Predict biomolecular complex structures from FASTA-like inputs to CIF and NPZ outputs.
npx skills add https://github.com/LiorZ/protein-design-skills --skill chai-lab
Please help me install this Agent Skill. Skill: chai-lab Source: https://github.com/LiorZ/protein-design-skills/tree/main/skills/chai-lab Command: npx skills add https://github.com/LiorZ/protein-design-skills --skill chai-lab
Enables end-to-end prediction of biomolecular structures for complexes and assemblies using multi-modal models, reducing manual trial-and-error in structure generation.
Provide a FASTA-like input with all chains and run chai-lab fold to produce CIFs, scores, and MSAs outputs in an output directory.
High-intent search queries and answers about installing and using this skill.
Protein complex structure prediction supports protein–protein, protein–ligand, and protein–DNA/RNA assemblies. It uses multi-modal all-atom models to handle glycosylated proteins and binder-validation workflows, outputting CIF structures with per-sample NPZ scores.
Yes, MSAs and templates are optional inputs that guide interface geometry and confidence scoring in complex prediction. You can also apply restraints to further shape the interfaces and scoring for protein–protein, protein–ligand, and protein–DNA/RNA assemblies.
Yes, the chai_lab package supports batch biomolecular structure prediction through the fold-batch CLI command. This enables batch campaigns by outputting CIF structures and per-sample NPZ scores suitable for downstream ranking and quality control.
You can run inference for glycosylated protein structure prediction using the run_inference or run_folding_on_context Python API functions. Provide a FASTA-like input with your glycosylated chains and optional restraints to generate CIF structures and NPZ scores.
Biomolecular structure prediction outputs CIF structure files and per-sample scores in NPZ format. These outputs facilitate downstream ranking and quality control for predicted protein complexes, protein–ligand assemblies, and binder-validation workflows.
Validating protein binders requires a FASTA-like input of the binder and target, optionally supported by MSAs, templates, and restraints. Running the fold command produces CIF structures and per-sample NPZ scores to assess binder confidence and interface geometry.
LiorZPredict biomolecular structures and affinities from YAML inputs via CLI diffusion workflow.
junior1pRank protein binder designs using ipSAE_min, LIS, and pDockQ scores.
InternScienceGenerate BioEmu conformational samples from sequences and extract per-frame PDBs.
zongtingweiPredict protein complex structures and generate CIF models with confidence scores.
