A*STAR researchers at the Genome Institute of Singapore (GIS) have created the Variant Network (VarNet), an innovative artificial intelligence (AI)-based technique that can examine and find cancer mutations (variants) among the millions of DNA fragments found in a tumor sample.
3D illustration of a method of DNA sequencing. Image Credit: Genome Institute of Singapore
It will act as a crucial compass in directing individualized treatment plans in the battle against cancer. VarNet can be used to evaluate mutations to customize treatment plans or comprehend cancer in both clinical and research settings. On July 22nd, 2022, the study was released in Nature Communications.
Mutations picked up throughout the course of a person’s lifespan are something that causes the hereditary condition known as cancer. To create individualized treatment plans—providing the correct treatment to the right patient at the right time—identifying these mutations has long been a problem that needs to be overcome. To deal with that problem, this research was created.
Without the need for specialist skills in cancer or genomics, VarNet employs deep learning, an AI technique, to find cancer mutations. Vast volumes of cancer sequencing data from Singaporean and foreign databases were used to train VarNet.
VarNet frequently outperforms current mutation detection algorithms in terms of precision when tested against real tumor benchmarks. Upstream analyses that could alter research results and clinical treatment choices are impacted by the correct identification of mutations in tumors.
We have been working on machine learning methods for some time to improve detection of cancer mutations. During this work, we learned that human experts were often involved in the process to validate selected high-confidence cancer mutations.
Dr Anders Skanderup, Study Corresponding Author and Group Leader, Laboratory of Computational Cancer Genomics
“Such human experts make decisions by inspecting images of DNA reads overlapping the potential mutations,” Dr Skanderup says.
However, while a human can only do this for a couple of mutations in a limited amount of time, an AI approach could potentially perform the same task across the entire 3 billion nucleotides in the human genome. This inspired us to leverage deep learning approaches that learn patterns in images, and develop a pure AI-based method for identifying mutations in cancer.
Dr Anders Skanderup, Study Corresponding Author and Group Leader, Laboratory of Computational Cancer Genomics
The first author of the study and Ph.D. candidate Kiran Krishnamachari stated that the system learned to identify mutations from the raw data in a way that a professional would do when manually examining prospective mutations. Krishnamachari is also an A*STAR Computing and Information Science scholar affiliated with GIS.
He stated, “This gave us the confidence that the system can learn relevant mutational features when trained on vast sequencing datasets, using our weak-supervision strategy that does not require excessive manual labeling.”
GIS Executive Director, Professor Patrick Tan commented, “Identifying cancer mutation is a critical step in developing precision medicine. VarNet demonstrates that deep machine learning can detect cancer mutations with an accuracy often exceeding existing state-of-the-art methods.”
Journal Reference
Krishnamachari, K., et al. (2022) Accurate somatic variant detection using weakly supervised deep learning. Nature Communications. https://www.nature.com/articles/s41467-022-31765-8