I was using BSseq package to detect DMRs in non-human organism (chicken), I have a few questions that have confused me for some time:
1. The BSseq package reference manual states that the parameter "h" in smoothing is the minimum smoothing window, however, the definition of "h" in "Analyzing WGBS with the bsseq package" seems to be half the smoothing window, as stated "h is half the minimum window with (the actual window width is either 2 times h or wide enough to contain ns covered CpGs, whichever is greater)" (Page 4 in the PDF file). I was a little confused, I wonder which statement is more accurate? If I want to set the smoothing window to 1000bp, should I set h to 500 or 1000?
2. The raw coverage of my data is 30X, but after read mapping, remove duplicates, keep only uniquely mapped reads, the coverage is 14~16X across samples. Is it reasonable to include few CpGs, say 10~30CpGs, in the smoothing window, since BSmooth is developed for low coverage WGBS data and our coverage is relatively higher?
Any suggestions will be appreciated !