RNA 'Phenocopy' Signatures Provide Targeted Treatment Response Clues in Cancer

Investigators at the University of Wisconsin-Madison and other centers consider the RNA phenocopy signatures associated with cancer driver gene mutations, demonstrating that alternative alterations that prompt similar gene expression patterns may help to predict targeted treatment response patterns. Using machine learning methods, along with available gene expression profiles from the Cancer Genome Atlas project, the team characterized phenocopy mutation signatures across more than 9,200 tumor samples and nearly 2,000 cell lines, flagging expression patterns that resembled those found in samples with targetable mutations in eight known cancer genes, including EGFR, BRAF, and ERBB2. When they brought in expression, mutation, and drug response data from efforts such as the "Genomic of Drug Sensitivity in Cancer," "Cancer Cell Line Encyclopedia," and "Cancer Dependency Map," meanwhile, the authors found that phenocopy signatures could help to predict treatment responses — results they explored further with data from several clinical studies. "We found that these signatures improved our ability to predict response to targeted therapies compared to DNA mutations alone," they report in npj Genomic Medicine, adding that "these phenocopy signatures predict responses in clinical cohorts and shift under the selective pressure of treatment.