13.8 C

Google tool uses AI to predict risk of genetic mutations



Researchers at Google DeepMind, the tech giant’s artificial intelligence (AI) division, on Tuesday unveiled a tool that predicts whether a genetic mutation will cause damage. This is an advance that could help research into rare diseases.

Pushmeet Kohli, vice president of research at Google DeepMind, said in a press conference that the results “take us one step further in recognizing the impact that AI is having on the natural sciences.”

The tool, called AlphaMissense, focuses on so-called missense mutations that affect a single letter of the genetic code.

The average human genome has 9,000 such mutations. They can be harmless or cause diseases such as cystic fibrosis, cancer, or damage brain development.

To date, 4 million of these changes have been observed in humans, but only 2% have been classified as disease-causing or benign.

There are a total of 71 million possibilities for this type of mutation. Google DeepMind tool reviewed them and with 90% accuracy he was able to predict 89%.

Each was assigned a score to indicate its risk of causing disease.

As a result, 57% were classified as probably benign, 32% as probably pathogenic, and the remainder as unknown.

The database is now open and available to scientists. Additionally, this Tuesday, a study on this issue was published in the journal Science.

AlphaMissense shows “better performance” than previously available tools, experts Joseph Marsh and Sarah Teichman said in a similar article published in Science.

Google DeepMind’s Jun Cheng admits that these predictions were “not really intended to be used only for clinical diagnosis.”

“However, we believe that our predictions could help increase diagnostic rates for rare diseases. They could also potentially help discover new disease-causing genes.” I believe that,” he added.

Researchers say this could indirectly lead to the development of new treatments.

The tool was trained using DNA from humans and related primates, allowing it to recognize the widest range of genetic variations.

Chen compared the process to learning a language. “When you replace a word in an English sentence, someone familiar with English can immediately tell whether the word replacement changes the meaning.”

Source: Diario.Elmundo

Subscribe to our magazine


━ more like this


Please enter your comment!
Please enter your name here