On June 25, 2025, Google DeepMind introduced AlphaGenome, a powerful AI model designed to interpret the human genome with unprecedented accuracy. AlphaGenome predicts how DNA mutations affect gene regulation by processing up to 1 million DNA base pairs at single-letter resolution—a significant leap in both scale and granularity compared to previous models like Enformer.
This new model can forecast a broad range of biological phenomena: from where genes start and end, to RNA splicing patterns, chromatin accessibility, and protein binding. It does so across hundreds of cell types and tissues using data from leading genomic databases like ENCODE and GTEx. A key highlight is AlphaGenome’s ability to model RNA splice junctions, critical for understanding rare diseases such as spinal muscular atrophy and cystic fibrosis.
Unlike previous task-specific models, AlphaGenome offers a unified architecture, achieving state-of-the-art performance across 24 of 26 variant-effect benchmarks. It allows researchers to test the impact of a mutation across multiple modalities with a single API call, accelerating disease research, synthetic biology applications, and fundamental genetic insights.