Hi AWG members,
Upon processing the first 4 MethylSeq datasets, we decided to make a few updates to the recently approved Methyl-Seq pipeline. You can find the latest updates here. Please click on the link and let us know if you have any issues with the updates no later than Sunday, Feb 23rd.
Updates include the following:
-
Added a filtering and normalization step prior to differential methylation analysis. For studies in which the coverage is fairly similar across the samples, this step will not change the results, but in the case of outliers it will improve the data quality. See Step 11e for the complete implementation details, in brief:
### Normalize the data ###
# First, filter samples by coverage to account for PCR bias or over-amplification #
meth_obj <- filterByCoverage(meth_obj, lo.count = 10, lo.perc = NULL, hi.count = NULL, hi.perc = 99.9)
# Normalize coverage between samples using a scaling factor derived from the median coverage distributions #
norm_meth_obj <- normalizeCoverage(meth_obj, method = "median") -
Combined all pair-wise differential methylation analyses into one output table for bases and one output table for tiles, rather than providing individual sets of files for each comparison and splitting between hypo and hyper methylated regions.
-
Provided the differential methylation output without filtering, rather than imposing significance cutoffs.
The new output format matches the output formats we provide for our other assay types.
You can review an example of the updated analysis outputs for OSD-47 here: MethylSeq_new_differential_methylation_outputs_240214 where you can see the new output files both with and without normalization applied.