Scientists at St. Jude Children’s Research Hospital are using a type of machine learning to put a new twist on an established technique. The researchers created MethylationToActivity (M2A), a framework for using DNA methylation data to reveal promoter activity and gene expression. The results were published online today in Genome Biology.
“My interest is in whether we can use deep learning to see how methylation changes are leading to promoter activity changes and gene expression changes for different types of tumors,” said corresponding author Xiang Chen, Ph.D., of St. Jude Computational Biology. “What we have done with M2A is create a method for integrating DNA methylation information around promoters to make it more readily interpretable.”
Deep learning is part of a broader family of machine learning methods based on creating artificial neural networks. DNA methylation data are typically used to provide biomarkers for tumor detection, subtyping and classification but haven’t been reliable for individual gene expression and promoter activity. Results from M2A infer promoter activities based on enrichment of two histone modifications: H3K4me3 and H3K27ac.
The researchers have rigorously tested their approach on a variety of adult and pediatric tumors, including leukemias and solid tumors. They used data from St. Jude sources such as the Pediatric Cancer Genome Project as well as public databases such as ENCODE. Contrary to the current belief that DNA methylation data are not reliable for tracking gene expression, the researchers successfully decoded complex DNA methylation information that reveals histone modifications and thus promoter and gene expression changes.
Leveraging cloud capability
The information obtained through M2A is comparable to that gained from the gold-standard technique for assessing promoter activity: ChIP-Seq. However, M2A has the added advantage of using DNA methylation data obtained from samples that aren’t candidates for ChIP-Seq due to cost or sample quality.
“It comes down to how to achieve functional interpretation of DNA methylation data, which up until now hasn’t been possible, at least to the level achieved by M2A,” said first author Justin Williams, Ph.D., of St. Jude Tumor Cell Biology. “If you’ve had to choose between performing DNA methylation or doing a ChIP-Seq experiment, now you could essentially get the results of both with just DNA methylation.”
Using M2A, individual researchers can get an accurate promoter activity inference via the DNA methylome in about 15 minutes. This is because the M2A framework is available to scientists worldwide through St. Jude Cloud. Through this platform, researchers can upload their DNA methylation data and receive a profile based on the histone modification of their choice.
The study’s other authors are Beisi Xu, Daniel Putnam, Andrew Thrasher, Chunliang Li and Jun Yang, all of St. Jude.
The research at St. Jude was funded in part by grants from the National Institutes of Health (P30CA021765) and ALSAC, the fundraising and awareness organization of St. Jude.