Entering edit mode
6.8 years ago
ATpoint
81k
Hello,
we downloaded data from a WGBS study. They performed µWGBS in several replicates per cell type, using 10-1000 cells per library prep. The coverage for each CpG per replicate is low (mean about 1.3x). Therefore, I think it makes sense to first merge all replicates per cell type, then to normalize for different coverages between the samples, followed by downstream analysis. Is this a reasonable way to treat low-coverage replicates in WGBS?
