Filter on WGMS poorly covered regions #39
Description
Hello,
I'm using DSS to analyze WGMS data with replicates starting from Bismark coverage.
Among the resulting DMRs I have a small region which results to be covered only by 1 or 2 reads per sample when looking at the bam files so the information supporting the differential methylation is not sufficient. In this case the problem comes from the occurence of repetitions in the region which lead to have many reads classified as multimapping and so they are discarded by Bismark.
I saw that in the DMLtest function only the zero covered positions are discarded and that it is not required that all replicates of a conditions have coverage for a position to be retained so the risk is to have false positive results as happened with this data.
I was wondering if you ever considered to introduce a parameter to be able to modify the filter on the coverage or if you would suggest to manage the input before running the analysis (e.g. filtering the coverage file from Bismark).
Thank you.
Best regards,
Sara