EasyCM

Snakemake pipeline for cell-type marker gene discovery from scRNA-seq pseudobulk data. Per cell type: paired DESeq2 (~ subtype + sample) vs all other cells, apeglm LFC shrinkage, then fGSEA against Reactome+KEGG. Consumes pre-made pseudobulk matrices from EasyPseudobulk.

All R scripts must run inside the EasyDE micromamba environment (provides DESeq2, apeglm, fGSEA).

---

Pipeline Overview

Six rules, single {celltype} wildcard. Rules 02–04 run in parallel across all cell types.

01  Validate config + inputs           (once)
02  Prepare pseudobulk matrices        (per celltype)
03  DESeq2 + apeglm LFC shrinkage      (per celltype)
04  fGSEA pathway enrichment           (per celltype, optional)
05  Aggregate results + status plot    (once, after all celltypes)
06  Pipeline summary + DEG bar chart   (once, final)

For each cell type, rule 02 builds a combined count matrix in which every sample appears twice — once as interest (cells of that type) and once as other (all remaining cells). DESeq2 then tests ~ subtype + sample to isolate the cell-type effect while blocking on donor. Positive log2FoldChange = upregulated in the cell type of interest.

---

Quick Start

# 1. Install
micromamba env create -f installation/EasyCM_install.yml

# 2. Place inputs
#    data/counts/cell_mtx/{CellType}_persample_RNA_counts.tsv
#    data/counts/allbut_mtx/{CellType}_persample_RNA_counts.tsv
#    data/sample_metadata.csv

# 3. Edit config
cp config/config.yaml config/my_config.yaml

# 4. Dry run, then run
micromamba run -n EasyDE snakemake -n --config pipeline_config=config/my_config.yaml
micromamba run -n EasyDE snakemake --profile profiles/local \
    --config pipeline_config=config/my_config.yaml

# 5. Inspect
column -t -s, results/pipeline_summary.csv

---

Project Layout

EasyCM/
├── Snakefile
├── config/config.yaml                <- edit this
├── profiles/{local,slurm}/config.yaml
├── data/
│   ├── counts/{cell_mtx,allbut_mtx}/ <- pseudobulk TSVs from EasyPseudobulk
│   └── sample_metadata.csv
├── resources/gsea_files/             <- GMT + gene exclusion lists
├── workflow/scripts/
│   ├── 01_validate.R
│   ├── 02_prepare_pseudobulk.R
│   ├── 03_run_deseq.R
│   ├── 04_run_fgsea.R
│   ├── 05_aggregate_results.R
│   ├── 06_pipeline_summary.R
│   └── utils/{io,logging,filter,validation}_utils.R
├── installation/EasyCM_install.yml
└── docs/

Edit	Where
Analysis parameters	config/config.yaml
Execution settings	profiles/*/config.yaml
Inputs	data/counts/, data/sample_metadata.csv

---

Documentation

Guide	Contents
Installation	Environment setup
Configuration	Config reference, input formats
Running	Snakemake invocation, manual step-by-step
Methods	Paired design, DESeq2, apeglm, fGSEA
Output	File tree, column definitions, status taxonomy
Troubleshooting	Common errors, logs

---

Pipeline Status Values

Rule 05 writes celltype_summary.csv with a status per cell type:

Status	Meaning
success	DESeq2 completed, marker genes found
success_no_significant	DESeq2 ran, no genes passed FDR
skipped_no_samples	Cell type absent or zero samples after filtering
skipped_min_cells	Too few samples after donor deduplication
skipped_preflight	Too few genes passed filtering
skipped	Other skip — see error_message
error	DESeq2 or upstream crashed
not_run	No log or output found

Rule 05 emits status_overview.pdf (status + DEG counts heatmap). Rule 06 emits pipeline_summary.pdf (DEG bar chart, ordered by marker count).