Hi, I'm using rtracklayer with RangedData and Rle to analyze numerical tracks from wig and bigwig files. Ideally I'd like to convert the RangedData that comes from import.bw into Rle format; however, the only way to do this (quickly) appears to be the IRanges function coverage(x, weight = 'score'), where x is a RangedData object. The problem is that coverage() only accepts integer weights. This works fine for integer tracks (e.g. counts) but does not readily work RangedData with a floating point "score". My interim solution for non-integer RangedData has been a hack: I multiply the score of the input RangedData by 10^(precision), run coverage, and then divide the Rle output by 10^(precision). This is an ok fix for now, however in the long term could cause ugly rounding problems (eg for tracks with more than 8 log orders magnitude difference between the smallest and largest score +/- high degree of interval overlap). Any suggestions for better approaches for dealing with this problem? Session info below. Thanks in advance, Marcin R version 2.14.0 (2011-10-31) Platform: x86_64-unknown-linux-gnu (64-bit) [[alternative HTML version deleted]]