Bioinformatics and machine learning-driven key genes screening for vortioxetine

Abstract

Vortioxetine is a pharmacological agent that acts as a serotonin modulator and stimulant, with safety and tolerability being important health issues. This study aimed to use bioinformatic and machine learning methods to find differentially expressed genes (DEG) between rats exposed to vortioxetine and matched controls. The GSE236207 dataset (Rattus norvegicus) was obtained from the National Center for Biotechnology Information (NCBI) and analyzed with R, followed by genetic ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analyses, and String's protein-protein interaction network was established to identify important genes. The original datasets were preprocessed in the second step by detecting and correcting missing and noisy data and then merged. After feature selection for the cleaned dataset, machine learning algorithms such as the K-nearest neighbors' algorithm, Naive Bayes, and Support Vector Machine (SVM) were used. In addition, an accuracy of 0.90 was observed with SVM. Leveraging these techniques, the study linked IGFBP7, KLRA22, PROB1, SHQ1, NTNG1, and LOC102546359 to vortioxetine exposure. The bioinformatic analysis revealed 18 upregulated genes and 27 downregulated genes, with all approaches identifying only one common locus, LOC102546359, responsible for noncoding ribonucleic acid (ncRNA) synthesis. The crucial point is that this locus bears no connection to any disease or trigger mechanism, thereby bolstering the safety of vortioxetine.