pRoteomics is the scientific analysis repository for the Exp9 spatial proteomics study: hippocampal neuronal neuropil, neuronal soma, and microglia/PVM-enriched ROI datasets.
It owns scientific inference. Manuscript prose and journal figure rendering are intentionally maintained in a separate repository — see Publication boundary.
pipeline.yml is the only active machine-readable source of truth for script
order, dataset support, inputs, outputs, and safe-rerun status. Start with
WORKFLOW.md for the plain-language scientific workflow, then use
RUN_ORDER.md as the detailed command reference.
Rscript run_dataset_pipeline.R --list-stages
Rscript run_dataset_pipeline.R --dataset microglia --stage modules_downstream --dry-run
Rscript run_dataset_pipeline.R --dataset all --stage integration --dry-runRun named stages from pipeline.yml after private inputs are available:
Rscript run_dataset_pipeline.R --dataset <dataset> --stage core
Rscript run_dataset_pipeline.R --dataset <dataset> --stage qc
Rscript run_dataset_pipeline.R --dataset <dataset> --stage enrichment
Rscript run_dataset_pipeline.R --dataset <dataset> --stage modules_wgcna
Rscript run_dataset_pipeline.R --dataset <dataset> --stage modules_downstream
Rscript run_dataset_pipeline.R --dataset all --stage integrationValid dataset families are neuron_neuropil, neuron_soma, and microglia.
Microglia data are region-only microglia/PVM-enriched ROI/local
microenvironment proteomics, not purified microglia.
The authoritative control compartment-marker validation is
analysis/qc/render_compartment_abundance_figures.R.
It uses reconstructed observed, non-imputed abundance and animal-level
descriptive summaries; it does not estimate purity, cell fractions,
deconvolution, copy number, or total hippocampal abundance.
This repository owns, end to end:
- preprocessing and protein/gene identifier mapping;
- quality control, missingness, marker fidelity and confounding checks;
- spatial systems validation, bilateral aggregation and CA2-SLM robustness;
- differential abundance and enrichment;
- WGCNA module and supermodule construction and interpretation;
- cell-type enrichment;
- spatial and differential networks;
- biological integration and behaviour/physiology coupling;
- canonical result tables and publication source-data exports;
- the scientific regression suite.
pipeline.yml active pipeline registry
WORKFLOW.md plain-language workflow guide
RUN_ORDER.md detailed command reference
run_dataset_pipeline.R registry-driven launcher
R/ reusable scientific function libraries
paths.R repository-root and path bootstrap
data_contracts/ dataset, identifier and module contracts
qc/ QC and compartment-marker helpers
statistics/ module statistics, WGCNA, evidence
spatial/ spatial identity, atlas and robustness
enrichment/ enrichment IO, GO themes, EWCE
networks/ network construction helpers
utilities/ paths, registry, validation, export, plotting
analysis/ runnable analysis entrypoints
preprocessing/ preprocessing and identifier mapping
qc/ QC, marker and confounding checks
spatial_validation/ spatial systems, bilateral, CA2-SLM
differential_abundance/ differential abundance and GO/GSEA
wgcna/ WGCNA construction and downstream
enrichment/ cell-type enrichment (EWCE)
spatial_networks/ spatial and differential networks
integration/ integration and behaviour coupling
publication_source_data/ PRIDE and publication source data
config/ frozen scientific configuration contracts
data/ raw, metadata and reference inputs
work/ regenerable intermediates, gitignored
results/ canonical scientific results
<domain>/<analysis>/<scope>/ tables/ plots/ models/ manifests/ reports/
exports/ frozen outward-facing bundles
publication_source_data/ the only manuscript-facing interface
tests/ private-data-independent tests
audits/ provenance and robustness audits
tools/ maintenance, export and reference utilities
archive/ superseded code kept for provenance
docs/ reviewer and maintenance documentation
pride_submission/ generated, gitignored deposition payload
Analysis stage identities describe function rather than history. The former
04_differential_expression_enrichment/ is now
analysis/differential_abundance/; manuscript-facing text should describe
these outputs as differential abundance and enrichment results.
This repository's responsibility ends at canonical publication source data plus a scientific provenance manifest:
exports/publication_source_data/<publication_id>/
exports/publication_source_data/manifest.csv
This repository has no publication responsibility beyond that point. The
manuscript repository (Exp9_manuscript, local/private) owns panel
composition, typography, panel dimensions, legends, SVG assembly, PDF/PNG/TIFF
export, journal figure naming and layout, and the submission bundle. It renders from a frozen copy of the
bundle above and never reads a live path in this repository, so a future
restructure here cannot break manuscript rendering.
Behaviour data for Figure 1 and Extended Data 5/9 are owned upstream by
topohl/MMMSociability and enter the
manuscript repository through their own frozen import bundle. They are not
re-analysed here.
The test suite enforces this boundary: no manuscript renderer, no manuscript prose and no journal assembly code may reappear in this repository.
- Workflow
- Command reference
- Documentation map
- Analysis entrypoints
- Results ownership
- Repository architecture
- Restructure plan and equivalence oracle
- WGCNA workflow
- Datasets
- Input contracts
- Output contracts
- Output layout: work, results, exports
- Repository name recommendation
- Microglia ROI interpretation
- Reviewer reproducibility
- PRIDE export
Active scripts are listed only in pipeline.yml. Scripts excluded from the
canonical automated run are tracked in its legacy section with an explicit
replacement and status, and are documented in docs/NAMING_MIGRATION.md.
analysis/publication_source_data/ is the active export module. It builds the
canonical publication source-data bundle, its manifest and the hashes of the
scientific source tables, and audits figure outputs for publication readiness.
It does not assemble journal figures, name or lay them out for a journal, or
build a submission package: those belong to
Exp9_manuscript/tools/package_journal_figures.R. Superseded
generations live under archive/ and must not be treated as canonical;
diagnostic and validation-only scripts live under audits/.
CI and reviewer dry-runs validate the registry, script availability, dataset capabilities, and table contracts without private raw data. Full scientific runs require the private source matrices, metadata, raw/vendor files, and local deposition payload described in the manuscript data availability statement.
Tobias Pohl