AlphaFoldFetch

AlphaFoldFetch is a small command-line tool for downloading AlphaFold structure files from UniProt accessions or UniProt-style FASTA files.

It is built for quick one-off downloads and simple batch workflows:

• pass one ID, many IDs, FASTA files, or stdin
• download PDB, CIF, or both
• optionally gzip the saved files
• tune concurrency for larger jobs

Benchmark

Local benchmark against the AlphaFold Homo sapiens proteome with 47,172 gzipped structures (pdb + cif).

Install

Run once with uvx:

uvx --from AlphaFoldFetch affetch P11388

Or install the affetch command:

uv tool install AlphaFoldFetch

AlphaFoldFetch requires Python 3.11 or newer.

Quick Start

Download the default outputs for one UniProt accession:

affetch P11388

Download several accessions:

affetch P11388 Q01320 P41516

Write files to an existing directory:

mkdir -p structures
affetch -o structures P11388

Download an uncompressed PDB only:

affetch -f p P11388

Read IDs from stdin:

printf "P11388\nQ01320\n" | affetch -

Usage

affetch [OPTIONS] UNIPROT...

Arguments:

• UNIPROT...: UniProt IDs, FASTA files, or - for stdin

Common options:

• --output, -o: output directory, default: current directory
• --file-type, -f: any combination of p, c, and z, default: pcz
• --model, -m: AlphaFold model version, default: 6
• --n-sync: concurrent download requests, default: 50
• --n-save: file writes submitted per batch, default: 500

Input Rules

AlphaFoldFetch accepts:

• UniProt accessions like P11388
• strings that contain a valid UniProt accession
• FASTA files ending in .fasta, .fas, .fa, or .faa
• - to read whitespace-separated input from stdin

FASTA parsing only keeps validated UniProt IDs from header lines.

Output Rules

The --file-type option is a compact set of letters:

• p: save PDB files
• c: save CIF files
• z: gzip the selected outputs

The default value is pcz, so affetch P11388 downloads gzipped PDB and CIF files for AlphaFold model 6:

AF-P11388-F1-model_v6.pdb.gz
AF-P11388-F1-model_v6.cif.gz

Examples:

• -f p: uncompressed PDB only
• -f c: uncompressed CIF only
• -f pc: uncompressed PDB and CIF
• -f pz: gzipped PDB only
• -f cz: gzipped CIF only

The output directory must already exist. IDs without an available AlphaFold file are skipped.

FASTA Files

Download structures from one or more UniProt FASTA files:

affetch UP000005640_9606.fasta
affetch plant_pgks.fasta mammalian_pgks.fasta bacterial_pgks.fasta

Only header lines are scanned for UniProt accessions.

Pipelines

From an AlphaFold DB search results CSV:

tail -n +2 results-csv.csv | while IFS='-' read -r f1 f2 f3; do echo $f2; done | affetch -

From getSequence:

getseq human top2a, mouse top2a, rat top2a | affetch -

Remember to pass - when reading IDs from stdin.

More Information

• Documentation: https://mansanlab.github.io/alphafoldfetch/
• Issues: https://github.com/mansanlab/alphafoldfetch/issues

Credits

Inspired by getSequence.