**5. Molecular typing based on metabolic characteristics**

Cell metabolism is downstream of gene regulation and protein action network, reflecting the terminal information of life activities. The liver is the largest metabolic organ of the human body, and metabolic reprogramming undoubtedly plays an important role in the occurrence and development of liver cancer [30]. Multi-omics research results show that glycolysis and fatty acid metabolism are up-regulated in liver cancer tissues, while liver-specific metabolic pathways are down-regulated in liver cancer tissues, such as gluconeogenesis, detoxification, bile acid metabolism, and urea-ammonia metabolism [23]. The combined markers of glycine cholic acid and phenylpropionate tryptophan identified based on metabonomics technology can accurately diagnose liver cancer 1 year in advance [31]. The high heterogeneity of the liver cancer mutation spectrum and expression spectrum will inevitably lead to the heterogeneity of its metabolome level. By constructing a genome-scale metabolic network model, liver cancer can be divided into iHCC type 1 to 3. iHCC1 showed the highest fluxes in the metabolism of amino acids, cofactors and coenzymes, pyruvate, fatty acid oxidation, carnitine shuttle, steroids, TCA, and oxidative phosphorylation. iHCC2 exhibited specific features including lower fatty acid biosynthesis and high glutamine metabolism, and β-catenin–associated up-regulated fatty acid oxidation. Finally, iHCC3 tumors were associated with multiple features of malignant tumors, including hypoxic behavior, epithelial-to-mesenchymal transition, higher fluxes in fatty acid biosynthesis, and a strong Warburg effect [32]. Whether tumor metabolic reprogramming is the initiating factor of cancer or the accompanying result of cancer, there is still much controversy. Preliminary research results show that amino acid metabolism-related genes such as proline synthase PYCR1 play an important role in the occurrence and development of liver cancer [33].

#### **6. Conclusion**

In recent years, many breakthroughs have been made in the treatment of liver cancer. Following sorafenib, lenvatinib, regorafenib, cabozantinib and combination therapies centered on immune checkpoints have come out to promote the progress of liver cancer drug treatment. However, due to the high heterogeneity of liver cancer, the overall effectiveness of the above drugs is still limited. Accurate molecular classification of liver cancer not only contributes to the decision-making of individualized diagnosis and treatment of liver cancer, and personalized drug treatment, but also greatly deepens clinicians' understanding of the complexity and heterogeneity of liver cancer, so as to formulate a more accurate and effective treatment strategy. The new molecular typing system should be closely integrated with clinical-pathological information, which can not only reflect changes at the molecular level but also have guiding significance for clinical diagnosis and personalized treatment or predicting prognosis. The author believes that with the progress and development of multi-omics technology, single-cell technology, tumor molecular visualization technology, and medical artificial intelligence, the molecular classification of liver cancer will become closer and closer to the essence of tumor biological characteristics, and ultimately achieve disease precision treatment.
