IGVF CDMPRA Barcode Mapping Pipeline

Snakemake workflow for processing MPRA barcode libraries, aligning sequencing reads, generating barcode–variant assignments, and building final reporter element count matrices.

---

Dependencies

This pipeline requires:

HPC modules

• cutadapt
• bwa
• subread (featureCounts)

System tools

• gzip
• sed
• python3
• R >= 4.0

Python packages

Used in Module/A1.barcode_mapping.py

R packages

Required in A2.Barcode_Digest.R and A5.Cleansing_Count_Matrix.R:

• data.table
• reshape2
• Biostrings
• optparse
• ggplot2 (optional)

---

Directory Structure

IGVF_CDMPRA_BCmapping/ ├── Snakefile ├── config.yaml ├── Module/ │ ├── A1.barcode_mapping.py │ ├── A2.Barcode_Digest.R │ └── A5.Cleansing_Count_Matrix.R ├── logs/ └── (generated output files)

---

Configuration (config.yaml)

output_dir: "/path/to/output" bcmapping_fastq: "/path/to/bcmap_input.fastq.gz" libfile: "/path/to/library_file.txt"

fastq_dir: "/path/to/trim_align_fastqs" samples:

• sample1
• sample2
• ... adapterseq: "ACTAGTACACTCCCC"

Required fields

Field	Description
output_dir	Directory to place all generated outputs
bcmapping_fastq	FASTQ file used for barcode mapping
libfile	Library design table used for variant-barcode mapping
fastq_dir	Directory containing FASTQ files for BWA alignment
samples	List of sample prefixes (without .fastq.gz)
adapterseq	3’ adapter sequence for trimming with cutadapt

---

Running the Pipeline

Dry-run (check workflow): snakemake -n -p

Run full workflow: snakemake -j 20 --rerun-incomplete

---

Workflow Overview

A1 — Barcode Mapping

Reads the barcode FASTQ, extracts every 4th-line sequence, and maps barcodes to variants.

Input:

• bcmapping_fastq
• libfile

Output:

• bcmap.txt

---

A2 — Barcode Digest

Processes bcmap.txt to generate:

• Reporter_assignment.tsv
• barcode.fasta
• barcode.saf
• BC_statistics.pdf

Output files:

• Reporter_assignment.tsv
• barcode.fasta
• barcode.saf
• BC_statistics.pdf

---

A3 — Build Count Matrix (BWA alignment)

For each sample:

• trim reads using cutadapt
• align with bwa aln / samse
• generate sample.sam

Output:

• {sample}.sam

---

A4 — featureCounts

Uses SAM files + barcode SAF to build the raw count matrix.

Output:

• MPRA_barcode_count.txt

---

A5 — Cleansing Count Matrix

Performs:

• removing rows with all-zero barcodes
• removing outlier barcodes
• selecting variants supported by ≥5 barcodes

Output:

• Counts_noOutliers.tsv
• Reporter_Element.tsv
• barcode_representation.pdf

---

Final Outputs

File	Description
Counts_noOutliers.tsv	Barcode counts with outlier removal
Reporter_Element.tsv	Final element × sample count matrix
barcode_representation.pdf	QC histogram of barcode-per-variant distribution
barcode.fasta	FASTA file of barcodes
barcode.saf	SAF annotation for featureCounts

---

Notes

• Pipeline is modular: each step can be rerun independently.
• All log files stored in logs/.
• FASTQ filenames must match samples: list in config.yaml.

---

Maintainer

Hyunggyu Min
UNC Chapel Hill — Hyejung Won Lab