Generate side-chain-aware protein binder designs from YAML-configured campaigns using BoltzGen diffusion.
npx skills add https://github.com/junior1p/ProteinClaw --skill boltzgen-junior1p
Please help me install this Agent Skill. Skill: boltzgen Source: https://github.com/junior1p/ProteinClaw/tree/main/skills/boltzgen Command: npx skills add https://github.com/junior1p/ProteinClaw --skill boltzgen-junior1p
BoltzGen removes the manual complexity of generating side-chain-aware protein binders by providing an all-atom diffusion workflow that produces geometry-precise designs from YAML specifications, enabling researchers to design binders against proteins, peptides, and small molecules with GPU-accelerated performance.
Design 50 side-chain-aware binders for target.cif using binder_config.yaml with the protein-anything protocol.
High-intent search queries and answers about installing and using this skill.
To design side-chain-aware protein binders, you use an all-atom diffusion workflow that generates geometry-precise structures from YAML specifications. This produces explicit side-chains for accurate interface geometry across proteins, peptides, and small molecules.
For a YAML-driven binder generation campaign, you need CIF or PDB files for your target proteins. The workflow uses entity-based YAML configuration to define targets and restraints, outputting design.cif and sequence.fasta files.
Yes, all-atom protein design requires a CUDA-enabled GPU for execution. You can run the diffusion workflow via Modal for cloud GPU orchestration or use a local installation for on-prem GPUs.
Yes, you can generate nanobody and antibody designs using side-chain-aware diffusion. The system supports multi-protocol configurations for protein, peptide, nanobody, antibody, and protein-small molecule binding scenarios.
An all-atom binder generation workflow produces design.cif, sequence.fasta, and metrics.json outputs. These files provide the generated structures, sequences, and quality control metrics needed for downstream structure validation and scoring.
adaptyvbioAutomate all-atom protein design with diffusion models via YAML configuration.
zongtingweiScore protein sequences and generate embeddings using the ESM2 model.
adaptyvbioGenerate protein backbones with RFdiffusion using contigmap and hotspot specifications.
zongtingweiDesign all-atom protein structures with side-chain packing around target binding sites.
