Google’s AI can Now Predict Genetic Diseases

There are 71 million possible missense variants in the human genome.

Breaking this down??

There are approximately 71 million different ways in which a single nucleotide change (mutation) in the DNA sequence could wreak havoc on a person’s biology.

An average person carries more than 9,000 of them.

Most are harmless, but some have been implicated in genetic diseases such as sickle cell anemia and cystic fibrosis, as well as more complex conditions like type 2 diabetes, which may be caused by a combination of small genetic changes.

How do we know which ones are actually dangerous?

The answer: Well largely, we don’t.

Of the 4 million missense variants that have been spotted in humans, only 2 percent have been categorized as either pathogenic or benign, through years of painstaking and expensive research. It can take months to study the effect of a single missense variant.

Google DeepMind has released a tool that can rapidly accelerate that process.

AlphaMissense is a machine learning model that can analyze missense variants and predict the likelihood of them causing disease with 90 percent accuracy—better than existing tools.

How Does it Work?

It's built upon DeepMind's AlphaFold, a model renowned for predicting the structures of countless proteins from their amino acid compositions. However, AlphaMissense takes a different approach. Instead of predicting protein structures, it operates more like a large language model, such as OpenAI's ChatGPT.

AlphaMissense has been trained on the language of human and primate biology, giving it a deep understanding of what normal sequences of amino acids in proteins should look like. When faced with a sequence that deviates from the norm, it can identify it, much like spotting an out-of-place word in a sentence.

Each of the 71 million possible missense variants is assigned a "pathogenicity score" ranging from 0 to 1. This score is based on AlphaMissense's knowledge of the effects of closely related mutations. The higher the score, the more likely a specific mutation is to cause or be associated with a disease.

For researchers seeking to unravel the mysteries of disease-causing missense variants, AlphaMissense provides a valuable resource.

Here’s the best part: DeepMind has made AlphaMissense's predictions freely available to the scientific community. Together with EMBL-EBI, they are also making these predictions more accessible for researchers through the Ensembl Variant Effect Predictor.

In addition to our look-up table of missense mutations, we’ve shared the expanded predictions of all possible 216 million single amino acid sequence substitutions across more than 19,000 human proteins.

DeepMind researchers worked with Genomics England, a government body that studies the growing pool of genetic data collected by the UK’s National Health Service, to verify the model’s predictions against real-world studies on already-known missense variants. The paper claims 90 percent accuracy for AlphaMissense, with 89 percent of variants classified.

--

Hello and welcome to Biotech Bulletin - your go-to source for the latest updates and comprehensive insights in biotech and genomics.

Enjoying the newsletter?

? Subscribe now to become part of 1K+ readers here on LinkedIn.

?? Please help us spread the word by sharing it with your colleagues and friends.

?? Also, consider?joining our Substack community?(2K+ industry professionals reading it via email).

??Until next time,

Cambrian Bioworks.