Google DeepMind Unveils AI Tool to Predict Harmful Genetic Mutations

A study published in the journal 'Science' revealed that the AI tool accurately predicted 89% of these mutations with a 90% precision rate. Each mutation was assigned a score indicating its potential to induce disease, termed 'pathogenic.'

Researchers at Google DeepMind, the tech giant's AI division, have unveiled an AI tool that predicts the potential harm of genetic mutations. This development claims to have the potential to significantly advance research into rare diseases.

The tool, named AlphaMissense, focuses on 'missense' mutations, which involve a single letter of the genetic code. The human genome typically contains around 9,000 of these mutations, some of which can be benign, while others can lead to diseases such as cystic fibrosis and cancer or affect brain development.

Despite the prevalence of these mutations, only 2% of the four million observed in humans have been classified as either disease-causing or harmless. There are approximately 71 million possible missense mutations in total.

AlphaMissense aims to address this knowledge gap. In a study published in the journal Science, it was revealed that the tool was able to predict 89% of these mutations with a 90% accuracy rate. Each mutation received a score indicating its potential to cause disease, referred to as pathogenic.

According to the results, 57% of the mutations were classified as probably benign, 32% as probably pathogenic, and the remainder remained uncertain. The tool's database has been made accessible to scientists, offering a valuable resource for further research.

Google DeepMind's AlphaMissense outperforms existing tools, as experts Joseph Marsh and Sarah Teichmann noted in an article also published in Science. While not intended for clinical diagnosis, the tool holds promise in increasing the rate of rare disease diagnoses and discovering new disease-causing genes. This, in turn, could lead to the development of new treatments.

The tool's training involved analysing the DNA of humans and closely related primates, enabling it to recognise widespread genetic mutations. Jun Cheng of Google DeepMind explained that the training process involved inputting 'millions of protein sequences' and teaching the tool to recognise regular protein sequences. It can then identify mutations and assess their potential harm.

Cheng drew an analogy to language learning, stating, "If we substitute a word from an English sentence, a person familiar with English can immediately see whether this word substitution will change the meaning of the sentence or not."

The introduction of AlphaMissense underscores the increasing role of AI in the natural sciences. As this tool becomes widely available to scientists, it may open new avenues for research into rare diseases, ultimately benefitting patients and advancing medical Science.

Google has also entered the insurance sector, utilising AI to streamline the process.

In a recent development, Google's Cloud unit developed a software tool called Claims Data Activator to assist health insurers in streamlining the prior authorisation process.

This innovative solution focuses on cleaning up data received from healthcare providers, regardless of its format, to expedite the review process. Leveraging various existing Google AI models, the software extracts information from documents and organises it in a standardised format, facilitating faster and more efficient human review.