[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DPWGJ6HPX4M5GC3VGHPFDM":3,"skill-same:01M6DPWGJ6HPX4M5GC3VGHPFDM":75},{"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":31,"publishedAt":49,"linkOnly":50,"markdownContent":51,"markdownBody":52,"frontMatter":53,"refName":56,"commitSha":57,"skillDirectory":5,"forksCount":14,"repositoryHtmlUrl":58,"repositoryHomepageUrl":6,"ownerHtmlUrl":59,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":60,"install":61,"resources":74},"01M6DPWGJ6HPX4M5GC3VGHPFDM","protein-design-workflow",null,"End-to-end guidance for protein design pipelines. Use this skill when: (1) Starting a new protein design project, (2) Need step-by-step workflow guidance, (3) Understanding the full design pipeline, (4) Planning compute resources and timelines, (5) Integrating multiple design tools.\nFor tool selection, use binder-design. For QC thresholds, use protein-qc.","huamu668\u002Fclawhub-openclaw-medical","huamu668","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F202473531?v=4","01M6DK1V7D7VP9PJ06Z1VXVK13","huamu","01M6DK1VBYG1BX5FT154AFZHAS",0,"MIT","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fprotein-design-workflow\u002FSKILL.md",[18,23,26],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",{"categoryNo":27,"categoryName":28,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"protein design workflow","workflow","design","pipelines","binder","end","full","project","compute","selection","guidance","use","tool","Tools","pipeline","planning","orchestration","2026-11-10 10:12:30",false,"---\nname: protein-design-workflow\ndescription: >\n  End-to-end guidance for protein design pipelines.\n  Use this skill when: (1) Starting a new protein design project,\n  (2) Need step-by-step workflow guidance,\n  (3) Understanding the full design pipeline,\n  (4) Planning compute resources and timelines,\n  (5) Integrating multiple design tools.\n\n  For tool selection, use binder-design.\n  For QC thresholds, use protein-qc.\nlicense: MIT\ncategory: orchestration\ntags: [guidance, pipeline, workflow]\n---\n\n# Protein Design Workflow Guide\n\n## Standard binder design pipeline\n\n### Overview\n```\nTarget Preparation --> Backbone Generation --> Sequence Design\n         |                     |                     |\n         v                     v                     v\n    (pdb skill)          (rfdiffusion)         (proteinmpnn)\n                               |                     |\n                               v                     v\n                        Structure Validation --> Filtering\n                               |                     |\n                               v                     v\n                         (alphafold\u002Fchai)      (protein-qc)\n```\n\n## Phase 1: Target preparation\n\n### 1.1 Obtain target structure\n```bash\n# Download from PDB\ncurl -o target.pdb \"https:\u002F\u002Ffiles.rcsb.org\u002Fdownload\u002FXXXX.pdb\"\n```\n\n### 1.2 Clean and prepare\n```python\n# Extract target chain\n# Remove waters, ligands if needed\n# Trim to binding region + 10A buffer\n```\n\n### 1.3 Select hotspots\n- Choose 3-6 exposed residues\n- Prefer charged\u002Faromatic (K, R, E, D, W, Y, F)\n- Check surface accessibility\n- Verify residue numbering\n\n**Output**: `target_prepared.pdb`, hotspot list\n\n## Phase 2: Backbone generation\n\n### Option A: RFdiffusion (diverse exploration)\n```bash\nmodal run modal_rfdiffusion.py \\\n  --pdb target_prepared.pdb \\\n  --contigs \"A1-150\u002F0 70-100\" \\\n  --hotspot \"A45,A67,A89\" \\\n  --num-designs 500\n```\n\n### Option B: BindCraft (end-to-end)\n```bash\nmodal run modal_bindcraft.py \\\n  --target-pdb target_prepared.pdb \\\n  --hotspots \"A45,A67,A89\" \\\n  --num-designs 100\n```\n\n**Output**: 100-500 backbone PDBs\n\n## Phase 3: Sequence design\n\n### For RFdiffusion backbones\n```bash\nfor backbone in backbones\u002F*.pdb; do\n  modal run modal_proteinmpnn.py \\\n    --pdb-path \"$backbone\" \\\n    --num-seq-per-target 8 \\\n    --sampling-temp 0.1\ndone\n```\n\n**Output**: 8 sequences per backbone (800-4000 total)\n\n## Phase 4: Structure validation\n\n### Predict complexes\n```bash\n# Prepare FASTA with binder + target\n# binder:target format for multimer\n\nmodal run