How Much Library Coverage Is Enough for a Genome-Wide Screen?

200–500 cells per guide at the point of selection is the working range for most genome-wide screens. For a 100,000-guide library that means 20–50 million cells at selection, and more at transduction.

Below ~100 cells per guide, guide abundance estimates become noisy enough that real effects are lost. Above ~1,000, returns diminish quickly and cost rises linearly.

The second column is the number people plan; the fourth is the number they usually need. Transduction efficiency, selection death, and splitting losses each reduce coverage, and the losses compound.

• • Transduction. At MOI 0.3, a substantial fraction of cells is untransduced.
• • Selection. Antibiotic selection kills untransduced cells, which is intended — but it also removes coverage if the starting number was marginal.
• • Splitting. Each passage that splits a culture below the coverage threshold resets the effective library, and guides lost early cannot be recovered.
• • Uneven growth. Fast-growing clones dominate over time, which is the biology you are measuring but also a source of noise for non-essential genes.
• • Harvest losses. Cell counting errors at harvest translate directly into coverage errors.

A practical rule: calculate coverage at the lowest point in the protocol, not the highest.

• • Sequence the library and the transduced pool at day 0; the guide distribution should approximate the library's, with no guide missing.
• • Report the fraction of guides with zero counts at day 0 — a leading indicator of insufficient coverage.
• • Track cell numbers at every step and report them, since coverage is a claim about cells, not about reads.

For focused libraries targeting a pathway, higher effect sizes mean lower coverage requirements, and 200 cells per guide can suffice. For essentiality screens, where effects are large, coverage can be lower; for subtle regulatory phenotypes, it must be higher.

State the assumption explicitly: coverage requirements scale inversely with expected effect size.

EDITGENE builds screening-ready lines with documented transduction behaviour and expansion capacity, and advises on coverage arithmetic for focused versus genome-wide libraries.

Tell us the library size and phenotype. Contact EDITGENE to request a project consultation or quote.

Coverage recommendations are planning ranges; requirements depend on cell type, doubling time, effect size, and selection scheme.

Coverage claims should be reported with cell counts at each bottleneck and the fraction of guides detected at day 0.

• • Shalem O, et al. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343:84–87.
• • Hart T, et al. Evaluation and design of genome-wide CRISPR/SpCas9 knockout screens. G3. 2017;7:2719–2727.
• • Doench JG. Am I ready for CRISPR? A user's guide to genetic screens. Nature Reviews Genetics. 2018;19:67–80.