ProDEC

A package to easily calculate descriptors of protein sequences and their common transforms (domain averages, auto-cross covariances, physicochemical distance transformations, and the fast Fourier transform).

Table of contents

• Installation
• Quickstart
• Documentation
• Citation
• Contributing
• License

📦 Installation

pip install prodec

Quickstart

from prodec import ProteinDescriptors, Transform, TransformType

sequence = 'MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTL'

# Load the catalog of available descriptors and pick one
pdescs = ProteinDescriptors()
zscales = pdescs.get_descriptor('Zscale Hellberg')

# Get raw per-residue descriptor values
raw_values = zscales.get(sequence)

# Apply a transform, e.g. domain averages over 5 domains
avg_zscale = Transform(TransformType.AVG, zscales)
avg_values = avg_zscale.get(sequence, domains=5)

Documentation

• docs/usage.md — full walkthrough, gap handling, non-standard amino acids, transform compatibility, and how to add new descriptors.
• docs/api.md — reference for ProteinDescriptors, Descriptor, Transform, and TransformType.

📖 Citation

If you use ProDEC in your work, please cite this repository:

Béquignon, O. J. M. ProDEC: a package to calculate protein sequence descriptors and their common transforms. https://github.com/OlivierBeq/ProDEC

The bundled descriptors and transforms themselves originate from the literature (e.g. Zscales by Hellberg et al., and the Raychaudhury descriptor) — see Descriptor.summary for the citation of a specific descriptor, and refer to the corresponding original publication when using it in your own research.

Contributing

Contributions are welcome — see CONTRIBUTING.md for how to set up a development environment and run the test/lint/type-check suite locally.

License

ProDEC is distributed under the MIT License.