modal_colabfold.py \\\n  --input-faa all_sequences.fasta \\\n  --out-dir predictions\u002F\n```\n\n**Output**: AF2 predictions with pLDDT, ipTM, PAE\n\n## Phase 5: Filtering and selection\n\n### Apply standard thresholds\n```python\nimport pandas as pd\n\n# Load metrics\ndesigns = pd.read_csv('all_metrics.csv')\n\n# Filter\nfiltered = designs[\n    (designs['pLDDT'] > 0.85) &\n    (designs['ipTM'] > 0.50) &\n    (designs['PAE_interface'] \u003C 10) &\n    (designs['scRMSD'] \u003C 2.0) &\n    (designs['esm2_pll'] > 0.0)\n]\n\n# Rank by composite score\nfiltered['score'] = (\n    0.3 * filtered['pLDDT'] +\n    0.3 * filtered['ipTM'] +\n    0.2 * (1 - filtered['PAE_interface'] \u002F 20) +\n    0.2 * filtered['esm2_pll']\n)\n\ntop_designs = filtered.nlargest(50, 'score')\n```\n\n**Output**: 50-200 filtered candidates\n\n## Resource planning\n\n### Compute requirements\n\n| Stage | GPU | Time (100 designs) |\n|-------|-----|-------------------|\n| RFdiffusion | A10G | 30 min |\n| ProteinMPNN | T4 | 15 min |\n| ColabFold | A100 | 4-8 hours |\n| Filtering | CPU | 15 min |\n\n### Total timeline\n- Small campaign (100 designs): 8-12 hours\n- Medium campaign (500 designs): 24-48 hours\n- Large campaign (1000+ designs): 2-5 days\n\n## Quality checkpoints\n\n### After backbone generation\n- [ ] Visual inspection of diverse backbones\n- [ ] Secondary structure present\n- [ ] No clashes with target\n\n### After sequence design\n- [ ] ESM2 PLL > 0.0 for most sequences\n- [ ] No unwanted cysteines (unless intentional)\n- [ ] Reasonable sequence diversity\n\n### After validation\n- [ ] pLDDT > 0.85\n- [ ] ipTM > 0.50\n- [ ] PAE_interface \u003C 10\n- [ ] Self-consistency RMSD \u003C 2.0 A\n\n### Final selection\n- [ ] Diverse sequences (cluster if needed)\n- [ ] Manufacturable (no problematic motifs)\n- [ ] Reasonable molecular weight\n\n## Common issues\n\n| Problem | Solution |\n|---------|----------|\n| Low ipTM | Check hotspots, increase designs |\n| Poor diversity | Higher temperature, more backbones |\n| High scRMSD | Backbone may be unusual |\n| Low pLDDT | Check design quality |\n\n## Advanced workflows\n\n### Multi-tool combination\n1. RFdiffusion for initial backbones\n2. ColabDesign for refinement\n3. ProteinMPNN diversification\n4. AF2 final validation\n\n### Iterative refinement\n1. Run initial campaign\n2. Analyze failures\n3. Adjust hotspots\u002Fparameters\n4. Repeat with insights\n","# Protein Design Workflow Guide\n\n## Standard binder design pipeline\n\n### Overview\n```\nTarget Preparation --> Backbone Generation --> Sequence Design\n         |                     |                     |\n         v                     v                     v\n    (pdb skill)          (rfdiffusion)         (proteinmpnn)\n                               |                     |\n                               v                     v\n                        Structure Validation --> Filtering\n                               |                     |\n                               v                     v\n                         (alphafold\u002Fchai)      (protein-qc)\n```\n\n## Phase 1: Target preparation\n\n### 1.1 Obtain target structure\n```bash\n# Download from PDB\ncurl -o target.pdb \"https:\u002F\u002Ffiles.rcsb.org\u002Fdownload\u002FXXXX.pdb\"\n```\n\n### 1.2 Clean and prepare\n```python\n# Extract target chain\n# Remove waters, ligands if needed\n# Trim to binding region + 10A buffer\n```\n\n### 1.3 Select hotspots\n- Choose 3-6 exposed residues\n- Prefer charged\u002Faromatic (K, R, E, D, W, Y, F)\n- Check surface accessibility\n- Verify residue numbering\n\n**Output**: `target_prepared.pdb`, hotspot list\n\n## Phase 2: Backbone generation\n\n### Option A: RFdiffusion (diverse exploration)\n```bash\nmodal run modal_rfdiffusion.py \\\n  --pdb target_prepared.pdb \\\n  --contigs \"A1-150\u002F0 70-100\" \\\n  --hotspot \"A45,A67,A89\" \\\n  --num-designs 500\n```\n\n### Option B: BindCraft (end-to-end)\n```bash\nmodal run modal_bindcraft.py \\\n  --target-pdb target_prepared.pdb \\\n  --hotspots \"A45,A67,A89\" \\\n  --num-designs 100\n```\n\n**Output**: 100-500 backbone PDBs\n\n## Phase 3: Sequence design\n\n### For RFdiffusion backbones\n```bash\nfor backbone in backbones\u002F*.pdb; do\n  modal run