases

摘要: Abstract Obsessive–Compulsive Disorder (OCD) is a chronic psychiatric condition distinguished by intrusive thoughts followed with the urge to perform repetitive actions. These behaviors are cyclic and uncontrollable for more than short period. Genetic factors have strong influence on developing mental illness. Cognitive Behavioral Therapy (CBT) antidepressants such as Selective Serotonin Reuptake Inhibitors (SSRIs) well-known treatment strategies effectively control condition. Antipsychotic medication alongside SSRI procedure exhibits better results OCD affected individuals. So, find efficacious medications, responsiveness of certain drugs under genetic level patients should be properly scrutinized. But, analyzing gene expressions computationally-intensive task. The extraction useful patterns from high dimensional information needs heavily built learning models. In this paper, discern drug-responsive coherent markers OCD, filter-ensemble fused feature selection model proposed. Furthermore, improve predictability model, an unsupervised deep learning-based method used. This algorithm captures relevant much possible finds logical representation input features. extracted features trained supervised machine algorithms. experimental work carried out in dataset accessed Gene Expression Omnibus (GEO) repository accession number GSE76611. outcome revealed significant active drug responsiveness. effectiveness becomes phenomenal when medications profound following patient’s identified will further helpful transform existing schemes towards personalized treatment.