Field Guide
Protein design
Generative and modular design, from backbone diffusion to the modelling you can run on a laptop.
Other chapters
X-ray and crystallisation
Cryo-EM
Fundamentals
Prediction and validation
Protein design
AI workflows and chemistry tools
Docking, screening and simulation
Try it
FAFFAbout ↗
forecasts where a protein is likely to stall on the road to a structure, from 335,771 real attempts in the Protein Structure Initiative arch…
ALPHABETTI ↗
a protein sequence logo in 3D, wrapped around its own predicted structure. The amino acid letters themselves, stacked by how strongly a lang…
ButtFold ↗
a protein folds in your browser and the trajectory becomes music: contacts forming are note onsets, hydropathy sets pitch, secondary structu…
TopPDBLX ↗
every crystallisation condition in the protein data bank, parsed and indexed: what worked, how often, and how little of it is specific to yo…
AlphaFraud ↗
Catch AlphaFold where the fold is wrong: weekly, automatically, on freshly deposited human structures
ChatPDB ↗
A protein-structure-literate AI assistant for analyzing proteins: RAG for facts, a QLoRA-tuned model for behaviour, and live Biopython/gemmi…
12 articles
Protein Hunter: Pioneering the Next Frontier of AI-Driven Protein Design
Protein Hunter: Pioneering the Next Frontier of AI-Driven Protein Design Protein Hunter: exploiting structure hallucination within diffusion for protein design Revolution…
Revolutionizing Drug Discovery: LatentX and the Future of Protein Binder Design
Revolutionizing Drug Discovery: LatentX and the Future of Protein Binder Design LatentX generates lab-functional protein binders with picomolar affinities and near-perfec…
The New Architects: How Generative AI Is Designing Novel Protein Scaffolds
The New Architects: How Generative AI Is Designing Novel Protein Scaffolds The past, present and future of de novo protein design. Nature, 2026 From renovation to archite…
Building Proteins Brick by Brick: How AI Is Assembling Novel Bio-Machines
Building Proteins Brick by Brick: How AI Is Assembling Novel Bio-Machines AI models like ProDomino are being developed to predict how to successfully combine different pr…
AI-Driven Inverse Protein Folding: Transforming Protein Design for Next-Gen Therapeutics
AI-Driven Inverse Protein Folding: Transforming Protein Design for Next-Gen Therapeutics Advanced AI models are solving the “inverse folding problem,” allowing scientists…
The Protein Architect: How Generative AI Is Building the Drugs of Tomorrow
The Protein Architect: How Generative AI Is Building the Drugs of Tomorrow Generative AI has evolved from predicting existing protein structures to designing entirely new…
Beyond Trial and Error: AI-Driven Design of Peptide Therapeutics
Beyond Trial and Error: AI-Driven Design of Peptide Therapeutics AI is transforming peptide drug discovery by replacing slow, trial-and-error methods with rapid, in silic…
What If Protein Structures Could Be Designed Like LEGO? Modular Approaches in 2025
What If Protein Structures Could Be Designed Like LEGO? New AI models are enabling a modular, “LEGO-like” approach to protein design, where proteins are assembled from de…
Compositional Protein Design: Building Proteins from 3D Ellipsoid Blueprints
Compositional Protein Design: Building Proteins from 3D Ellipsoid Blueprints ProtComposer is a new AI model that generates protein structures based on a user-defined layo…
Flow-Based Models: A New Paradigm for Protein Structure Generation
Flow-Based Models: A New Paradigm for Protein Structure Generation Flow-based generative models like Proteina are an emerging alternative to diffusion models for designin…
Generative AI for Protein Pocket Design: The Next Leap in Drug Discovery
Generative AI for Protein Pocket Design: The Next Leap in Drug Discovery New generative AI models like PocketGen can design novel protein binding pockets from scratch, ta…
Installation of RFdiffusion on MacOSX M1 architecture without GPU support
Installation of RFdiffusion on MacOSX M1 architecture without GPU support Rosetta/DGL Issues: Some advanced functionalities (like DGL-based graph operations) may still fa…