Generate ligand-aware protein sequences for defined binding pockets from PDB files.
npx skills add https://github.com/junior1p/ProteinClaw --skill ligandmpnn-junior1p
Please help me install this Agent Skill. Skill: ligandmpnn Source: https://github.com/junior1p/ProteinClaw/tree/main/skills/ligandmpnn Command: npx skills add https://github.com/junior1p/ProteinClaw --skill ligandmpnn-junior1p
This Skill generates protein sequences optimized around bound small molecules, metal ions, cofactors, and designed ligands to enable ligand-aware inverse folding and pocket-specific sequence design without manual sequence curation.
Run LigandMPNN on your protein-ligand PDB to generate 16 candidate sequences with sampling temperature 0.1.
High-intent search queries and answers about installing and using this skill.
Ligand-aware inverse folding generates protein sequences conditioned on HETATM ligand records in PDB structures. It optimizes sequence identity around bound small molecules, metal ions, and cofactors to enable precise pocket-specific design.
Ligand-aware inverse folding is a sequence design technique that conditions amino acid generation on defined protein-ligand binding pockets. It is needed for enzyme active site redesign, small-molecule binder generation, and metal coordination stabilization.
Yes, ligand-aware protein sequence design requires Python with GPU resources to sample multiple sequences per target. You also need input PDB files containing HETATM ligand records to define the binding pocket context.
Sequence sampling diversity is controlled using sampling temperature and sample-count parameters. Running LigandMPNN with temperature 0.1 generates 16 candidate sequences per target, allowing integration with downstream structure prediction workflows.
Yes, ligand-aware inverse folding applies to metal coordination site design by conditioning sequence generation on metal ion HETATM records. It stabilizes metal binding contexts alongside small-molecule and cofactor binding pocket engineering.
Ligand-aware protein sequence design requires PDB files with explicit HETATM ligand records and Python with GPU resources. It generates candidate sequences requiring downstream structure prediction and quality control for validation.
NVIDIA-BioNeMoEvaluate directories of PDB files by orchestrating refolding and structural analysis workflows.
K-Dense-AIPredicts protein-ligand binding poses from PDB or sequence and SMILES inputs.
zongtingweiPredict protein and complex structures with Boltz-1 / Boltz-2 and output CIF files.
PKU-YuanGroupConverts PDF documents to clean, structured Markdown.
