🧬 Off-target Analysis Pipeline for CRISPR/Prime Editing

Welcome! This pipeline analyzes off-target effects in CRISPR or Prime Editing experiments using paired-end NGS sequencing data with UMI (Unique Molecular Identifier) deduplication.

• 🔍 UMI-based deduplication for accurate quantification
• 🎯 Dual alignment strategy against both on-target and genome-wide references
• 🧹 Local alignment filtering to remove sequences similar to known contaminants
• ✂️ Soft-clipping analysis to identify and remove primer artifacts
• 🚀 Automated batch processing for multiple samples

• R1: [Illumina Adapter] - [S + 8bp UMI] - [Genome-specific sequence]
• R2: Complementary sequence with target region

For detailed oligo design and library preparation protocol, please refer to: 📖 GUIDE-seq Simplified Library Preparation Protocol

• Linux/Unix environment (tested on Ubuntu 22.04)
• Python 3.7+
• Bash shell
• At least 16GB RAM (32GB recommended for human genome alignment)
• ~50GB free disk space for reference genome and analysis

Install the following tools with specific versions:

# Create conda environment
conda create -n offtarget python=3.8
conda activate offtarget
# Install core dependencies with specific versions
conda install -c bioconda umi_tools=1.1.6
conda install -c bioconda minimap2=2.28
conda install -c bioconda samtools=1.18
conda install -c bioconda pysam=0.21.0
conda install -c conda-forge biopython=1.83
# Or use the provided environment file
conda env create -f environment.yml

Download the human reference genome (GRCh38):

# Download GRCh38 reference genome
wget ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCA/000/001/405/GCA_000001405.15_GRCh38/seqs_for_alignment_pipelines.ucsc_ids/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna.gz
# Decompress
gunzip GCA_000001405.15_GRCh38_no_alt_analysis_set.fna.gz
# Create minimap2 index (optional but recommended for speed 🚀)
minimap2 -d GRCh38.mmi GCA_000001405.15_GRCh38_no_alt_analysis_set.fna

The pipeline requires paired-end NGS sequencing data in FASTQ format:

• R1 file: *_R1_001.fastq[.gz] - Contains UMI and target sequence
• R2 file: *_R2_001.fastq[.gz] - Contains complementary sequence

File naming convention: {sample_id}_S{sample_id}_L001_R{1,2}_001.fastq[.gz]

Example:

85_S85_L001_R1_001.fastq.gz
85_S85_L001_R2_001.fastq.gz

For homology filtering, provide a tab/space-separated file:

name sequence alignment_score_threshold
donor_1000bp ATCGATCGATCG 130
control_seq GCTAGCTAGCTA 100

1. Clone the repository:

git clone https://github.com/yourusername/offtarget-analysis.git
cd offtarget-analysis

2. Make scripts executable:

chmod +x run_umi_pipeline_v5.sh

3. Configure paths (Important! 🎯):

# Copy and edit the configuration template
cp config.template config.sh
nano config.sh
# Update these paths to match your system:
# - SCRIPT_DIR: Path to Python scripts
# - REF_DIR: Path to reference genomes 
# - GRCH38_REF: Path to GRCh38 reference

./run_umi_pipeline_v5.sh <input_folder>

./run_umi_pipeline_v5.sh <input_folder> <reference_sequences.txt>

# Process all samples in the data directory
./run_umi_pipeline_v5.sh ./data/
# Process with homology filtering
./run_umi_pipeline_v5.sh ./data/ ./homology_sequences.txt

• Filters reads containing specified indicator sequences
• Default indicators:
  • R1: CTACAAGAGCGGTGAGt
  • R2: GGTGCCAGAGGTATTGGCGCTAGGGTCA

• Removes reads similar to known sequences using local alignment
• Checks both forward and reverse complement

• Extracts 9bp UMI from R1 reads
• Adds UMI to read headers

• Aligns to on-target reference
• Aligns to human genome (GRCh38)

• Retains properly paired reads
• Removes PCR duplicates based on UMI

• Removes primer artifacts
• Filters soft-clipped reads

• Counts editing events at each position
• Generates summary statistics

For each sample, the pipeline generates:

• dedup_on_target_{sample}_filtered_r1_filtered.bam - Final on-target alignments
• dedup_off_target_{sample}_filtered_r1_filtered.bam - Final off-target alignments
• *_align_sites.txt - Junction count tables

The *_align_sites.txt files contain:

chromosome position count
chr1 12345 10
chr2 67890 5

Edit run_umi_pipeline_v5.sh:

indicator_seq_r1="YOUR_R1_SEQUENCE"
indicator_seq_r2="YOUR_R2_SEQUENCE" 

• Alignment scoring: Edit filter_similar_sequences.py
• Soft-clipping threshold: Edit soft_clip_filter_out.py
• UMI pattern: Modify --bc-pattern in the pipeline script

🔴 "command not found" errors

• Check all dependencies are installed
• Activate conda environment: conda activate offtarget

🔴 Memory errors during alignment

• Reduce thread count (default: 8)
• Use pre-built minimap2 index

🔴 Low read counts after filtering

• Verify indicator sequences match your library
• Check reference genome paths
• Review filtering thresholds

MIT License

For questions or issues:

• 🐛 Open an issue on GitHub
• 📧 Contact: [bbakgosu@snu.ac.kr]

This pipeline leverages these excellent tools:

• UMI-tools
• minimap2
• SAMtools
• Biopython