We present a computational study of transcriptomic data of 6,000+ tissue samples of 14 cancer types,aiming to address the fundamental question: what may drive cancer cell division at a root level? Our analyses of the data point to that persistent disruption of the normal intracellular pH by altered cellular iron metabolism,specifically,increased Fenton reactions may represent a key stress that derives the disease evolution.Specifically,(1) repeated Fenton reactions in cytosol increase the cytosolic pH,to which the cells respond by generating net protons to maintain the pH stability through consuming glycolytic ATPs by continuous nucleotide synthesis and a few other processes;(2) repeated Fenton reactions in mitochondria increase the pH within the mitochondrial inner membrane,creating a proton gradient on the two sides of the membrane and hence driving mitochondrial ATP synthesis like what Andre Jagendorf demonstrated 50 years ago but through a novel pathway involving only part of the electron transport chain;(3) additionally,Fenton reactions also take place in the extracellular matrix and space,generating signals for cell cycle progression;and (4) simultaneous Fenton reactions in all these subcellular locations may represent a necessary condition for a cancer to take place as suggested by our analyses on cancer versus noncancerous inflammatory disease tissues.A model is developed to link all these to cancer initiation,under which oncogenic mutations are selected to keep pace with the cell-division rates dictated by the level of cytosolic Fenton reactions and the rate of nucleotide synthesis.Warburg effect and a number of other long-time open questions can be explained naturally using this model.