TopMarker: Computational Screening Biomarkers of Hepatocellular Carcinoma from Transcriptome and Interactome Based on Differential Network Topological Parameters

Yanqiu Wang,Tong Wang,Yi Cao,Xu Qiao,Xianhua Han,Zhi-Ping Liu

DOI: https://doi.org/10.1016/j.compbiolchem.2024.108166

IF: 3.737

2024-01-01

Computational Biology and Chemistry

Abstract:Identifying diagnostic biomarkers for cancer is crucial in the field of personalized medicine. The available transcriptome and interactome provide unprecedented opportunities and challenges for biomarker screening. From a systematic perspective, network-based medicine methods provide alternative approaches to organizing the available high-throughput omics data for deciphering molecular interactions and their associations with phenotypic states. In this work, we propose a bioinformatics strategy named TopMarker for discovering diagnostic biomarkers by comparing the network topology differences in control and disease samples. Specifically, we build up gene-gene interaction networks in the two states of control and disease respectively. The network rewiring status across the two networks results in differential network topologies reflecting dynamics and changes in normal samples when compared with those in disease. Thus, we identify the potential biomarker genes with differential network topological parameters between the control and disease gene networks. For a proof-of-concept study, we introduce the computational pipeline of biomarker discovery in hepatocellular carcinoma (HCC). We prove the effectiveness of the proposed TopMarker method using these candidate biomarkers in classifying HCC samples and validate its signature capability across numerous independent datasets. We also compare the discriminant power of biomarker genes identified by TopMarker with those identified by other baseline methods. The higher classification performances and functional implications indicate the advantages of our proposed method for discovering biomarkers from differential network topology.