BiomeAI

Simple Python package for Multiple Sequence Alignment search using NVIDIA's ColabFold.

Install

pip install biomeai

Quick Usage

from biomeai import ColabFoldMSA

# Initialize with API key
msa = ColabFoldMSA(api_key="your-api-key")
# Or set NVCF_RUN_KEY environment variable

# Your protein sequence
sequence = "MVPSAGQLALFALGIVLAACQALENSTSPLSADPPVAAAVVSHFNDCPDSHTQFCFHGTCRFL"

# 1. Use a single database
results = msa.search_sync(
    sequence=sequence,
    databases=["Uniref30_2302"]
)

# 2. Use multiple databases
results = msa.search_sync(
    sequence=sequence,
    databases=["Uniref30_2302", "PDB70_220313"]
)

# 3. Use all databases (Cascaded Search)
results = msa.search_sync(
    sequence=sequence,
    databases=["Uniref30_2302", "PDB70_220313", "colabfold_envdb_202108"]
)

# Save results (optional)
msa.save_results(results, "results.json")

Available Databases

• Uniref30_2302: Universal Reference Database
• PDB70_220313: Protein Data Bank
• colabfold_envdb_202108: Environmental Database

Additional Options

results = msa.search_sync(
    sequence=sequence,
    databases=["Uniref30_2302"],
    e_value=0.0001,        # E-value threshold
    iterations=1,          # Number of iterations
    output_formats=["a3m", "fasta"]  # Output formats
)

Async Support

import asyncio

async def search_msa():
    results = await msa.search(
        sequence=sequence,
        databases=["Uniref30_2302"]
    )
    return results

results = asyncio.run(search_msa())

Development

Install with development dependencies:

pip install biomeai[dev]

Run tests:

pytest

License

MIT License - see the LICENSE file for details.