Analyses of Cancer Atlas data gives classifications across tumor types

Analyses in which thousands of cancer specimens were reviewed suggest common traits exist across a dozen cancer tumor types, U.S. research suggests. Genomic alteration profiles were reported in two related papers published Thursday in Nature Genetics. The collaborative project provides new bioinformatics tools for facilitating larger collaborations, as well as reports findings that suggest commonalities across a range of tumors. In one study, Chris Sander of Memorial-Sloan Kettering in New York and colleagues classified 3,299 tumors from 12 cancer types from The Cancer Genome Atlas Pan-Cancer dataset. The authors used an algorithmic approach to separate tumors into those with primarily non-germline mutations and those with primarily copy-number alterations, in which the cell has an abnormal number of copies of a region of the genome. In the second study, Rameen Beroukhim of the Broad Institute in Boston and colleagues characterized non-germline copy number alterations in 11 cancer types and 4,934 primary cancer specimens from the same pan-cancer dataset, Nature said. They found whole-genome doubling in 37 percent of cancers, associated with higher rates of all non-germline copy number alterations. Nature said two accompanying commentaries reviewed the TCGA Pan-Cancer analysis project, as well as the software platform used to share data and collaborate on analysis and results.