foldseek
Original：🇺🇸 English
Translated
Structure similarity search with Foldseek. Use this skill when: (1) Finding similar structures in PDB/AFDB databases, (2) Structural homology search, (3) Database queries by 3D structure, (4) Finding remote homologs not detected by sequence, (5) Clustering structures by similarity. For sequence similarity, use uniprot BLAST. For structure prediction, use chai or boltz.
8installs
Added on
NPX Install
npx skill4agent add adaptyvbio/protein-design-skills foldseekTags
Translated version includes tags in frontmatterSKILL.md Content
View Translation Comparison →Foldseek Structure Search
Prerequisites
| Requirement | Minimum | Recommended |
|---|---|---|
| Python | 3.8+ | 3.10 |
| RAM | 8GB | 16GB |
| Disk | 10GB | 50GB (for local databases) |
How to run
Note: Foldseek can run locally or via web server. No GPU required.
Option 1: Web Server (Quick; rate-limited, use sparingly)
bash
# Upload structure to web server
curl -X POST "https://search.foldseek.com/api/ticket" \
-F "q=@query.pdb" \
-F "database[]=afdb50" \
-F "database[]=pdb100"Option 2: Local installation
bash
# Install Foldseek
conda install -c conda-forge -c bioconda foldseek
# Search PDB
foldseek easy-search query.pdb /path/to/pdb100 results.m8 tmp/
# Search AlphaFold DB
foldseek easy-search query.pdb /path/to/afdb50 results.m8 tmp/Option 3: Python API
python
import subprocess
import pandas as pd
def foldseek_search(query_pdb, database, output="results.m8"):
"""Run Foldseek search."""
subprocess.run([
"foldseek", "easy-search",
query_pdb, database, output, "tmp/",
"--format-output", "query,target,pident,alnlen,evalue,bits"
])
return pd.read_csv(output, sep="\t",
names=["query", "target", "pident", "alnlen", "evalue", "bits"])Key parameters
| Parameter | Default | Description |
|---|---|---|
| 0.0 | Minimum sequence identity |
| 0.001 | E-value threshold |
| 2 | 0=3Di, 1=TM, 2=3Di+AA |
| 300 | Max hits to pass through prefilter; reducing this affects sensitivity |
Databases
| Database | Description | Size |
|---|---|---|
| PDB clustered at 100% | ~200K structures |
| AlphaFold DB at 50% | ~67M structures |
| SwissProt structures | ~500K structures |
| CATH domains | ~50K domains |
Output format
# results.m8 (tabular)
query target pident alnlen evalue bits
query 1abc_A 85.2 120 1e-45 180.5
query 2def_B 72.1 115 1e-32 145.2Sample output
Successful run
$ foldseek easy-search query.pdb pdb100 results.m8 tmp/
[INFO] Loading database: pdb100 (194,527 entries)
[INFO] Searching...
[INFO] Found 127 hits
Top 5 hits:
1. 1abc_A - 85.2% identity, E=1e-45
2. 2def_B - 72.1% identity, E=1e-32
3. 3ghi_C - 68.5% identity, E=1e-28
4. 4jkl_A - 55.3% identity, E=1e-18
5. 5mno_B - 42.1% identity, E=1e-10Decision tree
Should I use Foldseek?
│
├─ What are you searching?
│ ├─ By 3D structure → Foldseek ✓
│ ├─ By sequence → Use BLAST (uniprot skill)
│ └─ Both → Run both, compare results
│
└─ What do you need?
├─ Find structural homologs → Foldseek ✓
├─ Remote homolog detection → Foldseek ✓
├─ Structural clustering → Foldseek ✓
└─ Functional annotation → Cross-reference with UniProtCommon use cases
Find similar designs
bash
# Compare your design to PDB
foldseek easy-search design.pdb pdb100 similar_natural.m8 tmp/Novelty check
bash
# Ensure design is novel (low similarity to known)
foldseek easy-search design.pdb afdb50 novelty.m8 tmp/
# Novel if: top hit identity < 30%Scaffold search
bash
# Find scaffolds for motif grafting
foldseek easy-search motif.pdb pdb100 scaffolds.m8 tmp/ \
--min-seq-id 0.0 -e 10Verify
bash
wc -l results.m8 # Number of hitsTroubleshooting
No hits: Lower e-value threshold, try larger database
Too many hits: Increase min-seq-id threshold
Slow search: Use smaller database
Error interpretation
| Error | Cause | Fix |
|---|---|---|
| Wrong path | Check database location |
| Malformed structure | Validate PDB format |
| Large database | Use more RAM or web server |
Next: Download hits with skill → use for scaffold design.
pdb