Over the decades, there have been many advances in developing treatments for different types of cancer, specifically in the field of cancer genomics. These sophisticated tools identify the driver mutations that cause the disease.
The epigenetic process of methylation usually regulates gene activity through alteration of the DNA structure without changing the genetic information. However, sometimes excessive methylation, known as hypermethylation, can occur next to a tumor suppressor gene (TSGs) and trigger cancer through the inactivation of these genes.
A new machine learning technique for detecting other modifications to DNA that has a similar effect has been developed by researchers at Weill Cornell Medicine, New York-Presbyterian, and the New York Genome Center (NYGC).
Read the original publication of this study here: [Discovery of candidate DNA methylation cancer driver genes]
Learn more about how machine learning techniques are detecting genetic mutations that drive cancer.
Discovery of Candidate DNA Methylation Cancer Driver Genes
Previously, researchers were challenged to distinguish the driver mutations that cause cancer from the “passenger” mutations that don’t affect cancer. Now, more sophisticated techniques have been developed that distinguish the two.
Researchers focused on the DNA methylation changes detected in tumor cells and that infer which ones are likely to drive tumor growth.
MethSig, a new algorithm, was developed by the team that looks at background methylation rates in a particular area of the genome. It estimates when methylation is likely to be a cancer driver.
The statistical hypothesizing framework accounts for the changing stochastic hypermethylation rates across the genome and between samples. MethSig estimates expected background DNAme changes, thereby allowing the identification of epigenetically disrupted loci, where observed hypermethylation significantly exceeds expectation, potentially reflecting positive selection.
The algorithm was applied to different tumor type DNA methylation maps, and it found a small number of cancer-driven events in each tumor. The patterns were consistent across patients and tumor types and beat other existing methods of prediction.
The team further validated the research by looking at the affected gene in chronic lymphocytic leukemia (CLL) cells and detected likely cancer-driving methylation changes much more sensitively and selectively than current methods.
Being able to map the epigenetic changes that contribute to tumor growth in certain cancers enables scientists to understand cancer origins better and thus optimize treatments for individual cases. If researchers can map the entire landscape of DNA methylation of cancer drivers and for different tumors, they can also expand the realms of treatment beyond genetics to include the critical dimension of epigenetic changes in cancer.
Takeaways:
- Researchers at Weill Cornell Medicine, New York-Presbyterian, and the New York Genome Center (NYGC) developed a new machine learning technique for detecting DNA methylation changes that drive cancer.
- The classifier developed using MethSig produced estimated risks for each patient, and researchers found that patients with higher estimated risks were more likely to have had worse outcomes
- The algorithm has great potential for improving cancer prognosis and treatment.