REFS platform will be used to identify genetic and molecular mechanisms of drug efficacy.
GNS Healthcare entered into a subcontract with SAIC-Frederick to support the latter’s prime contract with the NCI. GNS will analyze NCI data generated from the application of several well-known cancer treatments to the NCI-60 cancer cell line panel. The goal is to identify biomarkers and biological mechanisms that will lead to better matching of drugs to patients and new effective drugs in cancer.
The collaboration will utilize GNS’ supercomputer-driven REFS™ platform to build computer models to identify key genetic and molecular mechanisms of drug efficacy and resistance in cancer. REFS is composed of integrated machine learning algorithms and software that reportedly extract causal relationships from complex, multidimensional data and enable the simulation of billions of hypotheses.
The NCI-60 panel of human tumor cell lines was initially developed by NCI in the 1980s as an in vitro resource for drug discovery to replace the use of transplantable animal tumors in the screening of anticancer agents. Drugs to be used in GNS’ study include doxorubicin X2, Velcade® (bortezimib), paclitaxel, Sprycel® (dasatinib), Sutent® (sunitinib), and rapamycin.
“GNS is excited to be undertaking this radically new approach to unraveling mechanisms of cancer drugs that is complementary to the expert-driven but biased approaches that have been the standard in cancer research for decades,” says Iya Khalil, Ph.D., evp and co-founder of GNS.
In the first phase of the project GNS will utilize transcriptional profiling data previously collected by NCI from the application of the NCI-60 panel. GNS will utilize the REFS platform to reverse-engineer network models from the data that connect drug doses to transcriptional networks to endpoints.
The results from millions of in silico simulations of these models is expected to provide unique insights into the varied response that heterogeneous tumors exhibit to commonly used anticancer agents. GNS says that next it will build versions of the computer models that may be made available more broadly to cancer scientists for their own research via a web interface.