Google DeepMind releases AlphaGenome 2 for cell state DNA predictions
Google DeepMind has launched AlphaGenome 2, an AI model that predicts regulatory activity and variant effects in specific human cell types.
FTMQ SI, written by our newsroom0 views
Google DeepMind has released AlphaGenome 2, a DNA sequence model designed to predict how genetic variants affect specific human cell types. MarkTechPost, The Verge AI, and SiliconANGLE AI all report that the new system models the impact of changes in human DNA across target cell states. The technology aims to help researchers analyze cellular disease causes and identify potential paths to medical treatment. [1][2][4]
According to MarkTechPost, AlphaGenome 2 is structured as a U-Net style conditional sequence-to-function model featuring a convolutional encoder, an 11-block transformer, and a decoder. It processes 1 megabase of input DNA along with a target cell state. MarkTechPost also reports that the system adapts to unseen single-cell datasets without fine-tuning and recovers 30 percent of causal single-cell expression quantitative trait loci, compared to 16 percent for prior approaches. [1]
The model serves as the successor to the original AlphaGenome system, The Verge AI reports. While the earlier model averaged regulatory predictions across bulk tissues, AlphaGenome 2 resolves closely related cell types to provide targeted predictions, according to MarkTechPost and The Verge AI. [1][2]
DeepMind Technologies Limited, which operates as Google DeepMind, is an artificial intelligence research laboratory and subsidiary of Alphabet Incorporated founded in 2010. MarkTechPost notes that AlphaGenome 2 is currently available through the AlphaGenome API for non-commercial research purposes. [1][11]
In short
- Google DeepMind released AlphaGenome 2 to predict DNA variant effects per cell type.
- The model architecture uses a U-Net style setup with a convolutional encoder, an 11-block transformer, and a decoder.
- AlphaGenome 2 evaluates 1 megabase of input DNA alongside a specified cell state.
- The model is available via the AlphaGenome API for non-commercial research.
Sources
Every paragraph above points to the numbered items it rests on. Read the originals here.
- [1]Google DeepMind Releases AlphaGenome 2 for Cell-Type-Specific Variant EffectsMarkTechPost, 6d ago (the report this story comes from)
- [2]Google’s AlphaGenome 2 zooms in on the cellular roots of diseaseThe Verge AI, 6d ago
- [3]Google’s AlphaGenome 2 zooms in on the cellular roots of diseaseThe Verge, 6d ago
- [4]Google DeepMind releases new advanced robotics AI model and next-generation gene prediction modelSiliconANGLE AI, 6d ago
- [5]AlphaGenome 2: Deciphering how cells read DNA to better understand diseaseDeepMind Research News, 6d ago
- [6]Google DeepMind AI Verifies Signal Encryption Behind WhatsApp and Private Chats in Major Security BreakthroughConvergence Now, 6d ago
- [7]AlphaGenome 2: Deciphering how cells read DNA to better understand diseaseGoogle DeepMind, 6d ago
- [8]Google DeepMind releases AlphaGenome 2 to predict how DNA changes affect specific cell types and diseaseMLQ.ai, 6d ago
- [9]Google DeepMind's AlphaGenome 2 Maps DNA Changes Across Cell TypesThe Tech Buzz, 6d ago
- [10]Google DeepMind launches AlphaGenome 2 to predict DNA effects in single cell typesStartup Fortune, 6d ago
Background
- [11]Google DeepMind on Wikipedia
- [12]Ehlers–Danlos syndrome on Wikipedia
Our newsroom writes these reports with the help of software, from the 12 sources listed and nothing else, and checks them against those sources. Facts can still be wrong or move on; the originals are the record. Spotted a mistake? Write to daniel@monsterkong.com.
Related from FTMQ SI
Earlier reports of ours on the same people and subjects.
- Amazon and Google to control combined 30 percent Anthropic stake post IPOTop Stories, 5h ago
- Hugging Face releases open Carbon-X model for DNA and natural languageModels & Products, 23d ago
More in Models & Products
- Microsoft unveils Copilot Workspaces and ROI tools at Ignite 20265h ago
- Meta deploys LLM test system to catch software bugs in production1d ago
- Alibaba releases Qwen4-Max model with native vision MoE architecture1d ago
- NiCE launches next generation Cognigy platform for agentic contact center automation2d ago
- AI scanning tool uncovers dozens of Linux kernel bugs2d ago
- Cmpnd releases dspy-wsp for weight space programming in DSPy2d ago
Get the day in one email
Reports like this one, the top news of the last 24 hours, every morning. Free, one email a day; readers can comment under every report.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.

Comments
Loading
Join the conversation
Comments are open to readers of our daily email: the top news of the last 24 hours, every morning, free. Sign up and the comment box opens.
Already on the list? Enter the same address and we will send a sign-in link.
By signing up you agree to our terms and privacy policy. Unsubscribe any time.