Machine learning modeling and analysis of prognostic hub genes in cervical adenocarcinoma: a multi target therapy for enhancement in immunosurveillance

TL;DR

This paper uses machine learning to find key genes and drugs for treating endocervical adenocarcinoma, a deadly cervical cancer subtype.

Contribution

The study introduces novel biomarkers and drug targets for endocervical adenocarcinoma using in silico and in vitro approaches.

Findings

11,592 differentially expressed genes were identified, enriched in metabolic and signaling pathways.

Imatinib showed high binding affinity to BIRC5, confirmed by cell viability experiments.

Potential therapeutic targets and biomarkers were proposed for endocervical carcinogenesis.

Abstract

Endocervical adenocarcinoma (ECA) the fatal and intrusive subtype of cervical carcinoma is on rise from the last decade. Its improper detection leads to worst clinical outcomes that urges the discovery of novel biomarkers. Therefore, we proposed insilico and invitro based approches to identify key genes that could be used as potential targeted therapies. RNA-seq and gene expression data was operated via R-programming that identified 11,592 differential expressed genes which are mainly enriched in metabolic pathways, chemical carcinogenesis-receptor activation, amoebias, MAPK and PI3K-AKT signaling pathway. Clustering modules and hub genes were retrieved to design network of immune cells with varying expression using multiple statistical algorithms. The Drugs targeting hub genes were determined from Drug gene interaction database which was further categorized for docking and dynamics based simulations. Results indicate high binding affinity of Imatinib compound into active pockets of BIRC5 which is confirmed by cell viability lab experiment. Current study demonstrates novel biomarkers and therapeutic drugs for in depth understanding of endocervical carcinogensis.

The online version contains supplementary material available at 10.1007/s12672-025-02834-3.

Figures