Identification and validation of gene module associated with lung cancer through coexpression network analysis

• Weighted gene coexpression network analysis (WGCNA) was performed.
• A lung cancer-specific hub network successfully separated lung cancer from adjacent normal tissues.
• Fifteen candidate genes were used to stratify lung cancer patients and considered potential therapeutic targets.
• We tested and validated the robustness of our findings in six independent datasets.

Lung cancer, a tumor with heterogeneous biology, is influenced by a complex network of gene interactions. Therefore, elucidating the relationships between genes and lung cancer is critical to attain further knowledge on tumor biology. In this study, we performed weighted gene coexpression network analysis to investigate the roles of gene networks in lung cancer regulation. Gene coexpression relationships were explored in 58 samples with tumorous and matched non-tumorous lungs, and six gene modules were identified on the basis of gene coexpression patterns. The overall expression of one module was significantly higher in the normal group than in the lung cancer group. This finding was validated across six datasets (all p values < 0.01). The particular module was highly enriched for genes belonging to the biological Gene Ontology category “response to wounding” (adjusted p value = 4.28 × 10− 10). A lung cancer-specific hub network (LCHN) consisting of 15 genes was also derived from this module. A support vector machine based on classification model robustly separated lung cancer from adjacent normal tissues in the validation datasets (accuracy ranged from 91.7% to 98.5%) by using the LCHN gene signatures as predictors. Eight genes in the LCHN are associated with lung cancer. Overall, we identified a gene module associated with lung cancer, as well as an LCHN consisting of hub genes that may be candidate biomarkers and therapeutic targets for lung cancer. This integrated analysis of lung cancer transcriptome provides an alternative strategy for identification of potential oncogenic drivers.