> ## Content Index > Fetch the complete content index at: https://www.notatechguy.com/llms.txt > Use this file to discover other available public pages before exploring further. # AlphaGenome 2 maps DNA variants across hundreds of cell types - URL: https://www.notatechguy.com/alphagenome-2-maps-dna-variants-across-hundreds-of-cell-types/ - Published: 2026-11-12T07:16:39.000Z - Updated: 2026-11-12T07:16:39.000Z - Description: Google DeepMind's AlphaGenome 2 predicts how genetic variants affect hundreds of human cell types, but the launch ships without published benchmarks - Author: Marcello Babbili - Tags: Technology & AI, Google Google DeepMind released AlphaGenome 2 on November 11, a DNA sequence model that predicts how genetic variants affect hundreds of human cell types by linking single-cell biology with human genetics [S¹](https://deepmind.google/blog/alphagenome-2-deciphering-how-cells-read-dna-to-better-understand-disease/?ref=notatechguy.com)[P⁴](https://blog.google/intl/ko-kr/company-news/technology/alphagenome-2-kr/?ref=notatechguy.com). The launch ships with no published accuracy benchmarks, no peer review, and no evidence of clinical deployment. It is impossible to judge whether the predictions outperform existing genomic tools. **My read:** The capabilities listed in the GitHub release — expression-conditioned predictions, variant scoring, cell embeddings, and input gradients, match the toolset a working genomics lab would expect to use [P³](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com). What I cannot assess is whether the predictions are any good, because DeepMind has not published benchmark numbers alongside the announcement. The detail I would watch is whether independent labs can reproduce the variant effect predictions on known disease-associated loci, because that is where the model's value either holds or collapses. ### What the release actually contains The code landed on GitHub the same day as the blog post. Tom Ward, a DeepMind engineer identified as a core contributor, pushed version 2.0.0 to the google-deepmind/alphagenome repository on November 11 [P³](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com)[S¹](https://deepmind.google/blog/alphagenome-2-deciphering-how-cells-read-dna-to-better-understand-disease/?ref=notatechguy.com). The v2.0.0 release introduces a new API namespace (`alphagenome.models.v2`) that lets users generate predictions conditioned on either a reference sequence or single-cell expression data. The same API supports scoring of genetic variants and genomic intervals, along with utilities for extracting cell embeddings and computing input gradients, per the GitHub release notes [P³](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com). The release also bundles a `GeneExpressionData` object for handling gene-level expression output [P³](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com). The blog post is authored by research scientists Adam Gayoso and Natasha Latysheva, alongside Pushmeet Kohli, VP Science at Google DeepMind and Chief Scientist at Google Cloud, and Žiga Avsec, also a research scientist, with Demis Hassabis credited among the core research team [P⁴](https://blog.google/intl/ko-kr/company-news/technology/alphagenome-2-kr/?ref=notatechguy.com)[S¹](https://deepmind.google/blog/alphagenome-2-deciphering-how-cells-read-dna-to-better-understand-disease/?ref=notatechguy.com). ### What is missing DeepMind has not published predictive accuracy benchmarks, training dataset details, or model architecture specifications alongside the announcement. No comparison against existing variant effect predictors exists in the released material. The work has not been peer-reviewed, and the announcement includes no evidence of clinical or commercial deployment. The original AlphaGenome project page describes the genome as "the blueprint of life" and frames AI as a tool for understanding why diseases occur and how to treat them, but that page does not yet reflect the version 2 release or its capabilities [P²](https://deepmind.google/science/alphagenome/?ref=notatechguy.com). A computational biologist who wanted to test AlphaGenome 2 today could pull the v2.0.0 release from GitHub, load the API, and run variant scoring on a set of known disease-associated SNPs. Comparing the model's predicted effects against experimentally validated results from their own lab is the only thing that would tell them whether to trust it. Whether independent genomics labs publish reproduction studies against known disease loci is the checkpoint that would tell the field whether to trust the v2.0.0 API released on November 11 [P³](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com). --- *Sources: [S1 — AlphaGenome 2: Deciphering how cells read DNA to better understand dis](https://deepmind.google/blog/alphagenome-2-deciphering-how-cells-read-dna-to-better-understand-disease/?ref=notatechguy.com) · [P2 — AlphaGenome — Google DeepMind](https://deepmind.google/science/alphagenome/?ref=notatechguy.com) · [P3 — Releases · google-deepmind/alphagenome · GitHub](https://github.com/google-deepmind/alphagenome/releases?ref=notatechguy.com) · [P4 — 알파게놈 2(AlphaGenome 2): 질병 이해를 돕는 세포 속 DNA 해독](https://blog.google/intl/ko-kr/company-news/technology/alphagenome-2-kr/?ref=notatechguy.com) · [P5 — Google DeepMind](https://github.com/google-deepmind/?ref=notatechguy.com)* --- *Written from 5 sourced items, 5 of them primary.* ## More from Not A Tech Guy - [BotPsy 0.8B agent model beats Qwen 2B on five benchmarks](https://www.notatechguy.com/botpsy-0-8b-agent-model-beats-qwen-2b-on-five-benchmarks/) - [Google Gemini Decisions API: 150ms choices at $0.10 per million](https://www.notatechguy.com/google-gemini-decisions-api-150ms-choices-at-0-10-per-million/) - [PocketPal AI runs Llama on your phone offline, hits 9,200 stars](https://www.notatechguy.com/pocketpal-ai-runs-llama-on-your-phone-offline-hits-9-200-stars/)