Sequencing FASTQ reads were aligned to the hg19 human reference genome with Bowtie2 (52) with the following restrictions “--very-sensitive --score-min L,-0.6,-0.15” for mapping sensitivity and accuracy. Only uniquely mapped reads, and in the case of paired-end sequenced data, only uniquely mapped sequence pairs that were aligned concordantly, were used for subsequent analyses. In S100A8 or S100A9 ChIP-seq data, confident peaks were defined as fulfilling all the following restrictions: (i) MACS2 peak-calling q value ≤1 × 10−7 against sample-corresponding chromatin input background; (ii) peaks called by MACS2 were identified in at least two biological replicates; (iii) at least five raw reads were aligned at the peak region. Depending on the comparisons indicated, peak regions from different datasets were merged if overlapping at least 1 bp. Confident Pol II ChIP-seq peaks were defined as MACS2 paired-end peak-calling P value ≤1 × 10−8 against genomic background of 1-Mb genomic region around. In ChIP-seq data of histone modifications and DNase-seq data, confident sites were defined as previously described (21). Meta-gene and heatmap analyses were carried out with deepTools (53).