Machine learning analysis of lung squamous cell carcinoma gene expression datasets reveals novel prognostic signatures.

BACKGROUND: Lung squamous cell carcinoma (LUSC) patients are often diagnosed at an advanced stage and have poor prognoses. Thus, identifying novel biomarkers for the LUSC is of utmost importance. METHODS: Multiple datasets from the NCBI-GEO repository were obtained and merged to construct the complete dataset. We also constructed a subset from this complete dataset with only known cancer driver genes. Further, machine learning classifiers were employed to obtain the best features from both datasets. Simultaneously, we perform differential gene expression analysis. Furthermore, survival and enrichment analyses were performed. RESULTS: The kNN classifier performed comparatively better on the complete and driver datasets' top 40 and 50 gene features, respectively. Out of these 90 gene features, 35 were found to be differentially regulated. Lasso-penalized Cox regression further reduced the number of genes to eight. The median risk score of these eight genes significantly stratified the patients, and low-risk patients have significantly better overall survival. We validated the robust performance of these eight genes on the TCGA dataset. Pathway enrichment analysis identified that these genes are associated with cell cycle, cell proliferation, and migration. CONCLUSION: This study demonstrates that an integrated approach involving machine learning and system biology may effectively identify novel biomarkers for LUSC.

Machine learning assisted analysis of breast cancer gene expression profiles reveals novel potential prognostic biomarkers for triple-negative breast cancer.

Tumor heterogeneity and the unclear metastasis mechanisms are the leading cause for the unavailability of effective targeted therapy for Triple-negative breast cancer (TNBC), a breast cancer (BrCa) subtype characterized by high mortality and high frequency of distant metastasis cases. The identification of prognostic biomarker can improve prognosis and personalized treatment regimes. Herein, we collected gene expression datasets representing TNBC and Non-TNBC BrCa. From the complete dataset, a subset reflecting solely known cancer driver genes was also constructed. Recursive Feature Elimination (RFE) was employed to identify top 20, 25, 30, 35, 40, 45, and 50 gene signatures that differentiate TNBC from the other BrCa subtypes. Five machine learning algorithms were employed on these selected features and on the basis of model performance evaluation, it was found that for the complete and driver dataset, XGBoost performs the best for a subset of 25 and 20 genes, respectively. Out of these 45 genes from the two datasets, 34 genes were found to be differentially regulated. The Kaplan-Meier (KM) analysis for Distant Metastasis Free Survival (DMFS) of these 34 differentially regulated genes revealed four genes, out of which two are novel that could be potential prognostic genes (POU2AF1 and S100B). Finally, interactome and pathway enrichment analyses were carried out to investigate the functional role of the identified potential prognostic genes in TNBC. These genes are associated with MAPK, PI3-AkT, Wnt, TGF-ß, and other signal transduction pathways, pivotal in metastasis cascade. These gene signatures can provide novel molecular-level insights into metastasis.