Decoding Cervical Cancer Biomarkers: An Integrated Framework of Bioinformatics, Machine Learning, and Experimental Confirmation

Cervical cancer is the fourth most frequent cancer in females, with a high mortality rate globally. Persistent infection with high-risk, oncogenic human papillomavirus (HPV) types is a critical etiologic factor in the progression of the disease. Unfortunately, cervical cancer often remains undiagnosed until advanced stages, hence limiting treatment effectiveness. Therefore, identifying precise and significant biomarkers is crucial. High-throughput sequencing technologies have revolutionized targeted cancer therapy research by generating extensive data for analysis. This study employed bioinformatics and machine learning (ML) approaches to identify dysregulated genes with significant diagnostic value in cervical cancer, utilizing transcriptomics datasets. Seven potential diagnostic biomarker genes (<i>APOD</i>, <i>SPARCL1</i>, <i>AR</i>, <i>MCM2</i>, <i>NUSAP1</i>, <i>PLK1</i>, and <i>STIL</i>) were validated by a real-time polymerase chain reaction (RT-PCR) experiment. The ML models were developed using significantly differentially expressed genes (DEGs) involved in important pathways for cervical cancer. ML prediction models are available at https://github.com/PGlab-NIPER/CC_Pred.