phospy 1.4.0

Python native implementation of selected PhosR style workflows for phosphoproteomics.

PhosPy

PhosPy is a Python package for a small, supported set of phosphoproteomics workflows inspired by PhosR.

It is built for three main jobs:

• Preprocessing raw total and phospho tables with PhosphoDataset
• Kinase scoring and prediction with SimpleKinaseWorkflow
• Signalome analysis with SignalomeWorkflow

PhosPy is intentionally narrow. It is not a full PhosR replacement.

Install

PhosPy supports Python 3.10 and newer.

pip install phospy

Optional parquet support:

pip install "phospy[parquet]"

Start Here

Use PhosphoDataset when you want preprocessing only

from phospy.datasets import PhosphoDataset
from phospy.preprocessing import CorePreprocessingConfig

dataset = PhosphoDataset.from_files(
    "examples/data/total.tsv",
    "examples/data/phospho.tsv",
    phospho_encoding="utf-16le",
)

analysis_ready = dataset.preprocessing.run_analysis_ready(
    config=CorePreprocessingConfig(max_unmatched_fraction=0.1)
)

You get a validated, analysis-ready dataset that you can pass into downstream workflows.

Use SimpleKinaseWorkflow for the common end-to-end path

from phospy.api import PredictionRunConfig, SimpleKinaseWorkflow

with SimpleKinaseWorkflow().run(
    phospho="examples/data/simple_workflow/phospho.tsv",
    total="examples/data/simple_workflow/total.tsv",
    species="rat",
    prediction_config=PredictionRunConfig(
        min_substrates=1,
        min_motif_size=1,
        ensemble_size=2,
        top=3,
        inclusion=2,
        n_iterations=2,
        random_state=7,
    ),
) as result:
    pred_mat = result.pred_mat_result.to_frame(copy=False)
    weighted_activity = result.kinase_activity_result.weighted_activity

This path handles preprocessing, analysis-ready adaptation, bundled reference selection, prediction, and kinase activity analysis.

Use SignalomeWorkflow when you already have workflow outputs

The next example assumes you already have a result from SimpleKinaseWorkflow.run(...).

from phospy.api import SignalomeRunConfig, SignalomeWorkflow

signalome = SignalomeWorkflow().run_from_analysis_ready(
    dataset=result.analysis_ready_dataset,
    scoring_result=result.scoring_result,
    prediction_result=result.prediction_result,
    kinases_of_interest=list(result.pred_mat_result.kinase_names[:2]),
    config=SignalomeRunConfig(signalome_cutoff=0.5),
)

Minimum Input Expectations

PhosPy works with:

• total input as TSV
• phospho input as TSV
• predMat as CSV, with the first column used as the phosphosite index

Default required columns:

Input	Required columns
Total table	genes, group1 to group6
Phospho table	uid, gene_names, gene_p_site, localization_prob, centralized_sequence, p_group1 to p_group6

For full validation rules and common failure cases, see docs/validation.md.

CLI

PhosPy ships with a small CLI for file-based preprocessing and optional kinase activity analysis from an existing predMat.

phospy \
  --total examples/data/total.tsv \
  --phospho examples/data/phospho.tsv \
  --pred-mat examples/data/predMat.csv \
  --phospho-encoding utf-16le \
  --max-unmatched-fraction 0.1 \
  --outdir examples/output

Use phospy --help to see every option.

Runnable Examples

• examples/simple_workflow_demo.py
• examples/signalome_workflow_demo.py
• examples/signalome_map_demo.py
• examples/kinase_activity_analyzer_demo.py

Read Next

Beginner guides

• docs/api.md for the supported Python API
• docs/validation.md for input rules and common failures

Advanced and contributor docs

• docs/parity.md for parity scope and svm_mode
• docs/fixtures.md for fixture and trace rebuild commands
• docs/architecture/package-layout.md for package ownership and contributor layout rules
• docs/roadmap.md for likely next areas of work