Google DeepMind has launched AlphaGenome, a revolutionary AI model that predicts how DNA sequences regulate genes and how mutations affect biological processes. Released on June 25, 2025, this breakthrough system can analyze DNA sequences up to one million letters long and predict thousands of molecular properties with unprecedented accuracy.
AlphaGenome represents a significant advancement over previous genomic analysis tools by unifying long-range context analysis with base-level precision. The model predicts where genes start and end in different cell types, how they're spliced, and the amount of RNA produced. It can also evaluate how single-letter mutations might impact these processes, making it valuable for understanding disease mechanisms.
"It's a milestone for the field. For the first time, we have a single model that unifies long-range context, base-level precision and state-of-the-art performance across a whole spectrum of genomic tasks," said a DeepMind researcher. The system outperformed specialized models in 22 out of 24 genomic prediction benchmarks.
AlphaGenome joins other remarkable AI innovations announced this month. Chinese researchers have developed mosquito-sized drones measuring just 2 centimeters long for reconnaissance missions. Meanwhile, factory robots equipped with embodied large language models can now complete complex tasks in unpredictable environments through autonomous learning.
In healthcare, AI continues to save lives beyond genomics. A Chinese research team has created a deep learning model that predicts postpartum hemorrhage with 90% accuracy by analyzing placental MRI scans. This advancement could significantly reduce maternal mortality, as hemorrhage accounts for one in four maternal deaths worldwide.
Google DeepMind is making AlphaGenome available via API for non-commercial research, with plans to release the full model in the future. The technology could accelerate disease research, guide synthetic DNA design, and deepen our understanding of the genome's functional elements.