crispr-screen-analysis

CRISPR Screen Analysis

Overview

Pooled CRISPR genetic screens (genome-wide KO, CRISPRa, CRISPRi) identify genes essential for cellular fitness, drug response, or phenotype acquisition. This skill covers the complete workflow from raw sequencing reads to validated hits, using MAGeCK for statistical testing, DepMap integration for validation, and pathway enrichment for biological interpretation.

When to Use This Skill

• Analyzing MAGeCK count files from pooled CRISPR screens
• Computing gene-level essentiality scores (LFC, RRA, MLE)
• Calling hits in positive and negative selection screens
• Identifying synthetic lethal interactions
• QC of sgRNA representation and library dropout
• Pathway enrichment of CRISPR screen hits
• Integrating with DepMap data for cross-cell-line analysis
• Visualizing volcano plots, rank plots, and sgRNA distributions

Quick Start

MAGeCK Count (Read → Count Matrix)

# Count reads from FASTQ
mageck count \
    -l library.csv \
    -n screen_output \
    --sample-label "Day0,Day14_rep1,Day14_rep2" \
    --fastq Day0.fastq.gz Day14_rep1.fastq.gz Day14_rep2.fastq.gz \
    --sgrna-len 20 \
    --trim-5 ACCG \
    --pdf-report

MAGeCK Test (Hit Calling - RRA)

# Negative selection: essential genes drop out
mageck test \
    -k screen_output.count.txt \
    -t Day14_rep1,Day14_rep2 \
    -c Day0 \
    -n neg_selection \
    --gene-lfc-method median \
    --remove-zero both \
    --remove-zero-threshold 0 \
    --pdf-report

MAGeCK MLE (Maximum Likelihood Estimation)

# Design matrix for MLE
cat > design_matrix.txt << 'EOF'
Samples	baseline	treatment
Day0	1	0
Day14_rep1	0	1
Day14_rep2	0	1
EOF

mageck mle \
    -k screen_output.count.txt \
    -d design_matrix.txt \
    -n mle_output \
    --norm-method median

Python Analysis of MAGeCK Results

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

# Load MAGeCK gene summary
gene_summary = pd.read_csv("neg_selection.gene_summary.txt", sep="\t")
print(gene_summary.head())
print(f"Total genes tested: {len(gene_summary)}")

# Separate positive and negative selection results
neg_score = gene_summary[["id", "neg|lfc", "neg|fdr", "neg|rank"]].copy()
neg_score.columns = ["gene", "lfc", "fdr", "rank"]
neg_score["hits"] = neg_score["fdr"] < 0.1

print(f"Negative selection hits (FDR < 0.1): {neg_score['hits'].sum()}")
print(f"\nTop 10 essential genes:")
print(neg_score.nsmallest(10, "fdr")[["gene", "lfc", "fdr"]].to_string(index=False))