Metabolic reprogramming is a well-recognized hallmark of cancer, characterized by its remarkable flexibility in activating alternative pathways in the absence of specific regulators or substrates. Non-negative Matrix Factorization (NMF) was employed to identify tumor heterogeneity at the terms of metabolism, progression-related pathways and molecular subtypes of HCC based on gene set variation analysis (GSVA) and stratified survival analysis. The inherent heterogeneities of metabolism landscape which includes genomes, methelomic and transcriptomic data, proteomes, phosphoproteomes, mutational and immune microenvironment landscape between metabolic subtypes were performed based on multi-omics analyses. NMF has led three distinct survival-associated subtypes (NMF cluster 1, iHCC1; NMF cluster 2, iHCC2; NMF cluster 3, iHCC3) based on metabolic gene expression, and GSVA revealed remakeable metabolic differences. Additionally, multi-omics analysis further revealed the unique landscape of different metabolic subtypes, encompassing transcriptome, epigenetic and post-transcriptional modifications (PTMs) at both bulk and single cell seq-RNA level. Furthermore, 39 subtype-specific variables for identifying metabolic subtypes were screened using four feature selection algorithms and preliminarily validated on 8 machine learning models. We then built and verified a nomogram to guide the individualized strategy for HCC patients, utilizing a combination of metabolic signatures and clinical characteristics. Finally, we preliminarily identified the potential contribution of Aldolase B (ALDOB) in metabolic reprogramming triggered by epigenetic and PTMs. Overall, the research defined robust subtypes and further revealed potential targets linking metabolism with immune microenvironment and non-mutational epigenetic modifications, thereby advancing our understanding of metabolic heterogeneity for application in HCC diagnosis and clinical risk stratification.
Li et al. (Wed,) studied this question.