bio-crispr-screens-bagel-essentiality

What problem does it solve? CRISPR-Cas9 fitness screens produce per-sgRNA count data that must be converted into defensible per-gene essentiality calls; doing this by hand risks miscalibrated thresholds, non-reproducible results from unseeded resampling, and false tumor-suppressor calls from dropout-only screens. ## Core Features & Use Cases - End-to-end BAGEL2 workflow: Runs the fc + bf + pr pipeline to compute per-sgRNA fold changes, per-gene Bayes Factors against CEGv2/NEGv1 reference sets, and precision-recall curves for empirical threshold calibration. - Reproducibility and failure-mode guards: Enforces fixed -s seeds (unseeded runs flip dozens of gene calls), detects swapped -e/-n references (all-nan BF output), and restricts tumor-suppressor calls to enrichment-designed screens. - Cross-method reconciliation: Compares BAGEL2 hits against MAGeCK and drugZ, with per-sgRNA BF contribution analysis to catch guide-of-one artifacts. - Use Case: Given a Brunello-library dropout screen count matrix, compute fold changes versus the plasmid control, score all genes with a seeded bootstrap run, and report essentials at BF >6 calibrated by the screen's own PR curve. ## Quick Start Run BAGEL2 on my counts.txt with Plasmid as control, compute Bayes Factors against CEGv2 and NEGv1 with a fixed seed, and pick the BF threshold from the precision-recall curve.