Add output manifest for analysis artefacts

Author: Ekin-Kahraman

README.md

@@ -17,6 +17,7 @@ Reproducible bulk RNA-seq differential expression pipeline using DESeq2: QC, shr
 - **Extended: Sex-stratified interaction analysis** — Condition-by-sex interaction model (`~ condition * gender`) to identify genes with sex-differential transcriptional responses, complementing the original study's sex-adjusted analysis with a formal interaction test
 - Extracts full GEO covariates (viral load Ct, age, sex, sequencing batch) for covariate-aware analyses
 - Raw and shrunken DE outputs, analysis summary metrics, and git/session provenance are generated automatically
+- `results/tables/output_manifest.csv` records file sizes and checksums for committed figures and tables
 
 ## Workflow
 
@@ -42,7 +43,10 @@ GSE152075 (n=484, GEO)
     └──→ 09 Sex interaction ── ~ condition * gender → 12 sex-biased genes (9 male, 3 female)
          │
          ▼
-      07 Provenance ──── Git commit, session info, config, package versions → REPRODUCIBILITY.md
+      12 Manifest ────── File-size and checksum manifest for committed figures/tables
+         │
+         ▼
+      07 Provenance ──── Git commit, session info, config, package versions → results/session_info.txt
 ```
 
 ## Methods Overview
@@ -251,6 +255,7 @@ bulk-rnaseq-differential-expression/
 │   ├── 09_sensitivity_analysis.R
 │   ├── 10_viral_load_stratification.R  # Extended: high vs low viral load DE
 │   ├── 11_sex_stratified_analysis.R    # Extended: condition x gender interaction
+│   ├── 12_output_manifest.R     # Checksums for committed figures/tables
 │   └── config.R                 # Shared analysis thresholds and helpers
 ├── data/
 │   └── [RDS files]
@@ -277,6 +282,7 @@ source("scripts/08_pathway_diagram.R")
 source("scripts/09_sensitivity_analysis.R")
 source("scripts/10_viral_load_stratification.R")
 source("scripts/11_sex_stratified_analysis.R")
+source("scripts/12_output_manifest.R")
 ```
 
 ## Methods

REPRODUCIBILITY.md

@@ -7,6 +7,7 @@ This repository is set up so a reviewer can reproduce the analysis with a small
 - **Version pinning**: `renv.lock` pins CRAN + Bioconductor package versions.
 - **Pre-computed outputs**: key figures and tables are committed under `results/` for convenience and quick verification.
 - **Analysis summary**: `results/tables/analysis_summary.csv` captures the main counts used in the narrative.
+- **Output manifest**: `results/tables/output_manifest.csv` records file sizes and MD5 checksums for committed figures and tables.
 - **Pinned pathway snapshot**: `data/reference/kegg_hsa_pathway_*.tsv` freezes the KEGG human pathway universe used by enrichment.
 
 ## From a clean checkout (recommended)
@@ -47,6 +48,7 @@ After a successful run, you should see (among others):
 - `results/tables/deseq2_results_shrunken.csv`
 - `results/tables/full_cohort_deseq2_results.csv`
 - `results/tables/analysis_summary.csv`
+- `results/tables/output_manifest.csv`
 - `results/figures/volcano_plot.png`
 - `results/figures/sensitivity_lfc_scatter.png`
 - `results/figures/pca_plot.png`

results/tables/output_manifest.csv

@@ -0,0 +1,29 @@
+"path","bytes","md5"
+"results/figures/dispersion_plot.png",505694,"24ad93bc7ee8d56bf4ced6b98d4e1a57"
+"results/figures/go_dotplot.png",500804,"67fff951fbd53390334375ee84df802f"
+"results/figures/kegg_dotplot.png",443704,"fa05cd5ad9f3d1046aeebd83c6f73326"
+"results/figures/ma_plot.png",678515,"7110b781f954801398920463b3974f78"
+"results/figures/pathway_diagram.png",342418,"9e6345e0a436b411d72ebda295d7fd76"
+"results/figures/pca_plot.png",257242,"bb9aeb68d36a090f31825c308827910b"
+"results/figures/pca_scree.png",172959,"580e94c17877221ce49e121964ae7513"
+"results/figures/pca_sex_stratified.png",205728,"3c10f6413a02c9e684551fdbfa241d12"
+"results/figures/qc_library_size.png",302762,"e63c4f1a6a77c9f9e15a752737edf155"
+"results/figures/sample_distances.png",342140,"5658db50d4aa5bb498d5032af8258d13"
+"results/figures/sensitivity_lfc_scatter.png",654130,"cc8ab93c82653b6ffb68abca6f2b5f83"
+"results/figures/sex_interaction_volcano.png",619575,"ccf483f2c66562f1a5b1acc3df76f527"
+"results/figures/top50_heatmap.png",475753,"a6c0563044c6a7a261b1eb6893729883"
+"results/figures/viral_load_ct_distribution.png",134209,"2767a9d8a588e087e05c0710bf691836"
+"results/figures/viral_load_isg_correlation.png",524869,"ccaa04a2f4c716c6196587c3b81402a1"
+"results/figures/viral_load_volcano.png",543042,"18c4dc222a8fe5c518c58d755c3050ef"
+"results/figures/volcano_plot.png",579141,"b306cae707412da5cbeeee5285593256"
+"results/tables/analysis_summary.csv",451,"c36ab48fef1b91b3ffc2660501bed10a"
+"results/tables/deseq2_results_shrunken.csv",1634699,"f7997b9e23b8d8f333a62dcd40aa81d4"
+"results/tables/deseq2_results.csv",1629915,"55b5443669b9273ea803bba9c95d61b1"
+"results/tables/full_cohort_deseq2_results.csv",2762171,"86587ccc81941a606ee0e908eb2d2501"
+"results/tables/go_biological_process.csv",142506,"ec7ea045881bb700a6376aa84998033f"
+"results/tables/kegg_pathways.csv",9894,"65c571c2c112082722296692f9a5afcb"
+"results/tables/sex_analysis_summary.csv",167,"d1201d896d5e2d0780114ccfd20d1c98"
+"results/tables/sex_interaction_de_results.csv",2085625,"30c1138fa3ae35887084b605c2582cb4"
+"results/tables/top_genes.csv",670,"b5c3dfd5f7dcf078ac512af37fa36b68"
+"results/tables/viral_load_de_results.csv",2276277,"46d6160b57565f2a7244d0228517b17a"
+"results/tables/viral_load_summary.csv",162,"5affeb45ca4a2421202f05b31d2e21f9"

