I have a set of bed intervals (that correspond to genomic regions of open ATAC NFR regions). I also have ChIP-seq bigwig files for a histone mark that have been normalised to input with Deeptools2' bamCompare. I would like to use FeaturesCount to count the number of normalised reads in each interval. However, I need a bam file as an input. How can I convert from Bigwig or use another software.
I suggest that instead of trying to get counts from the bigWig files you instead use alignmentSieve to generate NFR BAM files. DESeq2 is then OK if you have particular regions you're interested in (e.g., peaks of NFR signal) but otherwise give CSAW a try. It's quite good at finding differentially accessible regions.