[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74MCFX2G6C2M6W789JX5B41":3,"skill-same:01M74MCFX2G6C2M6W789JX5B41":56},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":33,"commitSha":34,"skillDirectory":35,"forksCount":36,"repositoryHtmlUrl":37,"repositoryHomepageUrl":38,"ownerHtmlUrl":39,"ownerBlogUrl":6,"ownerDescription":40,"updatedAt":41,"install":42,"resources":55},"01M74MCFX2G6C2M6W789JX5B41","ligandmpnn",null,"Inverse-fold a backbone with ligand, nucleic-acid, and metal context using LigandMPNN (Dauparas et al. 2023, github.com\u002Fdauparas\u002FLigandMPNN). Reach for this skill to redesign the residues lining a binding pocket around a bound small molecule or cofactor, to design metal-coordinating sites where the geometry must be respected, or to get threaded designed-sequence PDBs out of any MPNN run.","pengqianhan\u002FAI-Human-Research-OS","pengqianhan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F4286531?v=4","01M74M3DT9HNZ6HPAK8TX69EX9","PengqianHan","01M74M3DTVSN2GXW0JH2W396FK",3,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fclaude-science-skills\u002Fskills\u002Fligandmpnn\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: ligandmpnn\ndescription: >\n  Inverse-fold a backbone with ligand, nucleic-acid, and metal context using\n  LigandMPNN (Dauparas et al. 2023, github.com\u002Fdauparas\u002FLigandMPNN). Reach for\n  this skill to redesign the residues lining a binding pocket around a bound\n  small molecule or cofactor, to design metal-coordinating sites where the\n  geometry must be respected, or to get threaded designed-sequence PDBs out of\n  any MPNN run.\nlicense: Apache-2.0\ncategory: biomodels\nmetadata:\n  display-name: LigandMPNN\n  # github.com\u002Fdauparas\u002FLigandMPNN\u002Fblob\u002Fmain\u002FLICENSE: MIT (© 2024 Justas\n  # Dauparas). verified 2026-06-30\n  third_party:\n    - kind: weights\n      name: LigandMPNN\n      license: MIT\n      terms_url: https:\u002F\u002Fgithub.com\u002Fdauparas\u002FLigandMPNN\u002Fblob\u002Fmain\u002FLICENSE\n---\n\n# LigandMPNN\n\nLigandMPNN extends the ProteinMPNN graph with non-protein atoms — small\nmolecules, nucleic acids, and metals are visible to the network — so it is the\nright inverse-folding tool whenever the design surface includes a bound ligand\nor cofactor that vanilla `proteinmpnn` would ignore. The same `run.py` is also\nthe most convenient runner for the other MPNN families because, unlike the\noriginal ProteinMPNN script, it threads designs back onto the input structure\nand writes PDBs alongside the FASTA. Code and weights are MIT\n(github.com\u002Fdauparas\u002FLigandMPNN). The model is small enough to run on CPU —\nfor a handful of designs on one structure that is seconds and usually faster\nthan dispatching, so the normal path is local with\n`pip install torch numpy biopython ProDy ml_collections dm-tree`; a GPU helps\nfor batched campaigns.\n\n## Running it\n\n```bash\npip install torch numpy biopython ProDy ml_collections dm-tree\ngit clone --depth 1 https:\u002F\u002Fgithub.com\u002Fdauparas\u002FLigandMPNN.git ligandmpnn\ncd ligandmpnn\nsed -i 's\u002Fnp\\.int\\b\u002Fnp.int64\u002Fg' openfold\u002Fnp\u002Fresidue_constants.py   # repo pins numpy 1.23; alias removed in >=1.24\nbash get_model_params.sh .\u002Fmodel_params\npython run.py \\\n  --model_type ligand_mpnn \\\n  --checkpoint_ligand_mpnn .\u002Fmodel_params\u002Fligandmpnn_v_32_010_25.pt \\\n  --pdb_path complex.pdb \\\n  --out_folder out \\\n  --batch_size 8 --number_of_batches 4 \\\n  --temperature 0.1 \\\n  --fixed_residues \"A45 A46 A47 A48\"\n```\n\nResidue selections are space-separated `{chain}{resnum}` tokens inside one\nquoted string (`\"A45 A46 B10\"`; insertion codes append directly, `\"B82A\"`).\nThat is the format for `--fixed_residues` and `--redesigned_residues`;\n`--bias_AA_per_residue` and `--omit_AA_per_residue` instead take a path to a\nJSON file whose keys use the same `{chain}{resnum}` form, and\n`--chains_to_design` is comma-separated (`\"A,B\"`). If you want to redesign only the pocket, naming the pocket residues\nin `--redesigned_residues` is usually shorter than fixing everything else.\n\nUnder `--out_folder` you get `seqs\u002F\u003Cstem>.fa` (headers carry\n`overall_confidence` and `ligand_confidence`), `backbones\u002F\u003Cstem>_{1..N}.pdb`\nwith the designed sequence threaded onto the input coordinates, and — with\n`--pack_side_chains 1` — full-atom packed models in `packed\u002F`. The threaded\nPDBs are the reason to prefer this runner even for protein-only jobs.