Peak calling and annotation.

Peaks were identified using MACS version 1.4.0rc2 according to the parameters stated below (26). The following command was used: macs14 –t <ChIP>.sam -c <input>.sam -n <output> -g mm -B -S. For analysis of CAP-D3 localization, only peak locations that had 10 or more tags, a P value of at most 10⁻⁵, and a fold enrichment of at least 4 were used for subsequent analyses. Reads and peak locations were visualized using Integrative Genomics Viewer (IGV) (57, 58).

Peak enrichment per genomic region was determined using CEAS and the mm9 RefSeq table supplied with the program (27). To determine the number and significance of peaks overlapping with other peaks or genomic elements, BEDTools was used (59, 60).

To create heat maps of reads, bamCoverage was first used to generate normalized bigWig files of ChIP and input reads using reads per genomic content (RPGC; 1× normalization), followed by bigwigCompare to subtract the normalized input signal from the ChIP signal (61, 62). computeMatrix was then used to calculate read enrichment scores at promoters with wild-type CAP-D3 peaks, and plotHeatmap was used to plot those scores.

Functional annotation of genes of interest was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (30). Genes were submitted to DAVID and analyzed using the functional annotation tool with the following categories selected: GOTERM_BP_DIRECT, GOTERM_CC_DIRECT, GOTERM_MF_DIRECT, UP_KEYWORDS, UP_SEQ_FEATURE, KEGG_PATHWAY, BIOCARTA, INTERPRO, PIR_SUPERFAMILY, and SMART. Functional annotation clustering was then used to cluster all of the gene enrichment results, with default parameters. The enrichment score for the annotation clusters represents the geometric mean, in –log scale, of the members’ P values in each corresponding annotation cluster. Functional annotation clusters were manually described based on their overarching theme.