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).
pip install prodecfrom 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)docs/usage.md— full walkthrough, gap handling, non-standard amino acids, transform compatibility, and how to add new descriptors.docs/api.md— reference forProteinDescriptors,Descriptor,Transform, andTransformType.
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.
Contributions are welcome — see CONTRIBUTING.md for how to set up a
development environment and run the test/lint/type-check suite locally.
ProDEC is distributed under the MIT License.