Fig. 2: IntAC and dU6:3 drive much of the performance increase in an updated Drosophila CRISPR screening library.

A Overview of machine-learning optimization of a new (‘v.2’) genome-wide CRISPR screening library for Drosophila. Elements created in BioRender. Viswanatha, R. (2025) https://BioRender.com/y0pxkoa. B Schematic representation of transfection of the v.2 library. 92,795 attB-flanked sgRNAs targeting Drosophila melanogaster genes were transfected together. sgRNA expression was driven by the strong dU6:3 promoter, and the library was co-transfected with the IntAC system (a plasmid expressing φC31 integrase and AcrIIa4) into cells. Cells were passaged for approximately two months, and then CRISPR sgRNA distributions were assessed by next-generation sequencing. Elements created in BioRender. Viswanatha, R. (2025) https://BioRender.com/2v5d13e. C sgRNAs common to v.1 and v.2 libraries were analyzed to isolate the effect of IntAC and dU6:3. D Comparison of 17,948 sgRNAs shows dramatically increased, consistent dropout of sgRNAs in v.2 relative to v.1. Plot shows normalized log2 fold-change of gene-targeting (blue) or intergenic (red) sgRNAs. E Gene-level analysis of the sgRNA log2-fold changes in D shows that the increased sgRNA dropouts translate to increased fitness gene calls (genes with negative Z-scores). F Cumulative distribution of false-discovery of fitness genes (genes with low expression, FPMK < 1) from v.1 (green) or v.2 (purple) gene Z-scores in E shows that IntAC and dU6:3 increase the detection of fitness genes. This figure was created in part with Biorender.