Computational Method Offers Significant Boost in Finding New Cancer Targets
Article Jul 13, 2016
A team from Massachusetts Institute of Technology and ARIAD Pharmaceuticals, USA, have developed a computational methodology to predict significant overlap between variants in inherited disease and existing cancer genetic databases. So far they have added new cancer associated genes, substantially expanded the therapeutically actionable population of patients harbouring activating mutations in ACVR1, and identified the first evidence of significant mutations in CDK4 in lung cancer and in SOS1 in melanoma.
Current identification of driver cancer genes from cohort studies is limited by statistical resolution. Published last month in PLOS Genetics (Zhao & Pritchard (2016) Inherited Disease Genetics Improves the Identification of Cancer-Associated Genes.PLoS Genet 12(6): e1006081.doi:10.1371/journal.pgen.1006081), this new study demonstrates how anyone with a database of amino acid variants from diverse methodologies (e.g., mutagenesis screens or biochemical analyses) can investigate cancer associations further and add statistical rigour to findings.
The team merged data from the pathogenic entries in the UniProt HUMSAVAR database and the entire set of tumour exomes featured on the cBioPortal and applied a computational model with a match score (per gene per study) to make comparisons between datasets with an additional boot strap approach to assess the confidence in the score, by deriving a signal-to-noise ratio.
This paper demonstrates how intelligent new methods such as this can boost development of new cancer therapies by making orphan diseases a little more common and identifying larger unmet medical needs.
Justin Pritchard commented “Of direct and timely biological interest is the first identification of ACVR1 mutations in adult cancers (specifically, uterine cancers). Kinase activating ACVR1 mutations are responsible for the extraordinarily rare and devastating Fibrodysplasia Ossificans Progressiva (FOP) also known as “stone man syndrome” for the manner in which it locks people in an ectopic skeleton. Last year activating mutations in the same gene were found in extraordinarily rare pediatric gliomas and reported in 4 simultaneous papers in Nature Genetics and profiled in Cancer Research.”
Navigating Drug Discovery with High-Throughput ScreeningArticle
Once the preserve of big pharma, high-throughput screening and the data it has produced is now widely available. Together with advances in artificial intelligence, it is providing broad opportunities in drug discovery.READ MORE
Treating Cancer: Tackling Drug ResistanceArticle
One of the biggest barriers to treating cancer is drug resistance, a phenomenon that remains only partially understood. While drug resistance has been a problem in cancer for years, new treatments are coming to light that could help overcome it...READ MORE
qPortal Offers a Data-Driven Approach to Biomedical ResearchArticle
We discuss the recently released data management platform qPortal with one of its creators, Christopher Mohr.READ MORE