pRoteomics

pRoteomics is a repository of R-based workflows for spatial and systems-level proteomics analyses.

The repository has evolved from a primarily clusterProfiler/GSEA workflow into a broader analysis framework integrating:

• spatial proteomics
• differential enrichment analysis
• EWCE cell-type enrichment
• WGCNA and module analyses
• spatial region/layer network analysis
• bootstrap-based network stability analyses
• behavior and physiology coupling
• PRIDE/ProteomeXchange package preparation support

The current focus is high-resolution spatial proteomics across hippocampal region/layer structures combined with systems-level network and behavioral analyses.

---

Repository structure

01_preprocessing/
02_id_mapping/
03_qc_exploration/
04_differential_expression_enrichment/
05_celltype_enrichment_EWCE/
06_modules_WGCNA/
07_spatial_networks/
08_behavior_physio_coupling/
09_pride_submission/
09_export_pride_journal/
90_testing/
99_deprecated/

Canonical data and result folders:

data/raw/
data/metadata/
data/external/
data/processed/<module>/
results/figures/<module>/
results/tables/<module>/
results/source_data/<module>/
results/logs/<module>/
results/reports/<module>/
pride_submission/

Use R/paths.R for repo-relative paths. Local overrides should go in config/*.local.yml or environment variables such as PROTEOMICS_PROJECT_ROOT; do not commit machine-specific paths.

---

Pipeline overview

raw proteomics matrices + metadata
→ preprocessing / metadata harmonization
→ UniProt and ID mapping
→ QC and exploratory structure analysis
→ differential enrichment and GSEA
→ EWCE cell-type enrichment
→ WGCNA and module analyses
→ spatial network analyses
→ bootstrap stability analyses
→ behavior and physiology coupling
→ PRIDE / ProteomeXchange package metadata generation

---

Main analysis modules

01_preprocessing

Purpose:

• metadata formatting
• matrix harmonization
• imputation
• merged metadata generation

Representative scripts:

• 01_impute.r
• 03_gct_extractR.r
• 04_format_metadata.r

---

02_id_mapping

Purpose:

• UniProt mapping
• clusterProfiler-compatible ID conversion
• WGCNA-compatible identifier harmonization

Representative scripts:

• 01_MapThatProt.r
• 02_MapThatProt_batch.r

The canonical contrast handoff is now 01_preprocessing/03_gct_extractR.r to 02_id_mapping/01_MapThatProt_batch.r, producing clusterProfiler-ready mapped files at data/processed/02_id_mapping/mapped/<dataset>/forward/per_file/. Run this per dataset family, such as neuron_neuropil, neuron_soma, or microglia.

---

03_qc_exploration

Purpose:

• PCA
• variance partitioning
• rank abundance analysis
• protein/peptide QC

Representative scripts:

• 03_pcaPlot.r
• 05_pcaPlot_v3.r
• 07_varPart.r

---

04_differential_expression_enrichment

Purpose:

• GO enrichment
• GSEA
• pathway comparison
• publication-style enrichment figures

Representative scripts:

• 01_clusterProfiler.r
• 02_compareGO.r
• 03_compare_pathways.r

01_clusterProfiler.r now writes a manifest at:

data/processed/04_differential_expression_enrichment/clusterProfiler/<dataset>/clusterProfiler_manifest.csv

02_compareGO.r consumes that manifest instead of recursively discovering arbitrary CSVs. It filters by the required dataset column so neuron neuropil, neuron soma, and microglia runs do not overwrite or mix with each other. This preserves ontology, comparison, route category, route unit, simplification state, plot-used status, input hashes and config hashes across the clusterProfiler to compareGO handoff.

Dry-run checks are available:

Rscript 04_differential_expression_enrichment/01_clusterProfiler.r --dry-run
Rscript 04_differential_expression_enrichment/02_compareGO.r --dry-run

---

05_celltype_enrichment_EWCE

Purpose:

• EWCE analyses
• spatial cell-type interpretation
• measured-proteome-aware enrichment workflows

Representative scripts:

• 01_EWCE_E9.r

01_EWCE_E9.r now uses canonical module folders and supports --dry-run.

