Science Matters: Welcome To The Oncobiome

For the better part of the past decade, computational scientists have been using machine learning to “try to extract value and meaning from data sets that might otherwise be obscure,” in the words of Atul Butte, director of the University of California, San Francisco’s Bakar Computational Health Sciences Institute and chief data scientist for the University of California Health System. A group at the University of California, San Diego (UCSD) has now uncovered such an obscure – and valuable – data set from existing human cancer genomic studies, teasing out diagnostic information from microorganisms that reside inside cancer patients.

Taking sequencing data from samples deposited into The Cancer Genome Atlas (TCGA) project, the researchers showed that microbial signatures derived from sequencing reads from those samples correlate with the emergence...