/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

DeepMind says its latest AlphaFold model can generate predictions for nearly all molecules in the Protein Data Bank and for ligands, nucleic acids, and more

Kyle Wiggers / TechCrunch :

TechCrunch Kyle Wiggers

Context & Ripple Effects

AlphaFold’s arc began with competitive protein-structure prediction and then reached a claimed near-comprehensive catalog of known protein structures. This report extends the stated ambition beyond proteins to molecules recorded in the Protein Data Bank, including ligands and nucleic acids.

That matters because many biological questions concern interactions among molecular components, not protein shape in isolation. It foreshadows the broader interaction-focused direction later detailed in AlphaFold 3’s account of cellular building blocks and their interactions.

First-order effects

  • DeepMind broadens AlphaFold’s claimed prediction scope from protein structures to nearly all Protein Data Bank molecules and to ligands and nucleic acids.
  • Researchers studying molecular complexes gain a potentially more unified prediction toolset rather than treating protein, ligand, and nucleic-acid questions as wholly separate modeling tasks.

Second-order effects

  • Competing scientific-AI teams face a higher bar: usefulness increasingly depends on modeling interactions across molecular types, not only predicting individual protein structures.
  • Drug-discovery and biology users may consolidate early-stage structural hypotheses around platforms that cover more of a complex, while still needing experimental validation for consequential results.

Third-order effects

  • If performance holds across these molecule classes, structural biology AI could shift from narrow prediction benchmarks toward general-purpose molecular-interaction systems.
  • The value in scientific AI would increasingly rest on access to broad biological datasets, compute, and validation workflows—assets that favor well-resourced labs but leave accuracy and practical adoption as the decisive tests.

The trend: AlphaFold is part of a shift from single-task scientific models toward AI systems that represent interacting components of biology.

Discussion

  • @sergeiiakhnin Sergei Yakneen on x
    I'm incredibly proud to share this joint work by @IsomorphicLabs and @GoogleDeepMind on the latest version of AlphaFold, demonstrating SOTA on a wide variety of tasks, including protein-ligand structure, nucleic acids, PPIs, and PTMs. [video]
  • @raulferrerdev Raúl Ferrer on x
    @demishassabis @IsomorphicLabs It's impressive to see the advances in AlphaFold: it is able to predict structures that go beyond proteins, to ligands, nucleic acids and more. It has enormous potential to accelerate drug discovery and biological understanding. Awesome to see AI un…
  • @tfgg2 Tim Green on x
    New! We've just put up a note evaluating the latest, in-development version of AlphaFold ("AlphaFold-latest"). This is a preview - development is still in progress - but performance across a wide range of tasks is striking. https://deepmind.google/... Highlights in the thread. 1/…
  • @ai_ctrl @ai_ctrl on x
    The amazing success of AlphaFold proves that we don't need godlike AI in order for AI to benefit humanity. Controlling and limiting frontier AI models that aspire to be superintelligent isn't going to hobble innovation.
  • @demishassabis Demis Hassabis on x
    Excited to share major progress on next generation of AlphaFold! Together with @IsomorphicLabs, we've significantly advanced accuracy & expanded coverage beyond proteins to other key biomolecular classes. Continuing to accelerate science at digital speed https://dpmd.ai/...
  • @moalquraishi Mohammed AlQuraishi on x
    Interesting status update from DeepMind on AlphaFold (just that, no model, paper, or code). All atom version in the works (similar to RFAA). Meaningful gains on small molecules but far from ‘solved’ (think AF1 vs AF2). Same w/nucleic acids and antibodies. [image]