CRISPR screens systematically test the function of every gene in the genome (or a curated subset) by introducing thousands of sgRNAs into a population of cells, applying a selective pressure, and measuring which sgRNAs drop out or accumulate. They’ve become the gold standard for unbiased functional genomics.
The three types of CRISPR screens
CRISPR knockout (KO) screens
Use Cas9 + sgRNAs to introduce loss-of-function mutations via NHEJ-mediated indels. Tests what happens when each gene is completely lost.
Best for: Identifying essential genes, drug resistance/sensitivity hits, signaling pathway components.
CRISPR interference (CRISPRi) screens
Use dCas9-KRAB to repress transcription. Targets the promoter region (typically -50 to +300 bp around the TSS).
Best for: Studying essential genes (where complete KO would kill the cell), tunable knockdown, repeated time points, non-coding RNAs.
CRISPR activation (CRISPRa) screens
Use dCas9 fused to transcriptional activators (VP64, p65, SAM complex) to upregulate gene expression. Targets the promoter region.
Best for: Identifying genes whose overexpression confers a phenotype — drug resistance, immune evasion, differentiation drivers.
Pooled vs arrayed screens
Pooled screens deliver all sgRNAs simultaneously to a cell population (typically via lentivirus at low MOI so each cell gets one sgRNA). The selection acts on the pool, and sgRNAs are quantified by NGS at the start and end. Most CRISPR screens are pooled.
Arrayed screens deliver one sgRNA per well across a plate. Each well is read out separately by imaging or other assays. More expensive but allows complex phenotypes (cell shape, localization, transcriptional response).
Common screen types by selection
| Screen type | Selection | Hits |
|---|---|---|
| Dropout (essentiality) | Cells passaged over time | sgRNAs that drop out target essential genes |
| Drug resistance | Cells treated with drug | sgRNAs that enrich confer resistance |
| Drug sensitivity | Cells treated with sublethal drug | sgRNAs that drop out confer sensitivity |
| Marker-based sorting | FACS-sort by reporter | sgRNAs that enrich in sorted populations regulate the reporter |
| Single-cell readout | Perturb-seq, CROP-seq | Each sgRNA’s transcriptional effect |
Library design considerations
- Library size: Genome-wide (~20,000 genes × 4–6 sgRNAs = 80,000–120,000 elements) or focused (kinome, transcription factors, druggable genome)
- sgRNAs per gene: 4–6 typical; more sgRNAs increase confidence in hits
- Coverage: 300–500× cell coverage per sgRNA at the start; lower coverage = noisier results
- Common libraries: Brunello, Brie, Bassik, GeCKO, Toronto Knockout (TKO), Doroshow CRISPRi
Standard workflow
- Design library in silico or order an existing one
- Clone library into lentiviral vector
- Generate lentivirus, titer carefully
- Transduce target cells at low MOI (~0.3) to ensure one sgRNA per cell
- Select with puromycin or other marker
- Apply selection (drug, time, FACS, etc.)
- Extract genomic DNA and PCR-amplify the sgRNA region
- Sequence on Illumina platform
- Analyze with MAGeCK, BAGEL2, drugZ, or similar tools
Analysis essentials
- Count sgRNA reads at start (T0) and end (T_final)
- Calculate log fold-change for each sgRNA
- Aggregate sgRNA-level changes to gene-level scores (multiple sgRNAs per gene strengthen confidence)
- Estimate FDR via permutation or modeling
- Validate top hits with individual sgRNAs and orthogonal assays
Common pitfalls
- Insufficient coverage — too few cells per sgRNA leads to drift
- Bottlenecks at transduction, selection, or passaging — lose representation of rare sgRNAs
- Off-target effects — confirm hits with at least 2–3 independent sgRNAs
- Cell-line copy number effects in KO screens — high copy regions cut multiple times, causing artifactual essentiality. Use copy-number-aware analysis (CERES, CHRONOS)
Single-cell CRISPR screens
Perturb-seq and CROP-seq combine pooled CRISPR perturbation with scRNA-seq readout — every cell gets a perturbation and a transcriptome. Reveals not just whether a perturbation matters but exactly how it changes the cell.
CRISPR screens are among the most powerful tools in functional genomics. Choose the screen type to match your question, design the library and selection carefully, and validate hits orthogonally before drawing conclusions.
