Expression profile of tumor educated platelets as biomarkers for diagnosis and early detection of cancer

Since liquid biopsy is less invasive than tissue biopsy, studies on liquid biopsy biomarkers for the early detection of cancer and diagnosis are taken into consideration. Expression profiles of tumor-educated platelets (TEP) in liquid biopsy can be used as one of the biomarkers. The use of classification machine learning models, according to the features space derived from the expression data of TEPs, has given us the ability to predict cancer. In this study, we evaluate different types of classification models namely SVM, LDA, logistic regression, boosting, classification tree, and random forest, on the expression profile of TEPs in 230 patients with breast, liver, colorectal, brain, pancreatic, and lung cancers, as well as the expression profile of these genes in 55 healthy individuals. These models were examined on the expression profile of 2000 high variance selected genes. Also, pathway enrichment analysis was performed on these genes by the GSEA preranked method. The results showed that linear SVM and polynomial SVM models have lower error rates than two-class models and linear SVM models have lower error rates than multi-class models. In general, the results of TEP expression profile classification and pathway enrichment analysis indicate that the expression profile of TEPs can be considered as candidate biomarkers.