alphafold-database
Access AlphaFold protein structures by UniProt ID
Also available from: davila7
Researchers need efficient access to computational protein structure predictions for drug discovery and structural biology studies. This skill provides direct access to AlphaFold DB's 200M+ AI-predicted protein structures, enabling retrieval by UniProt ID, download of coordinate files, and analysis of confidence metrics.
Download the skill ZIP
Upload in Claude
Go to Settings → Capabilities → Skills → Upload skill
Toggle on and start using
Test it
Using "alphafold-database". Download the AlphaFold structure for P00520 and analyze confidence
Expected outcome:
- AlphaFold ID: AF-P00520-F1
- Protein: Tyrosine-protein kinase ABL1 (Human)
- Sequence length: 1130 residues
- pLDDT Analysis:
- - Very high confidence (>90): 67% of residues
- - High confidence (70-90): 18% of residues
- Structure saved to: ./structures/AF-P00520-F1-model_v4.cif
Using "alphafold-database". Download E. coli proteome using Google Cloud
Expected outcome:
- taxonomy ID: 83333
- Downloading from: gs://public-datasets-deepmind-alphafold-v4/proteomes/
- Files matched: 4123
- Downloading proteome-tax_id-83333-*.tar (45 GB total)
- Progress: 45.2 GB / 45.2 GB (100%)
- Extracted 4123 structure archives to ./proteomes/
Security Audit
SafeThis is a legitimate scientific skill for accessing the AlphaFold protein structure database. All 244 static findings are false positives. The analyzer misinterpreted markdown code formatting (backticks), standard Python HTTP library usage, and documented public API endpoints as security threats. The skill uses safe Biopython library calls, standard requests to authorized EBI APIs, and subprocess with list-form arguments for Google Cloud access.
Risk Factors
🌐 Network access (2)
⚙️ External commands (1)
📁 Filesystem access (1)
Quality Score
What You Can Build
Retrieve protein structures for docking
Download target protein structures for computational docking studies and analyze binding site conformations.
Analyze prediction confidence
Evaluate pLDDT and PAE metrics to identify reliable structural regions for downstream analysis.
Build automated pipelines
Integrate AlphaFold access into computational workflows for large-scale protein analysis.
Try These Prompts
Download the AlphaFold structure for UniProt ID P00520 in mmCIF format and show the pLDDT confidence scores.
Download structures for P00520, P12931, and P04637. Compare their average pLDDT scores and identify high-confidence regions.
Download all AlphaFold predictions for E. coli (taxonomy ID 83333) using Google Cloud bulk access.
Create a Python script that takes a list of UniProt IDs, downloads their structures, extracts CA coordinates, and calculates inter-residue distances.
Best Practices
- Use Biopython for simple single-protein access (cleaner API than direct HTTP calls)
- Cache downloaded files locally to avoid repeated API requests and rate limits
- For bulk downloads over 100 proteins, use Google Cloud Storage instead of REST API
Avoid
- Avoid using shell=True with subprocess when calling gsutil (use list form instead)
- Do not ignore pLDDT scores when interpreting structures (low confidence regions may be unreliable)
- Avoid downloading individual files for whole proteomes (use tar archives from Google Cloud)