Brunello library (~77,441 sgRNAs, ~4 sgRNAs/gene, 1,000 non-targeting controls). Two biological screen replicates from independent transductions. MOI ~0.3 (≤30% transduction for single-integrant dominance). Cell coverage maintained at ≥500 cells per sgRNA at every passage (~4 x 10^7 cells minimum). Timepoints: T0 (just after puromycin selection, ~day 7) and T21 (14 days of competitive growth after T0). Genomic DNA from ≥4 x 10^7 cells per sample to preserve coverage in the sequencing input.
BSL-2+ for lentiviral work — perform all transductions in a dedicated Class II biosafety cabinet with enhanced practices; the Brunello library is replication-incompetent (3rd-gen) lentivirus but treat supernatants and transduced cells as infectious for the first 72 h. Decontaminate with fresh 10% bleach (≥30 min). Puromycin and blasticidin are toxic — handle with gloves and dispose as hazardous. Use a sealed-rotor centrifuge for spinoculation. Institutional Biosafety Committee approval required for lentiviral library use.
1,000 non-targeting control sgRNAs in the library serve as the neutral baseline (should not deplete; used for normalization). Core-essential gene set (e.g., Hart et al. reference essentials) is the internal positive control for dropout sensitivity; non-essential gene set is the internal negative control for specificity. Parental (untransduced) K562 + puromycin confirms selection stringency (complete kill). Two biological replicates control for transduction/growth variability. A T0 sample is the mandatory reference for fold-change normalization.
Library representation at T0 should retain >90% of sgRNAs detected with a Gini index <0.1 and read-count distribution skew ratio (top10%/bottom10%) <10, indicating intact coverage. Core-essential genes should show median log2FC < -2 by T21; non-targeting controls centered at log2FC ~0. Replicate-to-replicate sgRNA log2FC correlation Pearson r > 0.7. ROC-AUC separating known essentials from non-essentials should be >0.9.
Quantify sgRNA representation in a genome-wide Brunello CRISPR knockout screen at an early reference timepoint (T0) and a late dropout timepoint (T21) in Cas9-expressing K562 cells, and rank genes by negative selection (dropout) to recover known and novel fitness genes while preserving library coverage throughout.
Independent variable: time in competitive culture (T0 vs T21) for each sgRNA. Dependent variables: normalized sgRNA read counts, gene-level log2 fold-change, MAGeCK beta/RRA score, gene rank. Controlled variables: MOI (~0.3), cell coverage (≥500x maintained throughout), puromycin concentration (2 µg/mL), passage schedule, sequencing depth (≥300 reads/sgRNA), library lot, Cas9 expression maintained with blasticidin.
Maintaining >500x cell coverage per sgRNA at a low MOI (~0.3) will yield reproducible sgRNA abundance measurements such that essential genes (e.g., ribosomal, proteasomal, RNA Pol II subunits) drop out by >2 log2-fold-change at T21 versus T0, with the known core-essential gene set strongly depleted and non-targeting controls stable.
Trim/align reads to the Brunello sgRNA library with MAGeCK count (or count_spacers), producing a sgRNA count matrix. Assess QC: zero-count sgRNAs, Gini index, coverage. Run MAGeCK test (RRA) or MAGeCK-MLE with T0 as control and T21 as treatment, median-ratio normalization to non-targeting/all sgRNAs. Rank genes by negative-selection score and FDR. Benchmark with precision-recall against reference essential/non-essential gene sets. Visualize with rank plots and replicate-correlation scatter.
Poor library coverage at T0 (high Gini/many zero counts): coverage lost during selection — increase cell numbers to ≥1,000x and verify MOI. Transduction too high (>40%): multiple integrants confound — re-titer and reduce library volume. Weak dropout signal at T21: Cas9 activity low — verify with a control sgRNA (e.g., against a surface marker) editing assay and maintain blasticidin. Low replicate correlation (r < 0.5): bottlenecking — confirm coverage at every passage and increase sequencing depth. PCR fails on large gDNA input: split into more reactions (≤2.5 µg per 50 µL) and use Q5 Hot Start to suppress nonspecific products.
Gene-level significance from MAGeCK RRA permutation-based p-values with Benjamini-Hochberg FDR correction; call dropout hits at FDR < 0.05 and median log2FC < -1. Two biological replicates are analyzed jointly within MAGeCK-MLE (sgRNA as replicate measurements). Report precision-recall AUC and the number of recovered reference essentials. Coverage of ≥500 cells/sgRNA and ≥300 reads/sgRNA underlies the statistical power; with ~4 sgRNAs/gene the per-gene test integrates multiple guides to control noise.