modal_proteinmpnn.py \\\n    --pdb-path \"$backbone\" \\\n    --num-seq-per-target 8 \\\n    --sampling-temp 0.1\ndone\n```\n\n**Output**: 8 sequences per backbone (800-4000 total)\n\n## Phase 4: Structure validation\n\n### Predict complexes\n```bash\n# Prepare FASTA with binder + target\n# binder:target format for multimer\n\nmodal run modal_colabfold.py \\\n  --input-faa all_sequences.fasta \\\n  --out-dir predictions\u002F\n```\n\n**Output**: AF2 predictions with pLDDT, ipTM, PAE\n\n## Phase 5: Filtering and selection\n\n### Apply standard thresholds\n```python\nimport pandas as pd\n\n# Load metrics\ndesigns = pd.read_csv('all_metrics.csv')\n\n# Filter\nfiltered = designs[\n    (designs['pLDDT'] > 0.85) &\n    (designs['ipTM'] > 0.50) &\n    (designs['PAE_interface'] \u003C 10) &\n    (designs['scRMSD'] \u003C 2.0) &\n    (designs['esm2_pll'] > 0.0)\n]\n\n# Rank by composite score\nfiltered['score'] = (\n    0.3 * filtered['pLDDT'] +\n    0.3 * filtered['ipTM'] +\n    0.2 * (1 - filtered['PAE_interface'] \u002F 20) +\n    0.2 * filtered['esm2_pll']\n)\n\ntop_designs = filtered.nlargest(50, 'score')\n```\n\n**Output**: 50-200 filtered candidates\n\n## Resource planning\n\n### Compute requirements\n\n| Stage | GPU | Time (100 designs) |\n|-------|-----|-------------------|\n| RFdiffusion | A10G | 30 min |\n| ProteinMPNN | T4 | 15 min |\n| ColabFold | A100 | 4-8 hours |\n| Filtering | CPU | 15 min |\n\n### Total timeline\n- Small campaign (100 designs): 8-12 hours\n- Medium campaign (500 designs): 24-48 hours\n- Large campaign (1000+ designs): 2-5 days\n\n## Quality checkpoints\n\n### After backbone generation\n- [ ] Visual inspection of diverse backbones\n- [ ] Secondary structure present\n- [ ] No clashes with target\n\n### After sequence design\n- [ ] ESM2 PLL > 0.0 for most sequences\n- [ ] No unwanted cysteines (unless intentional)\n- [ ] Reasonable sequence diversity\n\n### After validation\n- [ ] pLDDT > 0.85\n- [ ] ipTM > 0.50\n- [ ] PAE_interface \u003C 10\n- [ ] Self-consistency RMSD \u003C 2.0 A\n\n### Final selection\n- [ ] Diverse sequences (cluster if needed)\n- [ ] Manufacturable (no problematic motifs)\n- [ ] Reasonable molecular weight\n\n## Common issues\n\n| Problem | Solution |\n|---------|----------|\n| Low ipTM | Check hotspots, increase designs |\n| Poor diversity | Higher temperature, more backbones |\n| High scRMSD | Backbone may be unusual |\n| Low pLDDT | Check design quality |\n\n## Advanced workflows\n\n### Multi-tool combination\n1. RFdiffusion for initial backbones\n2. ColabDesign for refinement\n3. ProteinMPNN diversification\n4. AF2 final validation\n\n### Iterative refinement\n1. Run initial campaign\n2. Analyze failures\n3. Adjust hotspots\u002Fparameters\n4. Repeat with insights\n",{"name":5,"tags":54,"license":15,"category":48,"description":55},[42,46,33],"End-to-end guidance for protein design pipelines. Use this skill when: (1) Starting a new protein design project, (2) Need step-by-step workflow guidance, (3) Understanding the full design pipeline, (4) Planning compute resources and timelines, (5) Integrating multiple design tools.\nFor tool selection, use binder-design. For QC thresholds, use protein-qc.\n","main","d3998984163a67a891ee8397e35adf79b14e872d","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical","https:\u002F\u002Fgithub.com\u002Fhuamu668","2026-03-12 11:49:42",{"source":8,"skillName":5,"commands":62,"prompt":72,"zipAvailable":73},[63,66,69],{"tool":64,"command":65},"npx","npx skills add huamu668\u002Fclawhub-openclaw-medical --skill protein-design-workflow",{"tool":67,"command":68},"bunx","bunx skills add huamu668\u002Fclawhub-openclaw-medical --skill protein-design-workflow",{"tool":70,"command":71},"pnpm","pnpm dlx skills add huamu668\u002Fclawhub-openclaw-medical --skill protein-design-workflow","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fprotein-design-workflow\u002FSKILL.md",true,[],[76,95,123,143,157,178],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":49},"01M6DR3JGAEX5G5FR7TK3D8NS8","zarr-python","Chunked N-D arrays for cloud storage. Compressed arrays, parallel I\u002FO, S3\u002FGCS integration, NumPy\u002FDask\u002FXarray compatible, for large-scale scientific computing pipelines.