---

06_modules_WGCNA

Purpose:

• WGCNA
• module scoring
• module preservation
• overlap-based module generation

Representative scripts:

• 01_WGCNA.r
• 02_WGCNAtraitpreservation.r
• 03_module_spatial_networks.r

Phase 3 canonicalized the safer helper/downstream scripts 02_module_spatial_networks.r, 03_overlap_modules.r, and 91_module_score_v0.0.2.r. The central 01_WGCNA v.2.0.0.r remains documented for a data-aware path refactor.

---

07_spatial_networks

Purpose:

• anatomical region/layer relationship networks
• differential network analysis
• bootstrap network validation
• chord/network visualization

Representative scripts:

• 01_network_spatial_relations.r
• 02_differential_networks.r
• 03_bootstrap_network_stability.r

Phase 4 canonicalized 01_network_spatial_relations.r, and Phase 3 canonicalized 02_differential_networks.r, 03_bootstrap_network_stability.r, 04_bootstrap_differential_network_stability.r, 05_bootstrap_differential_network_figures.r, and 06_chord_diagram.r. The producer writes the spatial object at data/processed/07_spatial_networks/network_spatial_relations/network_spatial_relations_objects.rds, which downstream network and behavior scripts consume.

---

08_behavior_physio_coupling

Purpose:

• connect proteomics and network structure with behavior and physiology
• movement/stress score integration
• systems-level phenotype coupling

Representative scripts:

• 01_correlate_proteomics_with_behavior.r
• 02_network_behavior_coupling.r

02_network_behavior_coupling.r now reads behavior files from data/external/behavior/, writes canonical outputs under results/*/08_behavior_physio_coupling/network_behavior_coupling/, and supports --dry-run.

---

09_pride_submission

Purpose:

• create a local PRIDE/ProteomeXchange package skeleton
• generate SDRF-style sample-to-file metadata
• generate a file manifest
• calculate MD5 checksums
• validate whether raw files, search outputs, processed results, and metadata are traceable

Main command:

source("09_pride_submission/00_make_pride_package.R")

This creates or updates local files under the gitignored PRIDE_package/ folder:

PRIDE_package/00_metadata/sdrf_proteomics.tsv
PRIDE_package/00_metadata/pride_file_manifest.tsv
PRIDE_package/00_metadata/checksum.md5
PRIDE_package/00_metadata/validation_report.tsv

Large raw/vendor mass spectrometry files should be uploaded to PRIDE, not committed to GitHub.

Detailed instructions are in:

09_pride_submission/README_PRIDE.md

09_export_pride_journal

Purpose:

• generate PRIDE-ready manifests
• stage sample metadata and SDRF-like metadata
• stage publication supplementary tables
• validate deposition readiness
• write methods/provenance summaries

Main scripts:

09_export_pride_journal/01_make_pride_manifest.R
09_export_pride_journal/02_make_sample_metadata.R
09_export_pride_journal/03_make_supplementary_tables.R
09_export_pride_journal/04_validate_pride_submission.R
09_export_pride_journal/05_make_methods_summary.R

Each export script supports --dry-run to validate expected folders and inputs without staging or writing analysis products.

---

Running the pipeline

A recommended execution order is documented in:

RUN_ORDER.md

The repository also contains:

• 90_testing/ for exploratory or developmental workflows
• 99_deprecated/ for archived legacy scripts retained for reproducibility

---

System requirements

Recommended:

• R >= 4.2
• RStudio or VS Code
• macOS/Linux preferred for large workflows

Core package ecosystem includes:

• tidyverse
• clusterProfiler
• enrichplot
• WGCNA
• limma
• EWCE
• ggplot2
• openxlsx
• igraph
• ggraph
• patchwork
• pheatmap
• ComplexHeatmap
• lme4/lmerTest
• mgcv

Additional dependencies vary between workflows.

---

Outputs

Typical outputs include:

• publication-ready SVG/PDF figures
• enrichment tables
• network edge tables
• bootstrap stability summaries
• module score matrices
• EWCE outputs
• source-data tables
• QC reports
• PRIDE package metadata, manifests, checksums, and validation reports

---

Reproducibility philosophy

The repository intentionally preserves:

• exploratory workflows
• older analysis versions
• testing scripts
• alternative implementations

This is done to maintain reproducibility of intermediate biological findings and figure-generation pipelines.

Older scripts are preferentially moved to 99_deprecated/ rather than deleted.

---

Author

Tobias Pohl