Supplementary Figure 4: Validation of genome-wide screen results using focused sublibraries and individual gene KO cell lines.

a, Comparison of genome-wide and focused sublibrary screens. Scatterplots of confidence scores for genes in the focused sublibrary batch retest screens and original genome-wide results across various substrates. Confidence scores are calculated by casTLE and are negative if the gene KO inhibits phagocytosis and positive if the gene KO promotes phagocytosis. Hits discovered in the genome-wide screen for a given substrate are highlighted in red. Genes that were present in the batch library but were not hits for the given substrate in the genome-wide screen are in gray. Note that the focused sublibrary screen reveals phenotypes for many genes that were missed in the genome-wide screens (false negatives). b, Representation of the number of hits for each focused sublibrary screen. Hits are defined as the gene KO having a nonzero effect on phagocytosis within a 95% credible interval. c, Summary of all microscopy-based zymosan validations. U937 cells containing the indicated guide were treated with pHrodo-labeled zymosan and phagocytosis was monitored over time using automated microscopy. The phagocytic index of the time point at 5 h is presented. Data are the average of four independent replicate wells, and error bars represent the s.e.m. d, Summary of all microscopy-based positive bead validations. U937 cells containing the indicated guide were treated with pHrodo-labeled positively charged beads and phagocytosis was monitored over time using automated microscopy. The phagocytic index of the time point at 5 h is presented. Data are the average of four independent replicate wells, and error bars represent s.e.m. e, Validation by magnetic separation of magnetized zymosan and positively charged beads for predicted phagocytosis-promoting hits ICAM1 and CCNC. U937 cells containing the indicated guide were treated with magnetized zymosan or magnetic positively charge beads. Cells were separated by magnetic column as conducted in genome-wide screens. Cells in the unbound and bound fractions were counted and the fold enrichment of the bound:unbound cell ratio was calculated. f, Validation by magnetic separation of magnetized zymosan and positively charged beads for the predicted substrate-specific hits PLEK and TLN1. U937 cells containing the indicated guide were treated with magnetized zymosan or magnetic positively charge beads. Cells were separated by magnetic column as conducted in genome-wide screens. Cells in the unbound and bound fractions were counted and the fold enrichment of the bound:unbound cell ratio was calculated. g, Validation of sgRNAs editing target genes in U937 cells. Genomic DNA from U937 cells containing the indicated guide was submitted to Sanger sequencing at the location of sgRNA homology. The percentage of edited cells at the sgRNA target location was calculated using the ICE software package. Sanger sequencing data are available upon request. h, Validation of sgRNAs editing target genes in clonal U937 and RAW 264.7 cells. Genomic DNA from U937 and RAW 264.7 pooled cell populations containing the indicated guide was single-cell sorted, clonally expanded and submitted to Sanger sequencing at the location of sgRNA homology. The percentage of edited cells at the sgRNA target location was calculated using the ICE software package. Sanger sequencing data are available upon request. i, Validation of sgRNAs editing target genes in primary human macrophage cells. Genomic DNA from primary cells containing the indicated guide was submitted to Sanger sequencing at the location of sgRNA homology. The percentage of edited cells at the sgRNA target location was calculated using the ICE software package. Sanger sequencing data are available upon request.