Multi-DDGNET model: exploring driver genes and mechanisms of Multiple disease states in the progression of NAFLD

王任琮 Ren-Cong Wang1, 莊佾軒 Yi-Shen Chung 2, 楊進木 Jinn-Moon Yang3

1, 3Department of Biological Science and Technology, National Yang Ming Chiao Tung University

3Institute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung University

The development of bioinformatics and omics data has provided opportunities to link genomics, disease, and clinical causality. However, identifying disease driver genes and mechanisms remains a challenge. Multi-time point disease data can be used to identify biomarkers and investigate disease progression mechanisms. Currently, gene-based methods have limitations, such as ignoring the interaction and correlation between genes. Pathway-based methods only induce the biological functions of the genome through statistical methods, such as gene enrichment analysis, and cannot evaluate the key driving pathways and their correlations.

Therefore, we propose a Multi-DDGNet (Multiple state disease driver gene network) model to simultaneously evaluate the biological pathway changes regulated by single genes and their communities at different time points, to identify the association of gene-pathway-disease progression, and to assess the importance of genes in disease progression through regulatory pathways. Using non-alcoholic fatty liver disease (NAFLD) and hepatitis as examples, where the disease driver genes and progression mechanisms remain unclear, we use the Multi-DDGNet model to investigate the biochemical pathways of healthy liver, non-alcoholic fatty liver (NAFL), and non-alcoholic steatohepatitis (NASH) progression and identify several biomarkers.

The analysis results and experiments show that the Multi-DDGNet model can help reveal the disease mechanisms of NAFLD progression, provide potential diagnostic biomarkers and drug targets.

Keyword: Multi-time, Pathway-based, Driver gene, NAFLD