Question about minimum coverage

[desmodus1984]

Hi,

I am interested in finding differentially methylated sites and DM- loci or regions, and the output of MethylDackel is what many programs need. To confidently call DMS or DML/R, it is recommended to only use those with a minimum depth of at least 10 reads. I did process my samples and later I found that there is a similar option for minimum depth: absolute minimum coverage.
When I read about it, I am a little confused, because it states: "absolute minimum coverage, here, defined as the number of methylation calls kept after filtering for MAPQ, phred score, etc.)". I am interested in the number of reads per CpG, and this is the number of methylation calls. It seems that changing it might bias the results, because in my case, I work with an insect with very low average methylation, different from mammals which have a bimodal distribution; in my case is one a very left-skewed beta-binomial distribution with a long tail.
Thus, tell me if I am wrong, setting for instance, setting --minDepth 10, will only retain sites with 10 methylation calls, and then I will lose plenty of sites because methylation levels are very low.

Looking forward to hearing from you how to tackle this issue.

Best regards;

Juan Pablo