Revolutionizing Cancer Genomic Medicine by AI and Supercomputer with Big Data

Abstract

We are running a cancer clinical sequence system based on whole genome/exome, RNA sequence and epigenome as research. When focused on hematology/oncology, it takes currently four days for a patient from signing informed consent (IC) to diagnosis. This process consists of IC, specimen collection, next-generation sequencer analysis, data analysis, interpretation/translation of mutations, determining the diagnosis combined with all pathological/clinical data and returning the result to the patient. Therapies are not only drugs but also hematopoietic stem cell transplantation. A pipeline Genomon for analyzing cancer genomes and RNA sequences by next-generation sequencers plays one of the key roles. It is running on the supercomputer system at Human Genome Center. The bottleneck of interpretation/translation was drastically resolved by employing IBM Watson for Genomics in harmony with our in-house human curation pipeline. We report how our system works as a conglomerate of oncologists, cancer biologists, bioinformatics experts augmented with Watson and Genomon.