\n\n## Model types — which one to pick\n\n| `--model_type` | sees | use |\n|---|---|---|\n| `ligand_mpnn` | backbone + ligand\u002FNA\u002Fmetal atoms | binding-pocket or active-site design |\n| `protein_mpnn` | backbone only | protein–protein; same weights as `proteinmpnn` |\n| `soluble_mpnn` | backbone only, soluble-trained | expression-biased prior; see `solublempnn` |\n| `*_membrane_mpnn` | backbone + membrane label | transmembrane designs |\n\nEach model type has its own `--checkpoint_\u003Ctype>` flag; the wrong pairing is\ncaught at load time, but the default checkpoint path is relative to the repo,\nso run from inside the clone or pass the absolute path.\n\n## ProDy compiles from source on py3.11 — `pip install` fails without a C compiler\n\n`run.py` imports ProDy unconditionally for ligand atom parsing. On py3.11 the\nprebuilt wheel is missing on PyPI, so `pip install ProDy` compiles from source\nand needs a working C\u002FC++ compiler. On Modal's `add_python` bases the default\n`CXX=clang++` points at a missing binary — `apt_install(\"build-essential\")`\nand export `CC=gcc CXX=g++` before the install. On most CPU-local Python\ndistributions the sdist builds in ~10 s if no wheel matches your Python.\n\n## Turning ligand context off changes the answer, not the model\n\n`--ligand_mpnn_use_atom_context 0` keeps the ligand-aware weights but masks the\nligand atoms at inference. That is useful for an ablation — the difference\nbetween context-on and context-off tells you how much the ligand is shaping the\ndesign — but it is **not** equivalent to running `protein_mpnn`, which uses a\ndifferent checkpoint trained without those features. For a fair protein-only\nbaseline, switch `--model_type`.\n\n## Stripped HETATM or a chain filter silently drops the ligand — the design comes back pocket-blind\n\nLigandMPNN does not warn when no ligand atoms are found; it just runs as if\n`--model_type protein_mpnn` had been picked. The two common ways this happens\nare an input PDB whose HETATM records were stripped by an upstream\nclean-up step, and `--parse_these_chains_only` naming the protein chains but\nnot the ligand's. If `ligand_confidence` in the FASTA header is missing or\nzero across every design, the model never saw the ligand — fix the input, do\nnot trust the sequences.\n\n## Errors worth recognizing\n\n| You see | It means \u002F do this |\n|---|---|\n| `ModuleNotFoundError: No module named 'tree'` | `pip install dm-tree` — the vendored openfold imports it unconditionally. |\n| `module 'numpy' has no attribute 'int'` | Run the `sed` patch on `openfold\u002Fnp\u002Fresidue_constants.py`, or pin `numpy\u003C1.24` (py≤3.11 only). |\n| `error: command 'clang' failed` while `pip install ProDy` | See the ProDy gotcha above — `apt_install(\"build-essential\")` and `env({\"CC\":\"gcc\",\"CXX\":\"g++\"})`. |\n| `FileNotFoundError` for `model_params\u002F...` | Checkpoints not fetched — run `bash get_model_params.sh .\u002Fmodel_params` from inside the clone. |\n\n---\n\n**Next:** fold the designs in complex with the ligand via `boltz` or `chai1`\n(both accept SMILES\u002FCCD) and filter on ipTM and ligand placement.\n","# LigandMPNN\n\nLigandMPNN extends the ProteinMPNN graph with non-protein atoms — small\nmolecules, nucleic acids, and metals are visible to the network — so it is the\nright inverse-folding tool whenever the design surface includes a bound ligand\nor cofactor that vanilla `proteinmpnn` would ignore. The same `run.py` is also\nthe most convenient runner for the other MPNN families because, unlike the\noriginal ProteinMPNN script, it threads designs back onto the input structure\nand writes PDBs alongside the FASTA. Code and weights are MIT\n(github.com\u002Fdauparas\u002FLigandMPNN). The model is small enough to run on CPU —\nfor a handful of designs on one structure that is seconds and usually faster\nthan dispatching, so the normal path is local with\n`pip install torch numpy biopython ProDy ml_collections dm-tree`; a GPU helps\nfor batched campaigns.\n\n## Running it\n\n```bash\npip install torch numpy biopython ProDy ml_collections dm-tree\ngit clone --depth 1 https:\u002F\u002Fgithub.com\u002Fdauparas\u002FLigandMPNN.git ligandmpnn\ncd ligandmpnn\nsed -i 's\u002Fnp\\.int\\b\u002Fnp.int64\u002Fg' openfold\u002Fnp\u002Fresidue_constants.py   # repo pins numpy 1.23; alias removed in >=1.24\nbash get_model_params.sh .\u002Fmodel_params\npython run.py \\\n  --model_type ligand_mpnn \\\n  --checkpoint_ligand_mpnn .