Validation of the platform with a Parkin localization screen, targeting kinases, phosphatases, and drug targets—gRNA pooled library (12,500 sgRNAs targeting 2,774 genes).(a) Schematic representation of the AI-PS platform. GFP-Parkin cells, pa-mCh, and dCas9-KRAB cells were transduced with a subgroup pooled sgRNA library. Cells with cytosolic GFP-Parkin (green color dispersed in cytosol) were photoactivated and sorted by flow cytometry and subsequently submitted to deep-sequencing analysis. (b) Flow cytometry sorted 385 cells from 77,114. Flow cytometry scatterplot representing the separation of the postscreen photoactivated from the inactivated cell population. BFP florescence signal y-axis in cyan, mCherry fluorescence signal x-axis in red. Total number of sorted cells was 2,227 (n = 3). (c) Fold-change threshold was computed from the noise model of nontargeting gRNA distribution. Mean Log₂-CPM fold change in the purple line, fold-change threshold at the red vertical line represents two SDs from the mean (n = 3). Data distribution was assumed to be normal, but this was not formally tested. (d) Enrichment plot comparing sgRNA abundance in the photoactivated sample following CCCP treatment to sgRNA abundance before treatment. Vertical red line set on log₂-fold change threshold; horizontal red line indicating the Benjamini-Hochberg corrected P value set on 5%. See also Table S1. The number of sgRNAs detected and filtered was 3,471 targeting 1,157 genes (n = 3). (e) Statistical power analysis by simulation on the 3,471 sgRNAs retrieved from the Parkin screen. The simulation was done using the R package PROPER (Wu et al., 2015). Effect size (log of fold-change) in the x-axis, power in the y-axis, the curved lines are color coded for the number of biologic repeats.