Cell lines with varying methylation levels

[have...@gmail.com]

Hi,

I am comparing methylation levels between a wild-type cell line and a cell line that has been confirmed to have little to no methylation. After running the samples through methylCRF and viewing the .bedGraph files in UCSC Genome Browser, I found that the methylation levels of both cell lines to be almost the same.

Would methylCRF have problems estimating methylation levels in a sample devoid of methylation?

[michael]

Hi,

That's interesting.  Yes, I think it might have a problem with very low levels of methylation.  Because MeDIP-seq read count varies between experiments, we normalize the 75th percentile of read counts to an arbitrary number.  It should still be able to pick up differences between cells, though.

When you checked on the browser, were the tracks auto or constant scaled?  Also, you can check your assumption about low/no methylation in your cell line by comparing just the MeDIP-seq coverage between the two cell types.   The MRE-seq coverage should change dramatically too. 

I hope this helps.

michael