Fig. 1
Genomic distribution of CFP1 relative to CpG islands, transcription start sites (TSSs) and marks of active and repressed chromatin. a Heatmap plots of human genomic 6-kb regions centred on TSSs (top) and CGIs (bottom) in erythroblasts (ERY) and Epstein-Barr virus transformed lymphoblast (EBV) cells. TSSs sorted in order of CFP1 coverage; CGIs sorted in order of size. Sequencing read depths shown for CFP1, H3K27me3, H3K27ac, H3K4me1, H3K4me3, DNase-seq, ATAC-seq and Pol II are net coverage after normalisation to 1x genome-wide and subtraction of an input data set similarly normalised. Gene expression (log2 RPKM) is shown to the right for TSSs in both cell types. b Location of CFP1 summits allocated in order of CGI TSS (< 1 kb), CGI genic, CGI intergenic, (non-CGI) TSS (< 1 kb), (non-CGI) genic and (non-CGI) intergenic. ERY, top; and EBV, bottom. c Venn diagrams showing CFP1 peaks within 1-kb of TSSs are strongly associated with H3K4me3 histone mark and poorly associated with H3K27me3 repressive histone mark. Cell types are ERY (upper) and EBV (lower). Public data sets: * NCBI GEO GSE36985, ** NCBI GEO GSE50893. d Mutual exclusivity of CFP1 and the H3K27me3 mark in ERY (upper) and EBV (lower) cell types