Machine learning is enhancing how to identify cancer mutations affecting precision medicine treatments
Researchers from the Institute for Research in Biomedicine (IRB) in Barcelona, Spain, have employed the technology to solve the problem of distinguishing the risks of cancer mutations.
Precision medicine depends on recognizing genetic or physiological differences between patients and tailoring treatment that is individualized to these differences. While precision medicine is being applied in all areas of medicine, cancer treatment has been a particularly relevant field due to its genetic roots.
Mutations Present an Obstacle for Oncology
A significant challenge to oncological precision medicine is that cancers mutate frequently. Some of these mutations are clinically significant and indicate specific personalized treatment is necessary, while other mutations are benign, random mutations that do not have any clinical effect.
“We started from the premise that we only get to observe some mutations because the tumor cells with this mutation guide the development of the tumor, and we questioned what distinguishes these mutations from other possible mutations,” Ferran Muiños, PhD, one of the primary researchers involved in this project, said in a news release. “Doing this analysis manually would be excessively laborious, but there are computational strategies that allow it to be organized systematically and efficiently.”
To help identify which cancer mutations are clinically relevant, researchers at IRB Barcelona developed a computational tool founded on the biological principle of natural selection. Led by Núria López-Bigas, PhD, head of the Biomedical Genomics Lab at IRB Barcelona, researchers created a machine learning model that analyzed cancer mutations and identified those that enhanced the survival and growth of cancer cells.
Machine learning is a subset of artificial intelligence technology.
This tool, dubbed BoostDM, builds on previous tools that the team had developed to identify genes that were responsible for the spread of cancer cells.
“BoostDM goes further: It simulates each possible mutation within each gene for a specific type of cancer and indicates which ones are key in the cancer process,” López-Bigas explained. “This information helps us to understand how a tumor is caused at the molecular level, and it can facilitate medical decisions regarding the most appropriate therapy for a patient.”
Digital Tools to Help Personalized Cancer Treatments
The research behind the creation of BoostDM has been published in the journal, Nature, and this tool has been added into the existing resources that the Biomedical Genomics Lab team has created.
These resources include a research database, called IntOGen, and a clinical decision-making tool called the Cancer Genome Interpreter that is designed specifically for oncological personalized medicine applications.
The BoostDM machine learning algorithm will be of particular interest to hospital and oncology groups practicing precision medicine because of its availability through the Cancer Genome Interpreter. This integration will empower clinicians to not only better recognize clinically relevant cancer mutations but also understand when an existing mutation is not a clinical concern.
As precision medicine continues to advance, especially in the area of oncology, understanding which cancer mutations matter while simultaneously being able to ignore benign mutations will be an important consideration for clinical providers. While much attention has been given to recognizing pathological mutations, concurrently identifying irrelevant mutations will enable hospitals and oncologists to provide more precise personalized cancer treatments.
—Caleb Williams
Related Information:
Institute for Research in Biomedicine (IRB Barcelona)
Machine learning fuels personalized cancer medicine