COSMIC: exploring the world's knowledge of somatic mutations in human cancer

TLDR

COSMIC is the world’s largest and most comprehensive catalogue of somatic mutations in human cancer. COSMIC curates mutation data manually from the literature, annotates each mutation to the human genome and affected genes, and correlates these annotations across disease and mutation types to provide precise disease and patient definitions. The v70 release documents over 2 million coding point mutations from more than one million tumor samples, 12 000 curated genomes, and millions of noncoding alterations, fusions, rearrangements, copy‑number changes and expression variants, offering unprecedented depth and breadth of mutation–phenotype relationships and biomarker insights.

Abstract

COSMIC, the Catalogue Of Somatic Mutations In Cancer (http://cancer.sanger.ac.uk) is the world's largest and most comprehensive resource for exploring the impact of somatic mutations in human cancer. Our latest release (v70; Aug 2014) describes 2 002 811 coding point mutations in over one million tumor samples and across most human genes. To emphasize depth of knowledge on known cancer genes, mutation information is curated manually from the scientific literature, allowing very precise definitions of disease types and patient details. Combination of almost 20 000 published studies gives substantial resolution of how mutations and phenotypes relate in human cancer, providing insights into the stratification of mutations and biomarkers across cancer patient populations. Conversely, our curation of cancer genomes (over 12 000) emphasizes knowledge breadth, driving discovery of unrecognized cancer-driving hotspots and molecular targets. Our high-resolution curation approach is globally unique, giving substantial insight into molecular biomarkers in human oncology. In addition, COSMIC also details more than six million noncoding mutations, 10 534 gene fusions, 61 299 genome rearrangements, 695 504 abnormal copy number segments and 60 119 787 abnormal expression variants. All these types of somatic mutation are annotated to both the human genome and each affected coding gene, then correlated across disease and mutation types.