Chemogenomic maps reveal a PRDX1-mediated iron-damage axis in the DNA damage response

This repository contains the bulk RNA-seq analysis from the above study: CRISPRi knockdowns of IREB2, MRGBP and PAX7 compared against a non-targeting control in A549 dCas9-KRAB cells, with differential expression of each knockdown versus the control.

Authors: O'Loughlin TA†,*, Arab A†, Misiukiewicz S, Montesano E, Yogodzinski C, Borah AA, Quarantotti V, Lou K, Rosen BS, Corn JE, Gianni D, Kabir S*, Forment JV*, Gilbert LA*

_{† equal contribution  ·  * corresponding author}

Samples

Sixteen paired-end libraries (Illumina NovaSeq X Plus), four replicates per condition:

Condition	sgRNA target	Replicates	Sample IDs
Control	non-targeting (NTC)	4	TONC1–TONC4
IREB2 knockdown	IREB2	4	TOIB1–TOIB4
MRGBP knockdown	MRGBP	4	TOMR1–TOMR4
PAX7 knockdown	PAX7	4	TOPX1–TOPX4

Data availability

Raw FASTQ files and the processed gene-level TPM matrix are deposited in NCBI GEO; the accession is provided in the associated publication. The completed submission manifest is data/DDRiPRDX1_RNAseq_filled.xlsx.

Read processing

FASTQs are quantified with nf-core/rnaseq (v3.18.0) against GRCh38 (GENCODE v47). The exact command is in command.sh, the pipeline input sheet in samplesheet.csv, and executor settings in config.json. Per-sample Salmon quant.sf files are the input to the differential expression analysis.

Differential expression

analysis/rnaseq_de.R (R 4.4.2) performs the following:

1. Imports Salmon quantifications with tximport using the sample metadata in data/samplesheet_Tomo.csv, aligning samples to count columns by name.
2. Filters low-expression genes (edgeR::filterByExpr) and applies TMM normalisation.
3. Fits a limma-voom linear model ~ replicate + gene_target, where replicate blocks for replicate effects and gene_target is releveled so NTC is the reference — each coefficient therefore tests a knockdown against the non-targeting control.
4. Computes moderated statistics with eBayes(robust = TRUE) and maps Ensembl gene IDs to symbols with org.Hs.eg.db.

Run it from the repository root (after the pipeline has produced results/star_salmon/):

Rscript analysis/rnaseq_de.R

Outputs:

• results_dge/DE_<target>_vs_NTC.csv — one table per knockdown, with gene ID, symbol, log2 fold change, and moderated statistics (logFC > 0 up-regulated, logFC < 0 down-regulated, relative to the non-targeting control).
• plots_dge/single_comparisons_volcano.png — one volcano panel per knockdown.

Repository layout

command.sh                          nf-core/rnaseq run command
config.json                         Nextflow executor configuration
samplesheet.csv                     pipeline input (16 samples)
data/samplesheet_Tomo.csv           sample metadata for the DE analysis
data/DDRiPRDX1_RNAseq_filled.xlsx   GEO submission manifest
analysis/rnaseq_de.R                differential expression analysis
results_dge/DE_<target>_vs_NTC.csv  per-knockdown DE tables
plots_dge/single_comparisons_volcano.png