run_all.R

@@ -27,17 +27,18 @@ run_script <- function(script_path, step) {
   })
 }
 
-run_script("scripts/00_get_data.R", " 1/11")
-run_script("scripts/01_qc.R", " 2/11")
-run_script("scripts/02_pca.R", " 3/11")
-run_script("scripts/03_deseq2.R", " 4/11")
-run_script("scripts/04_visualisation_volcano.R", " 5/11")
-run_script("scripts/05_model_diagnostics.R", " 6/11")
-run_script("scripts/06_enrichment.R", " 7/11")
-run_script("scripts/09_sensitivity_analysis.R", " 8/11")
-run_script("scripts/08_pathway_diagram.R", " 9/11")
-run_script("scripts/10_viral_load_stratification.R", "10/11")
-run_script("scripts/11_sex_stratified_analysis.R", "11/11")
+run_script("scripts/00_get_data.R", " 1/12")
+run_script("scripts/01_qc.R", " 2/12")
+run_script("scripts/02_pca.R", " 3/12")
+run_script("scripts/03_deseq2.R", " 4/12")
+run_script("scripts/04_visualisation_volcano.R", " 5/12")
+run_script("scripts/05_model_diagnostics.R", " 6/12")
+run_script("scripts/06_enrichment.R", " 7/12")
+run_script("scripts/09_sensitivity_analysis.R", " 8/12")
+run_script("scripts/08_pathway_diagram.R", " 9/12")
+run_script("scripts/10_viral_load_stratification.R", "10/12")
+run_script("scripts/11_sex_stratified_analysis.R", "11/12")
+run_script("scripts/12_output_manifest.R", "12/12")
 
 run_script("scripts/07_reproducibility.R", "Session")
 

scripts/12_output_manifest.R

@@ -0,0 +1,33 @@
+#!/usr/bin/env Rscript
+# Write a checksum manifest for committed result artefacts.
+
+dir.create("results/tables", recursive = TRUE, showWarnings = FALSE)
+
+figure_files <- list.files(
+  "results/figures",
+  pattern = "\\.(png|pdf)$",
+  full.names = TRUE
+)
+table_files <- list.files(
+  "results/tables",
+  pattern = "\\.csv$",
+  full.names = TRUE
+)
+table_files <- table_files[basename(table_files) != "output_manifest.csv"]
+
+artefacts <- sort(c(figure_files, table_files))
+if (length(artefacts) == 0) {
+  stop("No result artefacts found under results/figures or results/tables")
+}
+
+info <- file.info(artefacts)
+manifest <- data.frame(
+  path = gsub("\\\\", "/", artefacts),
+  bytes = as.numeric(info$size),
+  md5 = unname(tools::md5sum(artefacts)),
+  stringsAsFactors = FALSE
+)
+
+manifest <- manifest[order(manifest$path), ]
+write.csv(manifest, "results/tables/output_manifest.csv", row.names = FALSE)
+message("Wrote output manifest for ", nrow(manifest), " artefacts")

tests/testthat/test-smoke.R

@@ -96,6 +96,24 @@ test_that("key derived tables exist", {
   expect_gt(summary_df$shrunken_lfc_spearman[[1]], 0.8)
 })
 
+test_that("output manifest tracks committed result artefacts", {
+  manifest_path <- file.path(root, "results/tables/output_manifest.csv")
+  expect_true(file.exists(manifest_path))
+
+  manifest <- read.csv(manifest_path, stringsAsFactors = FALSE)
+  expect_true(all(c("path", "bytes", "md5") %in% names(manifest)))
+  expect_gt(nrow(manifest), 20)
+  expect_true(all(file.exists(file.path(root, manifest$path))))
+  expect_true(all(manifest$bytes > 0))
+  expect_true(all(grepl("^[0-9a-f]{32}$", manifest$md5)))
+
+  actual_sizes <- file.info(file.path(root, manifest$path))$size
+  expect_equal(as.numeric(manifest$bytes), as.numeric(actual_sizes))
+  expect_true("results/tables/analysis_summary.csv" %in% manifest$path)
+  expect_true("results/figures/volcano_plot.png" %in% manifest$path)
+  expect_true("results/figures/sensitivity_lfc_scatter.png" %in% manifest$path)
+})
+
 test_that("KEGG enrichment uses pinned human pathway references", {
   links_path <- file.path(root, "data/reference/kegg_hsa_pathway_links.tsv")
   names_path <- file.path(root, "data/reference/kegg_hsa_pathway_names.tsv")