I have an IRanges object which contains regions of interest and read counts for a given sample in the metadata column. I would like to iterate through each of the regions and get the counts for regions which are within 1,000 bp of the current region. I have tried the following but It is too slow and I wondered if there were any faster alternatives?

for (i in 1:length(regions)) {
    iranges <- ranges(regions)
    range <- iranges[i]
    start <- start(range) - 1000
    end <- end(range) + 1000
    hits <- findOverlaps(IRanges(start=start, end=end), regions)
}

For counting the overlaps, use countOverlaps():

countOverlaps(regions, maxgap = 1000)