Meet AlphaMissense: Google's New AI Tool that Can Identify Protein Mutations Likely to Cause Disease
A groundbreaking new AI tool launched by Google DeepMind called AlphaMissense can accurately predict how protein mutations lead to genetic disorders. The tool analyzes a type of mutation called "missense mutations" which involves a single amino acid change in a protein.
AlphaMissense was created using DeepMind's powerful AI system AlphaFold which can predict the 3D structure of nearly all known human proteins. It analyzes over 216 million possible single amino acid changes across 19,233 human proteins. This allowed the tool to classify 89% of observed missense variants1, predicting 57% as probably benign and 32% as probably pathogenic.
The team published their findings in the journal Science. They believe AlphaMissense can accelerate disease research by helping scientists uncover previously unknown pathogenic mutations and disease genes. It may also aid in diagnosing rare genetic disorders by prioritizing variants for testing.
Missense variants are important for studying many complex diseases caused by multiple genetic factors. While AlphaMissense does not predict structure changes from mutations, it derives a pathogenicity score using related protein sequences and structural context. This gives it 90% accuracy in classification, an important step towards understanding the effects of mutations on human health.
What is a missense variant?
A missense variant is a single letter substitution in DNA that results in a different amino acid within a protein. If you think of DNA as a language, switching one letter can change a word and alter the meaning of a sentence altogether. In this case, a substitution changes which amino acid is translated, which can affect the function of a protein.