Plasma biomarker panel for major depressive disorder by quantitative proteomics using ensemble learning algorithm: A preliminary study

Major depressive disorder (MDD) is a major international public health issue; thus, investigating its underlying mechanisms and identifying suitable biomarkers to enable its early detection are imperative. Using data -independent acquisition-mass spectrometry-based proteomics, the plasma of 44 patients with MDD and 25 healthy controls was studied to detect differentially expressed proteins. Bioinformatics analyses, such as Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analysis, Protein-Protein Interaction network, and weighted gene co-expression network analysis were employed. Moreover, an ensemble learning technique was used to build a prediction model. A panel of two biomarkers, L-selectin and an isoform of the Ras oncogene family was identified. With an area under the receiver operating characteristic curve of 0.925 and 0.901 for the training and test sets, respectively, the panel was able to distinguish MDD from the controls. Our investigation revealed numerous potential biomarkers and a diagnostic panel based on several algorithms, which may contribute to the future development of a plasma-based diagnostic approach and better understanding of the molecular mechanisms of MDD.