pd-netprox

Protein-network proximity analysis for Parkinson's disease candidate prioritization using STRING and GTEx substantia nigra expression.

Key Components of Repository

• A canonical, documented pipeline instead of multiple competing script paths.
• Offline seed mapping from local STRING alias files.
• A tracked analysis rerun path that works from the files stored in Git.
• A full rebuild path for the main network preprocessing if you also provide the raw STRING detailed network download locally.
• SHA256 checksums for the tracked reproducibility-critical inputs and outputs.

Quick start

conda env create -f environment.yml
conda activate pd-netprox
python scripts/verify_repository.py
python scripts/run_pipeline.py

If you also have the raw STRING detailed network file locally, the pipeline will rebuild data/string/ppi_edges.csv. Otherwise, reuses the tracked data/string/ppi_edges.csv :(

To refresh the checksum file after a fresh run:

python scripts/run_pipeline.py --update-checksums

Canonical pipeline order

1. scripts/extract_gtex_sn.py
2. scripts/map_seeds_to_ensp.py
3. scripts/process_string.py if the raw STRING detailed network is available
4. scripts/build_graph.py
5. scripts/compute_proximity.py
6. scripts/prepare_top20.py
7. scripts/scatter_plot.py
8. scripts/plot_ppi_subgraph.py

Tracked inputs and outputs

Tracked source or reduced-input files:

• data/expression/GTEx_Analysis_2017-06-05_v8_RNASeQCv1.1.9_gene_median_tpm.gct.gz
• data/expression/gtex_sn_expression.tsv
• data/seeds/pd_monogenic.txt
• data/seeds/pd_monogenic_ensp.txt
• data/seeds/pd_monogenic_seed_mapping.tsv
• data/string/9606.protein.aliases.v11.0.txt.gz
• data/string/9606.protein.info.v11.0.txt.gz
• data/string/ppi_edges.csv

Tracked canonical outputs:

• results/ranked_candidates.tsv
• results/top20_candidates.tsv
• results/top20_genes_for_enrichment.txt
• figures/proximity_vs_tpm_corr.png
• figures/ppi_subgraph_top20.png

Not tracked in normal Git history:

• The raw STRING detailed interaction file because it exceeds GitHub's normal file-size limits.
• Local-only large or unused files such as uncompressed alias tables, intermediate graph pickles, and archive material in data/NOT USED/.

Environment files

• environment.yml: Small runtime environment for the canonical pipeline.
• environment.exact.yml: Full export of the Windows conda environment used for this repository snapshot.
• requirements.exact.txt: Exact pip freeze --all output from the same environment.

Validation

Repository validation:

python scripts/verify_repository.py

Requires the raw STRING detailed network to be present too:

python scripts/verify_repository.py --full-rebuild

Repository layout

• scripts/: Canonical pipeline scripts only.
• scripts/legacy/: Archived exploratory or superseded scripts kept for provenance.
• data/: Tracked inputs, reduced inputs, and documented optional raw inputs.
• results/: Canonical tabular outputs tracked in Git.
• figures/: Canonical figures tracked in Git.
• paper/: Writing and manuscript material.

Reach out with questions