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fzarr-python\u002FSKILL.md",[],[83,84,85,86,87,88,35,89,90,91,92,93,94],"python","zarr python","storage","scientific computing","numpy","arrays","parallel","dask","cloud","gcs","s3","integration",{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":104,"publishedAt":49},"01M6DR101MEDE5XT2ZZ5VBEAA5","vaex","Use this skill for processing and analyzing large tabular datasets (billions of rows) that exceed available RAM. Vaex excels at out-of-core DataFrame operations, lazy evaluation, fast aggregations, efficient visualization of big data, and machine learning on large datasets. Apply when users need to work with large CSV\u002FHDF5\u002FArrow\u002FParquet files, perform fast statistics on massive datasets, create visualizations of big data, or build ML pipelines that do not fit in memory.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fvaex\u002FSKILL.md",[101],{"categoryNo":102,"categoryName":103,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化",[97,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,43,121,122],"machine learning","in memory","learning","csv","visualization","create","core","ml","big data","memory","visualizations","fit","statistics","do","build","files","operations","work",{"skillNo":124,"name":125,"nameZh":6,"description":126,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":49},"01M6DQVMB0GBKF2W44FKXVD1P7","torch-geometric","Graph Neural Networks (PyG). Node\u002Fgraph classification, link prediction, GCN, GAT, GraphSAGE, heterogeneous graphs, molecular property prediction, for geometric deep learning.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Ftorch-geometric\u002FSKILL.md",[],[130,107,131,132,133,134,135,136,137,138,139,140,141,142],"torch geometric","node","Neural","link","gat","molecular","graphs","deep","networks","graph","neural networks","deep learning","classification",{"skillNo":144,"name":145,"nameZh":6,"description":146,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":147,"categories":148,"keywords":149,"publishedAt":49},"01M6DQH2GZPZKEQ140GRCJ9746","tiledbvcf","Efficient storage and retrieval of genomic variant data using TileDB. Scalable VCF\u002FBCF ingestion, incremental sample addition, compressed storage, parallel queries, and export capabilities for population genomics.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Ftiledbvcf\u002FSKILL.md",[],[145,150,151,152,153,154,155,89,156,85],"data","genomics","sample","queries","ingestion","export","capabilities",{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":161,"categories":162,"keywords":165,"publishedAt":49},"01M6DQD9G958EM6NHJ994D0SK1","solublempnn","Solubility-optimized protein sequence design using SolubleMPNN. Use this skill when: (1) Designing for E. coli expression, (2) Optimizing solubility of designed proteins, (3) Reducing aggregation propensity, (4) Need high-yield expression, (5) Avoiding inclusion body formation.\nFor standard design, use proteinmpnn. For ligand-aware design, use ligandmpnn.","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fsolublempnn\u002FSKILL.md",[163,164],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":27,"categoryName":28,"taxonomyType":29,"taxonomyTypeDesc":30},[159,166,167,43,168,45,169,34,170,171,172,173,174,175,176,177],"inclusion","yield","modal","aggregation","proteins","solubility","proteinmpnn","ligandmpnn","ligand-aware","inverse-folding","expression","sequence-design",{"skillNo":179,"name":180,"nameZh":6,"description":181,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":182,"licenseFileUrl":6,"sourceHtmlUrl":183,"categories":184,"keywords":185,"publishedAt":49},"01M6DQ8W91A93MGGCEEW949HBC","scvelo","RNA velocity analysis with scVelo. Estimate cell state transitions from unspliced\u002Fspliced mRNA dynamics, infer trajectory directions, compute latent time, and identify driver genes in single-cell RNA-seq data. Complements Scanpy\u002FscVI-tools for trajectory inference.","BSD-3-Clause","https:\u002F\u002Fgithub.com\u002Fhuamu668\u002Fclawhub-openclaw-medical\u002Fblob\u002Fmain\u002Fscvelo\u002FSKILL.md",[],[180,186,187,188,189,190,191,40,192,193,194,45,150,195,196],"RNA","single-cell","infer","cell","transitions","time","velocity","scanpy","analysis","state","inference"]