DeepMind Unveils AlphaGenome AI to Predict Gene Mutations
AlphaGenome predicts effects of mutations in DNA sequences up to one million base pairs, used by 3,000 researchers from 160 countries via an API since June 2025.
- On Wednesday, Google DeepMind unveiled AlphaGenome, an AI tool to help unravel the human genome and potentially aid treatments, with the work detailed in a Nature study and earlier blog post.
- Because only around two percent codes for proteins, the human genome contains three billion letters, and the 'dark genome' comprises the remaining 98% of non-coding DNA with many disease-linked variants.
- AlphaGenome operates as a sequence-to-function model that predicts nucleotide changes' effects, analysing up to one million DNA letters and comparing mutated vs non-mutated sequences, trained on public human and mouse cell/tissue datasets.
- Since its launch seven months ago, nearly 3,000 scientists across 160 countries have started using AlphaGenome to study cancer and other diseases, with about 1 million API calls per day via DeepMind's free API.
- Experts praised the advance but warned of limits, with Ben Lehner calling AlphaGenome a 'breakthrough' yet DeepMind noting lower accuracy for long-range regulatory predictions beyond 100,000 letters, while Pushmeet Kohli urged, 'We hope researchers will extend it with more data.
76 Articles
76 Articles
Google DeepMind researchers have unveiled an artificial intelligence model called AlphaGenome, which can analyze a million DNA letters at once and predict 11 different biological processes, such as how DNA is packaged in a cell. The new tool will help understand how mutations even far from genes affect cell function and disease.
Google DeepMind Unveils New AI Tool To Understand DNA Mutations
Google DeepMind has revealed a new artificial intelligence system, AlphaGenome, which, the researchers say, could help scientists better interpret the human genome and understand how genetic changes affect biology.
Advancing regulatory variant effect prediction with AlphaGenome
Deep learning models that predict functional genomic measurements from DNA sequences are powerful tools for deciphering the genetic regulatory code. Existing methods involve a trade-off between input sequence length and prediction resolution, thereby limiting their modality scope and performance1–5. We present AlphaGenome, a unified DNA sequence model, which takes as input 1 Mb of DNA sequence and predicts thousands of functional genomic tracks up to single-base-pair resolution across diverse modalities. The modalities include gene expression, transcription initiation, chromatin accessibility, histone modifications, transcription factor binding, chromatin contact maps, splice site usage and splice junction coordinates and strength. Trained on human and mouse genomes, AlphaGenome matches or exceeds the strongest available external models in 25 of 26 evaluations of variant effect prediction. The ability of AlphaGenome to simultaneously score variant effects across all modalities accurately recapitulates the mechanisms of clinically relevant variants near the TAL1 oncogene6. To facilitate broader use, we provide tools for making genome track and variant effect predictions from sequence. AlphaGenome, a deep learning model that inputs 1-Mb DNA sequence to predict functional genomic tracks at single-base resolution across diverse modalities, outperforms existing models in variant effect prediction and enables comprehensive genomic analysis.
After developing AlphaFold, the Nobel-winning protein analysis AI in 2024, the DeepMind company presents AlphaGenome. At the disposal of scientists, this tool accelerates the interpretation of our DNA and predicts the impact of its variationsOur genome is a great manufacturing manual of a human being, made up of hundreds of forms of cells, from neurons to white blood cells, whose functions are as different as each other. Thanks to the Human Geno…
The vast majority of human genetics are riddles for researchers. Deepmind AI software can now better understand the function of DNA. Can major medical questions be solved with mutants in the computer in the future?
Coverage Details
Bias Distribution
- 52% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