\u002Fmodel_params\u002Fligandmpnn_v_32_010_25.pt \\\n  --pdb_path complex.pdb \\\n  --out_folder out \\\n  --batch_size 8 --number_of_batches 4 \\\n  --temperature 0.1 \\\n  --fixed_residues \"A45 A46 A47 A48\"\n```\n\nResidue selections are space-separated `{chain}{resnum}` tokens inside one\nquoted string (`\"A45 A46 B10\"`; insertion codes append directly, `\"B82A\"`).\nThat is the format for `--fixed_residues` and `--redesigned_residues`;\n`--bias_AA_per_residue` and `--omit_AA_per_residue` instead take a path to a\nJSON file whose keys use the same `{chain}{resnum}` form, and\n`--chains_to_design` is comma-separated (`\"A,B\"`). If you want to redesign only the pocket, naming the pocket residues\nin `--redesigned_residues` is usually shorter than fixing everything else.\n\nUnder `--out_folder` you get `seqs\u002F\u003Cstem>.fa` (headers carry\n`overall_confidence` and `ligand_confidence`), `backbones\u002F\u003Cstem>_{1..N}.pdb`\nwith the designed sequence threaded onto the input coordinates, and — with\n`--pack_side_chains 1` — full-atom packed models in `packed\u002F`. The threaded\nPDBs are the reason to prefer this runner even for protein-only jobs.\n\n## Model types — which one to pick\n\n| `--model_type` | sees | use |\n|---|---|---|\n| `ligand_mpnn` | backbone + ligand\u002FNA\u002Fmetal atoms | binding-pocket or active-site design |\n| `protein_mpnn` | backbone only | protein–protein; same weights as `proteinmpnn` |\n| `soluble_mpnn` | backbone only, soluble-trained | expression-biased prior; see `solublempnn` |\n| `*_membrane_mpnn` | backbone + membrane label | transmembrane designs |\n\nEach model type has its own `--checkpoint_\u003Ctype>` flag; the wrong pairing is\ncaught at load time, but the default checkpoint path is relative to the repo,\nso run from inside the clone or pass the absolute path.\n\n## ProDy compiles from source on py3.11 — `pip install` fails without a C compiler\n\n`run.py` imports ProDy unconditionally for ligand atom parsing. On py3.11 the\nprebuilt wheel is missing on PyPI, so `pip install ProDy` compiles from source\nand needs a working C\u002FC++ compiler. On Modal's `add_python` bases the default\n`CXX=clang++` points at a missing binary — `apt_install(\"build-essential\")`\nand export `CC=gcc CXX=g++` before the install. On most CPU-local Python\ndistributions the sdist builds in ~10 s if no wheel matches your Python.\n\n## Turning ligand context off changes the answer, not the model\n\n`--ligand_mpnn_use_atom_context 0` keeps the ligand-aware weights but masks the\nligand atoms at inference. That is useful for an ablation — the difference\nbetween context-on and context-off tells you how much the ligand is shaping the\ndesign — but it is **not** equivalent to running `protein_mpnn`, which uses a\ndifferent checkpoint trained without those features. For a fair protein-only\nbaseline, switch `--model_type`.\n\n## Stripped HETATM or a chain filter silently drops the ligand — the design comes back pocket-blind\n\nLigandMPNN does not warn when no ligand atoms are found; it just runs as if\n`--model_type protein_mpnn` had been picked. The two common ways this happens\nare an input PDB whose HETATM records were stripped by an upstream\nclean-up step, and `--parse_these_chains_only` naming the protein chains but\nnot the ligand's. If `ligand_confidence` in the FASTA header is missing or\nzero across every design, the model never saw the ligand — fix the input, do\nnot trust the sequences.\n\n## Errors worth recognizing\n\n| You see | It means \u002F do this |\n|---|---|\n| `ModuleNotFoundError: No module named 'tree'` | `pip install dm-tree` — the vendored openfold imports it unconditionally. |\n| `module 'numpy' has no attribute 'int'` | Run the `sed` patch on `openfold\u002Fnp\u002Fresidue_constants.py`, or pin `numpy\u003C1.24` (py≤3.11 only). |\n| `error: command 'clang' failed` while `pip install ProDy` | See the ProDy gotcha above — `apt_install(\"build-essential\")` and `env({\"CC\":\"gcc\",\"CXX\":\"g++\"})`. |\n| `FileNotFoundError` for `model_params\u002F...` | Checkpoints not fetched — run `bash get_model_params.sh .\u002Fmodel_params` from inside the clone. |\n\n---\n\n**Next:** fold the designs in complex with the ligand via `boltz` or `chai1`\n(both accept SMILES\u002FCCD) and filter on ipTM and ligand placement.