biocentral_api - Python Client

Python API package for biocentral.

Installation

pip install biocentral-api

Basic Usage

Embedding protein sequences:

from biocentral_api import BiocentralAPI, CommonEmbedder

biocentral_api = BiocentralAPI()

# ProtT5
embedder_name = CommonEmbedder.ProtT5
reduce = True
sequence_data = {"Seq1": "MMALSLALM"}
result = biocentral_api.embed(embedder_name=embedder_name, reduce=reduce, sequence_data=sequence_data,
                              use_half_precision=False).run()
print(result)

For more examples, please refer to the examples folder.

Service Limitations

The following limitations are currently in place:

• Sequence Limit: The embed() API has a limit of 1,000 sequences per request. See the amylase example for how to deal with datasets that exceed this limit.
• Rate Limiting: Most endpoints have a quite strict rate limiting to distribute the available resources as fair as possible. All affected endpoints have a built-in retry mechanism to handle the rate limiting.
• 431 Forbidden: Sometimes, the DDoS Protection of the API will send a mistaken 431 forbidden response. If you encounter this, please just try to re-run your script.

Citation

Please cite our paper if you are using the biocentral API in your work:

@Article{Franz2026,
  author    = {Franz, Sebastian and Olenyi, Tobias and Schloetermann, Paula and Smaoui, Amine and Jimenez-Soto, Luisa F. and Rost, Burkhard},
  journal   = {Journal of Molecular Biology},
  title     = {biocentral: embedding-based protein predictions},
  year      = {2026},
  issn      = {0022-2836},
  month     = jan,
  pages     = {169673},
  doi       = {10.1016/j.jmb.2026.169673},
  groups    = {[JMB] biocentral: embedding-based protein predictions, swc_bo_engineering},
  publisher = {Elsevier BV},
}