ai4protein Biopython
Overview
Library skill for parsing and analysis. Prefer hub download tools when fetching remote data; use Biopython to parse the saved files.
Project Tools to prefer first
| Goal | Hub tool / skill |
|---|---|
| Read FASTA | read_fasta |
| PDB → sequences | pdb_chain_sequences, get_seq_from_pdb_chain_a, pdb_dir_to_fasta |
| NCBI sequence/BLAST | download_ncbi_sequence, download_ncbi_blast (ncbi_sequence) |
| UniProt | uniprot_database tools |
| AlphaFold structure | alphafold_database tools |
| Custom parse/plot | agent_generated_code + this skill |
Module map (on-demand references)
| Topic | Load with relative_path |
|---|---|
| SeqIO / sequences | references/sequence_io.md |
| Alignments | references/alignment.md or references/legacy_guide.md |
| Entrez / NCBI | references/databases.md — set email + NCBI_API_KEY |
| Bio.PDB | references/structure.md or references/legacy_guide.md |
| Phylogenetics | references/phylogenetics.md |
| Full cookbook | references/legacy_guide.md |
read_skill(skill_id="biopython", relative_path="references/databases.md")
Entrez setup (when needed)
from Bio import Entrez
import os
Entrez.email = "you@example.com"
if k := os.environ.get("NCBI_API_KEY"):
Entrez.api_key = k
Common mistakes
- Pulling huge GenBank XML into chat instead of
download_ncbi_* - Reimplementing AlphaFold download with Biopython instead of hub tools
- Missing Entrez email (NCBI requirement)
References
- See Module map above; full cookbook:
references/legacy_guide.md