\n",{"name":5,"license":15,"category":24,"metadata":25,"description":32},"biomodels",{"third_party":26,"display-name":29},[27],{"kind":28,"name":29,"license":30,"terms_url":31},"weights","LigandMPNN","MIT","https:\u002F\u002Fgithub.com\u002Fdauparas\u002FLigandMPNN\u002Fblob\u002Fmain\u002FLICENSE","Inverse-fold a backbone with ligand, nucleic-acid, and metal context using LigandMPNN (Dauparas et al. 2023, github.com\u002Fdauparas\u002FLigandMPNN). Reach for this skill to redesign the residues lining a binding pocket around a bound small molecule or cofactor, to design metal-coordinating sites where the geometry must be respected, or to get threaded designed-sequence PDBs out of any MPNN run.\n","main","e71245365df78e06cd166602744bf2491f79910e","research-skills-hub\u002Fclaude-science-skills\u002Fskills\u002Fligandmpnn",0,"https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS","https:\u002F\u002Fpengqianhan.github.io\u002FAI-Human-Research-OS\u002F","https:\u002F\u002Fgithub.com\u002Fpengqianhan","Ph.D. candidate of The University of Auckland, New Zealand","2026-11-14 18:08:16",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":54},[44,47,50],{"tool":45,"command":46},"npx","npx skills add pengqianhan\u002FAI-Human-Research-OS --skill ligandmpnn",{"tool":48,"command":49},"bunx","bunx skills add pengqianhan\u002FAI-Human-Research-OS --skill ligandmpnn",{"tool":51,"command":52},"pnpm","pnpm dlx skills add pengqianhan\u002FAI-Human-Research-OS --skill ligandmpnn","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fclaude-science-skills\u002Fskills\u002Fligandmpnn\u002FSKILL.md",true,[],[57,64,71,78,85,92],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M74MWYTHT498SVMN823896T7","workflow-skill-creator","Distills a completed user workflow or interaction into a reusable agent skill. Use when the user asks to turn their workflow, interaction, or multi-step process into a skill, or when they say \"make this a skill\", \"create a skill from what we just did\", \"package this workflow\" or similar. Do not use for creating skills from scratch without an existing workflow (use a generic skill-creator for that).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fworkflow_skill_creator\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M74MWWVCFZ8HQS8840F1V1FT","uv","Checks whether the uv Python package manager is installed and installs it if missing. Ensures uv is on PATH. Use when another skill requires uv as a prerequisite.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fuv\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M74MWRH5BTSWZ2C5PCTD5FV9","uniprot-database","Access protein metadata, function, taxonomy, and sequences across UniProtKB, UniParc, and UniRef. Use when searching for proteins, mapping identifiers, or retrieving functional annotations and publications. Don't use for sequence alignment, protein folding, or sequence similarity search (use specialized skills for those tasks).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Funiprot_database\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M74MWN1D3ER9F1FZPWH8HQC1","unibind-database","Queries the UniBind database for experimentally validated transcription factor (TF) binding sites. Use when retrieving direct TF-DNA interaction datasets, downloading binding site coordinates (BED\u002FFASTA) for local analysis, or listing available datasets by species, cell line, or TF name. Don't use to query specific intervals, locations, genes, motif models or expression data.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Funibind_database\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M74MWHHB5QZY3NR2FYA1EK4N","ucsc-conservation-and-tfbs","Fetch Evolutionary Conservation scores (phyloP, phastCons) and Transcription Factor Binding Sites (TFBS) from the UCSC Genome Browser. Use when analyzing whether genomic variants or regions are evolutionarily conserved, functionally important, or bounded by TF regulators across major projects (ENCODE, JASPAR, ReMap).","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fucsc_conservation_and_tfbs\u002FSKILL.md",[],[],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":30,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":19},"01M74MWERYHSY87P21FDMVE6BA","string-database","Query the STRING database for protein-protein interactions (PPIs), functional enrichment, and homology. Use when the user asks about interactions between specific proteins, interaction evidence, confidence scores, protein interaction partners, or pathway enrichments.","https:\u002F\u002Fgithub.com\u002Fpengqianhan\u002FAI-Human-Research-OS\u002Fblob\u002Fmain\u002Fresearch-skills-hub\u002Fscience-skills\u002Fstring_database\u002FSKILL.md",[],[]]