In 2024, researchers found a upcoming driver of inflammatory bowel disease (IBD): it was a gene called ETS2 that is overactive in immune cells owing to a DNA variant that predisposes people to the condition1. This variant had previously confounded researchers because it lies in a ‘gene desert’ nowhere near protein-coding sequences.
Industry observers note that that discovery had involved gene-editing and other sophisticated molecular-biology techniques, and it identified potential fresh IBD treatments. Now, researchers at Google DeepMind in London have replicated aspects of the finding using an artificial-intelligence model, and they have done this without lifting a pipette. Their model is called AlphaGenome 2.
Sources indicate that “They’ve managed to figure out the answer in 30 minutes, and it took us a year and a half of experiments,” states James Lee, an immunologist and geneticist at the Francis Crick Institute in London, who led the 2024 study. “It’s reassuring from our perspective.”.
According to recent reports, the AlphaGenome 2 model, which was declared today, is an update to an AI tool that predicts the molecular effects that ‘non-coding’ DNA sequences have, such as on the expression of other genes. These sequences comprise 98% of the human genome, but they are among the hardest to decipher. The latest model makes these predictions in hundreds of mouse and human cell types.
Two months ago, Google DeepMind released the AlphaGenome Atlas: these were predictions for every possible single-letter DNA change to the human genome — a total of nine billion variants.
These and other predictions were made with a tool called AlphaGenome 1 and came with a major caveat: they did not specify the precise kind of cell — be it antibody-making B cells or pancreatic β-cells — in which a variant was likely to act. “Having something that cannot predict cell-type specificity is a bit of a problem, which is why it hasn’t been hugely useful for us,” notes Lee.
But researchers know that many non-coding DNA sequences — including the variant that Lee’s team mapped, which is part of a sequence called a gene enhancer — tend to be active in specific types of cell.
“If you want to understand variants associated with genetic diseases, then you have to ask yourself what are the cell types that are affected,” says Žiga Avsec, a DeepMind researcher who leads the AlphaGenome team.
To develop an AI model that could decipher variant effects in diverse cells, Avsec’s team trained AlphaGenome 2 on single-cell gene activity and other measurements made in more than 100 million cells, across 68 tissues. AlphaGenome 1, by contrast, was fed data collected in ‘bulk’ samples of heterogeneous cells.
The result is a richer, more fine-grained representation of the differences between cells, says Adam Gayoso, a Google DeepMind scientist and co-author of a preprint describing the model on the organization’s website.
Industry observers note that alphaGenome 2 also isn’t limited to making predictions in the 637 types of human cell and 311 kinds of mouse cell it was trained on. Researchers can provide the model with new single-cell gene-activity data — such as those from cells in a disease state — and generate gene-variant predictions. This could help researchers to identify variants underlying disease more promptly, says Gayoso.
doi: https://doi.org/10.1038/d41586-026-03673-0.
According to recent reports, stankey, C. T. et al. Nature 630, 447–456 (2024).
Article PubMed Google Scholar.
Human replisomes direct a dedicated leading-strand primase for fork restart.
The molecular basis of replication fork restart via repriming by PRIMPOL.
Domain-alteration chimeras rewire oncogenic complexes in prostate cancer.
Massive atlas of brain region linked to a host of diseases is the most detailed yet.
Move over BMI: upcoming obesity definition is better for predicting your health.
New culprit in early Alzheimer’s: blood–brain barrier blocks protective proteins.
Recent findings suggest that NBIC, Peking University, seeks Assistant Professors and Research Track faculty in imaging, probes, spatial omics and digital life. Open until filled.
National Biomedical Imaging Center, College of Future Technology, Peking University.
Job Title: Deputy Editor, Nature Progress Bioengineering Location: Shanghai, Beijing, Latest York, Philadelphia - hybrid working model Closing date: 2...
Shanghai, Beijing, New York, Philadelphia - hybrid working model.
The SoQ invites highly qualified faculty member in quantum sciences & technology and related fields.
School of Quantum at the University of Chinese Academy of Sciences (UCAS).
Join HZAU's global faculty team to advance research with competitive benefits.
Huazhong Agricultural University (HZAU).
The School of Life Science and Technology (SLST) invites applications for numerous tenure-track and tenured positions.
