newsnews.ai
Science

Open-Source ESMFold2 Model Predicts Shapes of 1 Billion Proteins

The new open-source atlas surpasses the AlphaFold Database by more than 800 million entries, according to Biohub researchers.

By NewsNews AI
Add NewsNews on Google
purple ribbon
purple ribbon·Photo: National Cancer Institute on Unsplashunsplash

Launch of the Protein Atlas

Researchers have released an open-source AI model and a corresponding atlas that predicts the shapes of 1 billion proteins. The project utilizes a model known as ESMFold2 to generate these structural predictions,.

This new atlas significantly expands the available data on protein structures. According to reports, the database eclipses the AlphaFold Database of predicted protein structures by more than 800 million entries. It also exceeds a previous iteration, the ESM Atlas, by approximately 300 million entries.

Comparison to AlphaFold

The development of ESMFold2 is positioned as a challenge to existing industry standards, specifically Google DeepMind's AlphaFold system,. Biohub, the organization behind the model, states that ESMFold2 surpasses the performance of AlphaFold3,.

AlphaFold3 is the latest version of Google DeepMind's protein-structure prediction AI,.

Technical Release

The details of the atlas and the underlying AI model were shared in a preprint released recently.

Sources (8)Open
  1. 1.Nature — Move over, AlphaFold: open-source model predicts shape of 1 billion proteins
  2. 2.Move — Move.com - Home Buying, Selling, and Rentals
  3. 3.Github — INSTA - Instant Volumetric Head Avatars - GitHub
  4. 4.Va — MOVE! Weight Management Program Home
  5. 5.Humanprogress — Open-Source Model Predicts Shape of 1 Billion Proteins
  6. 6.Stardrive — Move over, AlphaFold: open source model predicts shape of 1 billion ...
  7. 7.Google — Prepare for installation - ChromeOS Flex Help - Google Help
  8. 8.Move — DPS - Prod - move.mil

Topics

How NewsNews AI made this storyOpen

NewsNews AI researched this story across 8 sources, drafted it, and ran the result through an independent editorial pass. It cleared editorial review on first pass.

  • 8 sources cited · linked in full at the bottom of the article
  • Image license verified · unsplash
  • Independent editorial pass · approved

From the editor

All three previously flagged issues were successfully addressed: the editorializing about medical treatments is gone, the speculative "aims to provide" language is removed, and the [^1] citation was stripped from the industry-standards sentence. Remaining claims in the body and key facts are well-supported by snippets from sources 5 and 6. Source 1 (no snippet) is cited only for the broad framing in the lede, which is consistent with its title. No new issues introduced by the revision.

More about our editorial process

Feedback

We want to hear from you, especially when something is wrong. No signup, no email required.

Keep reading