AlphaGenome unifies multimodal prediction, megabase-scale sequence context, and base-pair resolution in a single deep learning framework that processes 1 Mb DNA sequences to predict thousands of functional genomic tracks across 11 modalities, achieving state-of-the-art performance on 24 of 26 variant effect prediction benchmarks including superior prediction of splicing disruption, gene expression changes, and chromatin accessibility alterations compared to specialized single-modality models.

Deep learning models that predict functional genomic measurements from DNA sequence are powerful tools for deciphering the genetic regulatory code. Existing methods trade off between input sequence length and prediction resolution, thereby limiting their modality scope and performance. We present AlphaGenome, which takes as input 1 megabase of DNA sequence and predicts thousands of functional genomic tracks up to single base pair resolution across diverse modalities – including gene expression, transcription initiation, chromatin accessibility, histone modifications, transcription factor binding, chro- matin contact maps, splice site usage, and splice junction coordinates and strength. Trained on human and mouse genomes, AlphaGenome matches or exceeds the strongest respective available external models on 24 out of 26 evaluations on variant effect prediction. AlphaGenome’s ability to simultaneously score variant effects across all modalities accurately recapitulates the mechanisms of clinically-relevant variants near the TAL1 oncogene. To facilitate broader use, we provide tools for making genome track and variant effect predictions from sequence.