Fig. 2

A Summary table of samples sequenced in E14 mouse embryonic stem cells indicating replicate (Rep), ChIP, total aligned reads, percentage duplication and number of peaks. B Analysis pipeline for calling bivalent peaks. C Genome browser view of reChIP datasets including IgG–IgG reChIP control (grey), in-line total H3K4me3 (green, rows 2 and 3), in-line total H3K27me3 (red, rows 4 and 5) and bivalent reChIP for H3K4me3 followed by H3K27me3 (K4-K27, purple, rows 6 and 7) or vice-versa (K27-K4, blue, row 8 and 9). Two biological duplicates (R1 and R2) are shown for all but IgG–IgG libraries. CpG islands are denoted by orange bars. Bivalent regions are highlighted in yellow. D FRiP scores showing proportion of reads within peaks for each individual sample. IgG–IgG (grey) is shown for each set of peaks to get background levels. E Comparison of peaks called using our reChIP method compared to in silico overlap of independently derived total ChIP-seq datasets F Single ChIP-qPCR for H3K4me3 (green) or H3K27me3 (red) at a H3K4me3 region (Gapdh1) or three bivalent regions (Tlx1, Pou4f1, Dlx3). Bottom panel shows corresponding reChIP controls: IgG–IgG (black), H3K4me3-IgG (light green), H3K27me3-IgG (light red). G Schematic of enrichment of first IP into the second IP when IgG antibody is used as the second IP in reChIP experiments. H reChIP-qPCR analysis of a H3K4me3-only region (Gapdh1) and three bivalent regions (Tlx1, Pou4f1 and Dlx3) in control (dark bars) versus cells treated with Tazemetostat (Taz) to reduce global H3K27me3 levels (light bars)