Can BoltzGen design binders for small molecules?

Yes, BoltzGen includes a specific 'protein-small_molecule' protocol designed specifically for creating protein pockets that interact with ligands and small molecule targets.

What hardware is required to run BoltzGen?

It is recommended to use high-memory GPUs like the NVIDIA L40S or A100 (48GB+ VRAM). The skill simplifies this by supporting cloud execution via Modal to avoid local hardware limitations.

What file formats does BoltzGen use?

It uses YAML for configuration and primary outputs are in CIF format, which is the modern standard for macromolecular structures, though these can be converted to PDB for use in legacy tools.

How does BoltzGen differ from RFdiffusion?

While RFdiffusion is excellent for backbone generation, BoltzGen provides all-atom diffusion, meaning it designs side-chains and backbones simultaneously for higher precision, especially near binding sites.

BoltzGen All-Atom Protein Design

Name: BoltzGen All-Atom Protein Design
Author: adaptyvbio

byadaptyvbio

•

Data Science & ML

Generates high-precision all-atom protein structures and binders using advanced diffusion models for side-chain aware design.

BoltzGen is a specialized Claude Code skill for researchers and bio-engineers that facilitates all-atom protein design using BoltzGen diffusion models. Unlike backbone-only methods, it provides side-chain aware design from the start, making it ideal for creating binders for small molecules, ligands, and complex protein interfaces. By leveraging YAML-based configurations and Modal for cloud-scale compute, it allows users to execute sophisticated design protocols—including peptide, nanobody, and antibody design—directly within their development workflow.

Key Features

01Entity-based YAML configuration for complex binding site specifications.

02Cloud-scale execution via Modal with automated GPU provisioning.

0375 GitHub stars

04Comprehensive output including CIF structures, confidence metrics, and ranked results.

05All-atom diffusion for precise side-chain and backbone generation.

06Multi-protocol support for proteins, peptides, small molecules, and antibodies.

Use Cases

01Designing binders for small molecule ligands or specific protein pockets.

02Generating cyclic peptides with precise disulfide bond constraints.

03Engineering nanobody and antibody CDR regions for therapeutic development.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptyvbio/protein-design-skills boltzgen

For use in Claude.ai and ChatGPT

Key Features

01Entity-based YAML configuration for complex binding site specifications.

02Cloud-scale execution via Modal with automated GPU provisioning.

0375 GitHub stars

04Comprehensive output including CIF structures, confidence metrics, and ranked results.

05All-atom diffusion for precise side-chain and backbone generation.

06Multi-protocol support for proteins, peptides, small molecules, and antibodies.

Use Cases

01Designing binders for small molecule ligands or specific protein pockets.

02Generating cyclic peptides with precise disulfide bond constraints.

03Engineering nanobody and antibody CDR regions for therapeutic development.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add adaptyvbio/protein-design-skills boltzgen

For use in Claude.ai and ChatGPT