Skip to content

pynetworkin

DOI PyPI version PyPI - Python Version PyPI - Downloads Docker Image Version GitHub Release GitHub Actions License

Kinase–substrate network prediction — a modernised implementation of the NetworKIN algorithm.

Overview

pynetworkin predicts kinase–substrate relationships by combining:

  • Motif scoring via pynetphorest (NetPhoREST atlas)
  • Network context scoring using the STRING v12.0 protein interaction network
  • False-negative recovery to rescue pairs missed by motif scoring alone
  • Live data sources: OmniPath REST API and STRING API with local caching

Quick Start

pip install pynetworkin
pynetworkin predict input.fasta --output results.tsv

Or with uv:

uvx pynetworkin predict input.fasta -o results.tsv --refresh

Key Features

Feature Description
Motif scoring PyNetPhoREST atlas with >500 kinase models
Network context STRING v12.0 combined scores
FN recovery Context-proximity based recovery
Live data OmniPath REST, STRING API, local caching
CLI Typer + Rich progress bars and summary tables
Output TSV and Cytoscape SIF formats

Architecture

See ARCHITECTURE.md for full technical details.