Cancer metabolism is critical for understanding the mechanism of tumorigenesis, yet the understanding is still challenging. We studied gene-metabolism regulatory interactions and quantified the global driving forces for cancer-metabolism dynamics as the underlying landscape and probability flux. We uncovered four steady-state attractors: a normal state attractor, a cancer OXPHOS state attractor, a cancer glycolysis state attractor, and an intermediate cancer state attractor. We identified the key regulatory interactions through global sensitivity analysis based on the landscape topography. Different landscape topographies of glycolysis switch between normal cells and cancer cells were identified. We uncovered that the normal state to cancer state transformation is associated with the peaks of the probability flux and the thermodynamic dissipation, giving dynamical and thermodynamic origin of cancer formation. We found that cancer metabolism oscillations consume more energy to support cancer malignancy. This study provides a quantitative understanding of cancer metabolism and suggests a metabolic therapeutic strategy.

癌症新陈代谢对于理解肿瘤发生的机制至关重要，但是仍然难以理解。我们研究了基因代谢调节相互作用，并定量分析了癌症代谢动力学的全球驱动力，作为潜在的格局和概率通量。我们发现了四个稳态吸引子：正常状态吸引子，癌症OXPHOS状态吸引子，癌症糖酵解状态吸引子和中间癌症状态吸引子。我们通过基于地形的全球敏感性分析确定了关键的监管互动。确定了正常细胞与癌细胞之间糖酵解转换的不同景观地形。我们发现正常状态到癌症状态的转换与概率通量和热力学耗散的峰值相关，提供癌症形成的动力学和热力学起源。我们发现，癌症新陈代谢的振荡会消耗更多的能量来支持恶性肿瘤。这项研究提供了对癌症代谢的定量理解，并提出了一种代谢治疗策略。