Google DeepMind researchers have unveiled a new AI model designed to unravel how changes in our DNA affect different cells, giving scientists a more detailed look at a notoriously difficult problem that could hold important clues to the causes of diseases and point the way to new treatments.
AlphaGenome 2 is the successor to DeepMind’s AlphaGenome genome prediction AI model, which launched last year. While almost every cell in the human body carries the same DNA, different types of cells—from heart muscle cells and skin cells to immune cells that fight infection—turn different genes on and off, and those patterns can change depending on what a cell is doing. As a result, the same genetic variant can have very different effects depending on the type and state of the cell.
The new model is designed to capture those differences, predicting the effects of genetic variants on individual cell types and states. Google said in a blog post published Wednesday that it offers much more detailed insight than broader models like AlphaGenome, potentially revealing disease-relevant effects in specific cells that would otherwise go undetected.
AlphaGenome 2 was trained on more than 100 million single cell profiles across 68 tissues to give you a more detailed view. DeepMind says the model effectively turns static cell atlases (huge repositories of data that map and catalog individual cells in detail) “into a predictive engine for human genetics.” According to the company, AlphaGenome 2 identifies nearly twice as many variants that can change gene expression in cells than its previous next-generation model.
Notably, the researchers found that AlphaGenome 2 could also make predictions about new cell types and biological states without needing to retrain them. It could also identify gaps in the experimental data, which could help scientists decide what to investigate next in the lab.
For Adam Gayoso, a Google DeepMind research scientist involved in the project, the big breakthrough is “linking human genetics to cell biology.” AlphaGenome 2 offers researchers a way to drill down into genetic variants identified in large populations, to see what effects they may have at the cellular level, and potentially connect those effects to diseases, he said. The edge.
Žiga Avsec, head of genomics at Google DeepMind, said The edge This specificity is important if the goal is to discover how to treat a particular disease. “When you develop drugs, you target specific proteins and cell types,” he explained, which makes having this type of granularity vital.
Building a model like AlphaGenome 2 has only recently been made possible by new technologies that can capture individual cells and their states in high resolution and create the vast data sets needed to train it, Avsec said. DeepMind is now working to improve its genetic models by generating more data, including for cellular states that are not currently well represented. Avsec said future versions could also be trained with data “from multiple species,” giving the models a broader view of biology. In the long term, he said, models like AlphaGenome 2 could contribute to efforts to create “virtual cells,” digital simulations designed to mimic biological systems.
Although Avsec said it sees AlphaGenome 2 as a successor to AlphaGenome, it won’t replace it anytime soon. Its more detailed predictions are computationally expensive and “more expensive” to produce than those of its predecessor, he said. Those demands would have made it impractical to use AlphaGenome 2 to produce the Whole Genome Atlas covering “all possible DNA letter changes in the human genome” that DeepMind published in September.
Google says AlphaGenome 2 will be available for non-commercial research through the AlphaGenome API starting today, with a limited “private preview” for commercial research.
The model is the latest in a growing series of efforts by Google and rival AI labs to use the technology to address important problems in science and medicine. Pushmeet Kohli, who heads DeepMind’s AI for scientific research, has hinted that more impressive scientific results are on the way and the organization is expected to soon open an automated materials science laboratory in the UK. Meanwhile, Anthropic has been experimenting with drug design and giving Claude a wet lab, while OpenAI has been eliminating an impressive number of long-standing problems in mathematics and related fields.
