BioProcess-miR

Local, reproducible biological knowledge base (BioProcessDB) and disease-agnostic process → gene → microRNA analysis (BioProcess-miR) — v1.0.0.

What BioProcess-miR does

• Builds and queries a local DuckDB + Parquet knowledge base (HGNC, GO, Reactome, WikiPathways, optional KEGG under safeguard, miRTarBase).
• Resolves biological processes, expands membership, links experimentally supported miRNA–target evidence, runs explicit-universe enrichment, applies transparent prioritization policies, and generates provenance-rich reports.

What it does not do

• Disease-specific scoring, SH-SY5Y logic, expression integration, network propagation, ML ranking, therapeutic recommendation, or causal inference.
• Opaque composite “best miRNA” scores by default (components_only_v1).
• Redistribution of restricted KEGG pathway content.

Architecture

Canonical gene key: HGNC ID. Schema version 1.1.0 (independent of package 1.0.0).

Installation

pip install bioprocessdb
pip install "bioprocessdb[stats]"       # scipy / statsmodels
pip install "bioprocessdb[reporting]"   # matplotlib PNG export
pip install "bioprocessdb[all]"         # stats + reporting + dev tooling

Python 3.11+. Optional groups: stats, reporting, notebook, test, dev, docs, all.

Quick example

# Offline demonstration database (synthetic/minimal fixtures — NOT production)
python scripts/build_demo_database.py --overwrite
bioprocessdb --config demos/demo_v1/config/default.yaml analyze \
  --record GO:GO:0006979 \
  --include-descendants \
  --universe mirtarbase-eligible-targets \
  --policy components_only_v1 \
  --report html \
  --output results/examples/oxidative_stress

bioprocessdb --config demos/demo_v1/config/default.yaml run examples/oxidative_stress.yaml

from bioprocessdb import (
    BioProcessDB,
    ProcessEngine,
    ProcessMiRNALinker,
    MiRNATargetEnrichment,
    MiRNAPrioritization,
    BioProcessMiRReporter,
)

db = BioProcessDB("demos/demo_v1/database/bioprocess.duckdb")

Outputs

Under results/: process queries, process–miRNA links, enrichment, prioritization, and HTML/Markdown/JSON reports with manifests and fingerprints.

Provenance

Every analysis records package/schema/module versions, database build_id, source versions, query parameters, and result fingerprints. See docs/reproducibility.md.

Limitations

• Quality of results depends on local database completeness and upstream evidence.
• Demo fixtures are tiny and synthetic — not for biological conclusions.
• KEGG is optional and restricted; exports are gated.
• Enrichment significance ≠ biological importance or therapeutic utility.

Citation

See CITATION.cff and docs/data_licenses.md. Cite BioProcess-miR software separately from HGNC, GO, Reactome, WikiPathways, miRTarBase, and KEGG (when used).

Documentation

• Installation · Quickstart
• Public API · CLI reference
• Tutorials · Release notes v1.0.0
• CHANGELOG · SECURITY · CONTRIBUTING

License

MIT for project code. Upstream databases retain their own terms — notably KEGG redistribution restrictions. See docs/data_licenses.md.