View Single Post
Old 05-25-2016, 02:17 AM   #121
fkrueger
Senior Member
 
Location: Cambridge, UK

Join Date: Sep 2009
Posts: 614
Default

In that case you might add in the other scenarios and see whether it makes a big difference.

When you look at non-CG methylation levels in general (such as from the summary report), do you see very high levels that are indicative of conversion problems?

Frankly we got mixed results from using this method of looking at the filled-in position. Sometimes the values were very low (e.g. around 0.2% for the Booth et al data), but it sometimes came back with 25% methylation at that position which was clearly some sort of artefact since the overall level of non-CG methylation was 1% or so. So yea I would be a little careful with the values you get from looking at this position. If you take more global values such as (non-CG?) methylation levels over CpG islands as a measure, or possibly methylation of chrMT you might get better estimates for non-conversion. Cheers, Felix
fkrueger is offline   Reply With Quote