A Story of Data-Mining Based Discovery

Dr. Ying Xu is a professor and Georgia Research Alliance Eminent Scholar of bioinformatics. This seminar talk was given on February 19, 2021 at the Institute of Bioinformatics, University of Georgia.

Is it possible to derive the driving forces and associated mechanisms of cancer formation and evolution from the cancer omics data? An evolutionary framework is needed to accomplish such ambitious goals. However, little can be borrowed and utilized from the existing cancer theories to guide such efforts since too many cancer-related questions cannot be addressed by them.

In this lecture, I will outline a “stress-adaptation” framework, through which the many seemingly very unusual and counter-intuitive behaviors of cancer can be interpreted as the adaptive steps to the specific stressors that such cells uniquely encounter. Knowing that the cost to become cancerous is very high, say, consuming 30 to 50-fold more glucose than their matching normal cells and over 95% of them die soon after they are created, we expect that the cost for staying unchanged would be even higher.

Our basic hypothesis is: cancer is a survival process under such (to-be-identified) stress, in which cells must divide as otherwise they will die, 100%. With this guiding hypothesis, we have discovered that many of the cancer biology questions have to be addressed at the basic chemistry (or physical) homeostasis level. Our data analyses and modeling have revealed that the affected cells must make fundamental changes in their cells’ definition, for them to overcome the persistently disrupted homeostasis, potentially the cancer-defining stressors, which are ultimately due to chronic inflammation and local iron accumulation. Specifically, they need to transform themselves from cells of a multi-cellular organism (i.e., a human being) to a unicellular “organism” to enable the key survival pathways, which is accomplished through selection of mutations in large numbers of genes of specific functions. A variety of cancer behaviors can be explained in terms of this transformation and associated reprogrammed metabolisms.

This is an extremely exciting and challenging puzzle-solving problem; and we welcome interested people to join our effort to develop, together, a fundamentally novel theory of cancer evolution.