Work

Integrative machine learning reveals hidden and emerging co-regulatory gene networks for multi-phase glioblastoma outcome prediction

Year: 2025

Type: article

Abstract: BACKGROUND: Glioblastoma (GBM) is a highly prevalent and aggressive type of brain tumor characterized by profound molecular complexity and poor prognosis. While conventional biomarker studies focus on highly significant genes or proteins associated with cancer outcomes, the contribution of gene-gene co-regulation to GBM progression remains unclear. METHODS: This study employs an application of integrative machine learning approach, utilizing a high-dimensional transcriptomic profile that considers gene-gene co-regulations to identify predictive gene networks involved in GBM occurrence and 1-year survival prediction. We further integrate these network models with both empirical protein-protein interaction (PPI) data and random walk-based information flow analysis across the PPI landscape. RESULTS: This dual-layered approach uncovers gene modules that bridge transcriptional co-regulation with functional connectivity at the protein level. This integration highlighted several hub genes, including both strong and weak (e.g. BSN, RHOC, ANXA1, CSF1R, and ITGAM), that emerged as key molecular connectors involved in critical GBM processes such as immune response and neuronal signaling. Notably, these hub genes also exhibited cross-disease associations with traits including gut microbiome composition, type 2 diabetes, coronary artery disease, and other cancers, underscoring their systemic biological relevance. CONCLUSION: Overall, our findings through the computational approach underscore the significance of co-regulatory gene networks in GBM biology. It also demonstrates how integrating transcriptomic and protein-level interactions can refine prognostic modeling, advance biomarker discovery, and inform future therapeutic development.

Source: European Journal of Cancer

DOI: 10.1016/j.ejca.2025.116197

Authors: Md Tamzid Islam, Fengwei Yang, Stephan Komladzei, Murshalina Akhter, Mihaela E. Sardiu, Yanming Li

Institutions: University of Kansas Medical Center, The University of Kansas Cancer Center

Language: English

Topic: Bioinformatics and Genomic Networks

Subfield: Molecular Biology

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

SDG: Good health and well-being

Open Access status: hybrid

Estimated APC paid: $4,270