bio-crispr-screens-mageck-analysis

Version Compatibility

Reference examples tested with: MAGeCK 0.5+, matplotlib 3.8+, numpy 1.26+, pandas 2.2+

Before using code patterns, verify installed versions match. If versions differ:

• Python: pip show <package> then help(module.function) to check signatures
• CLI: <tool> --version then <tool> --help to confirm flags

If code throws ImportError, AttributeError, or TypeError, introspect the installed

package and adapt the example to match the actual API rather than retrying.

MAGeCK CRISPR Screen Analysis

"Analyze my pooled CRISPR screen with MAGeCK" → Count sgRNA reads, normalize across samples, and rank genes by enrichment or depletion using the MAGeCK robust rank aggregation algorithm.

• CLI: mageck count → mageck test for standard analysis
• CLI: mageck mle for multi-condition designs

Count sgRNAs from FASTQ

Goal: Quantify sgRNA representation from raw sequencing data.

Approach: Map FASTQ reads to the sgRNA library sequences with MAGeCK count, producing a normalized count matrix and QC summary across all samples.

# Count reads mapping to sgRNA library
mageck count \
    -l library.csv \
    -n experiment \
    --sample-label Day0,Treated1,Treated2,Control1,Control2 \
    --fastq Day0.fastq.gz Treated1.fastq.gz Treated2.fastq.gz Control1.fastq.gz Control2.fastq.gz \
    --norm-method median

# Output files:
# experiment.count.txt - normalized counts
# experiment.count_normalized.txt - normalized counts
# experiment.countsummary.txt - QC summary

Library File Format

# library.csv (tab-separated)
sgRNA_ID	Gene	Sequence
BRCA1_1	BRCA1	ATGGATTTATCTGCTCTTCG
BRCA1_2	BRCA1	CAGCAGATACTTGATGCATC
TP53_1	TP53	CCATTGTTCAATATCGTCCG
...

MAGeCK Test (RRA Algorithm)

Goal: Identify genes significantly enriched or depleted between treatment and control conditions.

Approach: Run MAGeCK test with robust rank aggregation, which ranks sgRNAs by fold change, tests whether per-gene sgRNA rankings deviate from uniform, and reports gene-level significance with FDR correction.

# Compare treatment vs control
mageck test \
    -k experiment.count.txt \
    -t Treated1,Treated2 \
    -c Control1,Control2 \
    -n results \
    --norm-method median \
    --gene-test-fdr-threshold 0.25

# Output files:
# results.gene_summary.txt - gene-level results
# results.sgrna_summary.txt - sgRNA-level results

MAGeCK MLE (Maximum Likelihood)

Goal: Estimate gene effects in complex experimental designs with multiple conditions or covariates.

Approach: Define a design matrix specifying sample-condition relationships, then run MAGeCK MLE which fits a generalized linear model to estimate per-gene beta scores (effect sizes) for each condition.

# Create design matrix
# design.txt:
# Samples    baseline    treatment
# Day0       1           0
# Control1   1           0
# Control2   1           0
# Treated1   1           1
# Treated2   1           1

mageck mle \
    -k experiment.count.txt \
    -d design.txt \
    -n mle_results \
    --norm-method median

# Output: mle_results.gene_summary.txt with beta scores

Interpret Results

Goal: Extract significant essential and resistance genes from MAGeCK output.

Approach: Load the gene summary table, filter by negative-selection FDR for dropout/essential genes and positive-selection FDR for enriched/resistance genes, and rank by MAGeCK score.

import pandas as pd

# Load gene summary
genes = pd.read_csv('results.gene_summary.txt', sep='\t')

# Negative selection (dropout/essential)
essential = genes[(genes['neg|fdr'] < 0.05)].sort_values('neg|rank')
print(f'Essential genes (dropout): {len(essential)}')
print(essential[['id', 'neg|score', 'neg|fdr']].head(20))

# Positive selection (enrichment/resistance)
resistant = genes[(genes['pos|fdr'] < 0.05)].sort_values('pos|rank')
print(f'Resistance genes (enriched): {len(resistant)}')
print(resistant[['id', 'pos|score', 'pos|fdr']].head(20))

Visualize Results

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

genes = pd.read_csv('results.gene_summary.txt', sep='\t')

# Volcano plot
fig, ax = plt.subplots(figsize=(10, 8))

x = genes['neg|lfc']
y = -np.log10(genes['neg|fdr'])

colors = ['red' if fdr < 0.05 else 'gray' for fdr in genes['neg|fdr']]
ax.scatter(x, y, c=colors, alpha=0.5, s=10)

# Label top hits
top_hits = genes[genes['neg|fdr'] < 0.01].nsmallest(10, 'neg|rank')
for _, row in top_hits.iterrows():
    ax.annotate(row['id'], (row['neg|lfc'], -np.log10(row['neg|fdr'])))

ax.axhline(-np.log10(0.05), linestyle='--', color='black', alpha=0.5)
ax.set_xlabel('Log2 Fold Change')
ax.set_ylabel('-log10(FDR)')
ax.set_title('MAGeCK Negative Selection')
plt.savefig('mageck_volcano.png', dpi=150)

MAGeCK Pathway Analysis

# Gene set enrichment on screen results
mageck pathway \
    -g results.gene_summary.txt \
    -c go_biological_process.gmt \
    -n pathway_results \
    --pathway-fdr-threshold 0.25

Time-Course Screens

# Compare multiple timepoints
mageck mle \
    -k timecourse.count.txt \
    -d timecourse_design.txt \
    -n timecourse_results

# Design matrix for time course:
# Samples    baseline    day7    day14
# Day0       1           0       0
# Day7_R1    1           1       0
# Day7_R2    1           1       0
# Day14_R1   1           0       1
# Day14_R2   1           0       1

CRISPR Activation (CRISPRa) Screens

# For CRISPRa, focus on positive selection
mageck test \
    -k crispra.count.txt \
    -t Activated1,Activated2 \
    -c Control1,Control2 \
    -n crispra_results

# Hits are genes where activation causes phenotype
# Use pos|fdr and pos|score columns

MAGeCK-VISPR (Visualization)

# Generate interactive report
mageck-vispr run \
    -n vispr_report \
    -c config.yaml

# config.yaml example:
# experiment: screen_name
# assembly: hg38
# species: homo_sapiens
# targets: library.csv
# sgrnas: experiment.count.txt
# samples:
#   - Day0
#   - Treated1

Related Skills

• screen-qc - Quality control before MAGeCK
• hit-calling - Alternative hit calling methods
• pathway-analysis/gsea - Downstream enrichment analysis