Astrocytes Paper — Data, Code and Reproducibility

Data and code are publicly available at https://github.com/abotlp/Astrocytes_paper. All analysis outputs are organized under Astrocytes_project/GUILDifyTools-main/astrocytes and are built on the BIANA‑derived human interactome in Astrocytes_project/BIANA_phy.

Run GUILDify locally (seeds → ranked network + enrichment)

Run from the repository root with repo‑relative paths. Provide HGNC gene symbols separated by semicolons.

Example (generalized):

python GUILDifyTools-main/scripts/run_guildify_local.py \
  -i "SEED1; SEED2; SEED3" \
  -n BIANA_phy \
  -o GUILDifyTools-main/local_guildify/<run_name> \
  -s netcombo -re 3 -it 2 \
  -t 9606 -ns BIANA_phy -ti all \
  -dr BIANA_phy/drug_info.txt

Notes

• -i: semicolon‑separated HGNC symbols (e.g., for astrocytes overexpressed).
• -n: network directory (here, BIANA_phy).
• -o: output folder created for this run.
• -s: scoring method (netcombo, netscore, netzcore, netshort, or diamond).
• -re, -it: repetitions/iterations for netscore/netcombo/netzcore.
• -t: taxonomy ID (9606 for human).
• -ns: network source label (BIANA_phy).
• -ti: tissue filter (all).
• -dr: drug–target info file path (repo‑relative).

Dependencies

• Python 3, numpy, networkx, scipy, statsmodels.

Where to find key outputs

Per‑run outputs (phenotypes and astrocyte seeds)

• Folder: GUILDifyTools-main/astrocytes/<run> where <run> is one of:
  • alzheimer, cardiovascular, cognitive_impairment, diabetes, inflammation, oxidative_stress
  • astrocytes_overexpressed, astrocytes_underexpressed
• Key files:
  • seeds.txt — seed genes actually recognized in the network
  • guild_scores.txt — consensus ranking across methods
  • Enrichment: enrichment.GObp.*, enrichment.GOmf.*, enrichment.Reactome.* (Bonferroni and FDR)
  • Top‑k subnetworks: subnetwork.sif.{1,2,5} and corresponding network_plot.png{1,2,5} (when present)

Cross‑network overlaps (shared cores)

• Folders: GUILDifyTools-main/astrocytes/network_overlap_over5, .../network_overlap_under5, .../network_overlap_over2, .../network_overlap_under2
• Kept files (cleaned for publication):
  • protein_overlap_results.txt — shared gene list with seed status/source
  • function_overlap_results.txt — Reactome pathway overlap for the shared set
  • subnetwork_genes.sif, subnetwork_genes.json — core overlap network

Mutation analyses (FoldX ΔΔG + ClinVar)

• Folder: ddg_analysis/
  • ddg_docking.csv — docking‑derived complex ΔΔG summary
  • ddg_homology.csv — homology/template‑derived complex ΔΔG summary
  • clinvar.csv — ClinVar variant table used for mapping

BIANA network assets

• Folder: BIANA_phy/ — node/edge tables, annotations, and resources used by GUILDify.
• Some very large files are not tracked in Git; host externally (e.g., Zenodo/OSF) and add the link here when available.

Citation and contact

Please cite the associated manuscript and this repository if you use the data or code. For questions or issues, open a GitHub issue in this repo.