Multi-Tensor Decompositions for Personalized Cancer Diagnostics, Prognostics, and Therapeutics

Опубликовано 31 января 2020, 16:26

Join this webinar to learn about the development of novel, multi-tensor generalizations of the singular value decomposition and their use in the comparisons of tumor and normal genomes of cancer patients. The talk will describe examples from brain, lung, ovarian, and uterine cancers. By using the complex structure of the datasets, rather than simplifying them as is commonly done, these mathematical frameworks enable the separation of genome-scale patterns of DNA copy-number alterations—which occur only in the tumor genomes—from those that occur in the genomes of normal cells in the body and variations caused by experimental inconsistencies. These patterns predict survival and response to treatment statistically better than, and independent of, the best indicators in clinical use and existing laboratory tests. The talk will also describe a recent, retrospective clinical trial that validates the brain cancer pattern. Researchers have recognized recurring alterations as a hallmark of cancer for over a century and observed them in these cancers’ genomes for decades. However, nobody succeeded in identifying copy-number subtypes predictive of patients’ outcomes before this work. The data had been publicly available, but the patterns remained unknown until the data were modeled by using the multi-tensor decompositions. This demonstrates that the decompositions underlie a mathematically universal description of the genotype-phenotype relationships in cancer that other machine learning methods miss.

Presenter:

Dr. Orly Alter, Utah Science, Technology, and Research (USTAR) Associate Professor of Bioengineering and Human Genetics, the Scientific Computing and Imaging Institute and the Huntsman Cancer Institute, University of Utah