ABSTRACT: Genome-wide occupancy analysis of TBX5, NKX2-5 and GATA4 in differentiating WT, Nkx2-5KO (NKO), Tbx5KO (TKO) and Nkx2-5;Tbx5KO (DKO) cells at the cardiac precursor (CP) and cardiomyocyte (CM) differentiation stages. Analysis of TBX5, NKX2-5 and GATA4 occupancy a and gene expression in WT, Tbx5KO, Nkx2-5KO and DoubleKO precursor (CP) and mature (CM) in vitro differentiated cardiomyocytes. We performed ChIP-exo experiments for NKX2-5, TBX5 and GATA4 in differentiating WT, Nkx2-5KO (NKO), Tbx5KO (TKO) and Nkx2-5;Tbx5KO (DKO) cells at the CP and CM stages. ChIP-exo was performed according to methods published previously (Boyer et al., 2005; Serandour et al., 2013; Wamstad et al., 2012) using specific antisera to TBX5 (sc-17866 XS, lot no. B1213), NKX2-5 (sc-8697 XS, lot no. B1213), and GATA4 (sc-1237 XS, lot no. B1213) on chromatin isolated from 10 mill. cells. Re-ChIP-exo was performed from 40 mill cells, combining methods published previously (Serandour et al., 2013; Shankaranarayanan et al., 2011). ChIP-exo Analysis: Barcoded libraries were single-end 50bp sequenced on an Illumina HiSeq 2500 instrument. Reads were trimmed using the fastq-mcf program (Aronesty, 2011), and then aligned to the mm9 mouse genome assembly using bowtie2 (ChIP-seq, ChIP-exo) (Langmead and Salzberg, 2012). Samtools was then used to retain reads that had a mapq score of 30 or greater (Li et al., 2009), thereby ensuring that each mapped uniquely to the genome assembly. The 5' -most position of reads that mapped to the reference strand and the 3â??-most position of reads that mapped to the non-reference strand were identified for each read as the actual edges of each exonuclease-treated fragment. The genome was divided into 20bp genomic bins and a normalized tag density was calculated for each genomic bin 'i' as follows: tag density(i)=([#tags within 75bp of i] X [total# genomic bins])/(total #of tags) To identify broad regions of binding, bins with tag densities of greater than 100 were merged to generate a peak list for each sample. Within 1kb of each region, strand-specific single-base-resolution tag densities were calculated for each dataset by dividing each region into 1bp bins, then counting the number of tags within 5bp of each bin. For each region of binding, the footprint for each bound region was defined as the span from the peak position of '+' strand binding to the peak position of '-' strand binding as seen from the high-resolution tag densities. For subsequent analyses, any overlapping footprint for replica/factor/stage/genotype was merged to generate an average footprint list. Only average footprints present in at least two replicas of the same factor, stage and genotype were considered (Final Average Footprints).
