Scientists have long struggled to find the exact genetic mutations that cause tumors to grow. Now, a new tool from Google is helping change that. It uses advanced artificial intelligence to spot cancer-related mutations in tumor DNA more accurately than ever before.

Understanding the Challenge of Detecting Mutations

One of the biggest hurdles in cancer research is identifying somatic variants. These are genetic changes that happen after birth and can drive cancer growth. Unlike inherited mutations, somatic variants can be caused by environmental factors like UV rays or random DNA errors during cell division. Detecting these mutations is tricky because they often exist at very low levels within tumors, sometimes even below the accuracy of current sequencing methods.

Genome sequencing can reveal many genetic differences in cancer cells, but distinguishing real mutations from errors introduced during sequencing has been a major challenge. This difficulty has limited researchers’ ability to pinpoint the exact mutations responsible for different cancers, slowing down the development of targeted treatments.

How Google’s DeepSomatic AI Works

DeepSomatic is an AI-powered tool designed to solve this problem. It analyzes both tumor samples and normal cells from the same patient to identify genuine cancer-driving mutations. The process begins by converting raw genetic data into images that represent various data points, making it easier for the AI to interpret complex information.

A convolutional neural network then examines these images to distinguish true somatic mutations from sequencing errors and inherited variants. This allows the system to produce a list of mutations likely responsible for the tumor’s growth. The ability to analyze both tumor-only and paired samples makes DeepSomatic versatile for different research and clinical settings.

The Impact on Cancer Research and Treatment

DeepSomatic’s high accuracy is partly thanks to a large training dataset called CASTLE. Created by Google and partners at UC Santa Cruz and the National Cancer Institute, this dataset includes genetic information from tumor and normal cells in several breast and lung cancer samples. Such high-quality data helps the AI learn to recognize true mutations more effectively.

This technology represents a significant step forward in understanding cancer’s genetic drivers. With more precise mutation detection, doctors can develop personalized treatment plans targeting the specific mutations present in each patient’s tumor. The open access to DeepSomatic and its training data encourages researchers worldwide to use AI to further advance cancer research and improve therapies.

As scientists continue refining DeepSomatic, the hope is that future versions will identify mutations even more accurately. This progress could lead to better, more targeted treatments and ultimately improve outcomes for cancer patients. The breakthrough shows how AI is transforming cancer research and opening new paths to fight the disease more effectively.

Inspired by

• https://www.artificialintelligence-news.com/news/google-ai-tool-pinpoints-genetic-drivers-of-cancer/

Sources

• nature.com
• cybersecuritycloudexpo.com