Fig. 1: Genome-wide CRISPR loss-of-function screen to identify factors that affect the insulator function of CTCF, complemented with biochemical approaches.

a, Layout of genetic and biochemical approaches for identification of candidates influencing the insulation function of CTCF. b, Layout of the genetic loss-of-function screen that separates MNs with a CTCF-boundary disruption from those with an intact boundary. RA, all-trans-retinoic acid; SAG, smoothened agonist. c, Rank of genes underrepresented in ESCs compared to the plasmid library. Cutoff line indicates FDR < 0.05. d, Rank of genes underrepresented in MNs compared to ESCs. Cutoff line indicates FDR < 0.05. e, Rank of genes overrepresented in double-positive MNs compared to mCherry-positive MNs in four genome-wide screens. Top candidates are listed for each screen (all candidates are listed in Supplementary Dataset 2). One of the top candidates is indicated on the plot in each independent screen. Lib., library. f, Venn diagram showing the overlap of CTCF-boundary-related candidates identified in four independent screens (two for library A and two for library B). P value cutoff = 0.05. g, Crosslinked FLAG-CTCF ChIP-MS in ESCs and MNs results in identification of known CTCF interactors and novel proteins. The peptide counts in FLAG-CTCF immunoprecipitations were normalized to control FLAG immunoprecipitations in untagged cells. The list is ranked based on CTCF immunoprecipitation/control ratios in MNs. IP, immunoprecipitation.