Possibly primer issues. Are your primer pairs different for the untreated DNA and bsDNA? Perhaps your primers to the bsDNA are not working well in your specific samples due to partial methylation.

Hi Ross,

no the primer have been designed for amplifying only treated DNA and indeed they can not amplify the untreated DNA. They should instead detect both methylated and unmethylated BS treated DNA

Better you try to optimize on denature step with alternative 1 or 2 based on manual protocol of EZ DNA mehylation gold kit. Previously i tried with common denature step but my PCR product didn't work well and i got better result with alternative 1. May be you should blast your primer first with database to predict which are your sequence target

Hi Ross,

When I develop primers I always develop 2 F and 2 R primers both shifting a few bases. I noticed that the efficiency of the primers varies greatly with the TAQ that you use and the annealing AND extension temperatures. I use the qiagen pyromark kit which has a pretty good efficiency. I always start with optimization with unmethylated DNA, since this seems to be most troublesome to optimize. I start with a gradient from 60 to 50 degrees annealing and strangely enough an extension temperature of 63 most often seems to be the best temperature. I also do not include CpGs at the 5' anymore.

The cycling condtions are mostly 30s den, 30 an, 1 min ext. If you see a nice specific band with unmethylated DNA, try it with methylated DNA.

Often I use the HRM QPCR kit of qiagen (similar family of TAQ) to see if the Cq values are within a few cycles between unmeth and meth DNA, AND one or two samples, after optimization. You can also check the melting curve here to see if you have to distinct templates between meth and unmeth DNA.

Always include a calibration curve of unmeth towards meth DNA with your analysis. Although the efficiency of unmeth vs meth DNA seems OK with PCR, their can always be a preference towards either templates in your samples, biasing your results.

Hope this helps a bit.

Jorke

Hi Jorke this help a lot! I was just wondering a couple of things. With the HRM qpcr kit do you perform real time pcr or for HRM? The HRM instrument we had in the institute is out of work since a while so it is not possible for me to use it, but I was planning to perform a qPCR thanks to which I expect to compare PCR efficiencies of full methylated and unmethylated DNA, shouldn't I? Do you have any suggestion for qPCR?

I'm using now an applied biosystem sybr green master mix for performing qPCR (I haven't seen the results yet), but if you could give me more details in the product you have been using it would be great.

I have to say that it was for me kind unexpected but, as in your experience, my unmethylated DNA is less prone to be amplified compared to the full methylated one. 

Hi Gianluca,

Yes I did, since we have one. Generally, if you have a QPCR machine, you can easily run a melting curve even without a HRM option. You should be able to distinguish methylated from unmethylated DNA.

Although the PCR efficiency will differ, as Mikeal also suggests, best thing is actually make a standard curve from methylated towards unmethylated DNA and always include that. With bisulfite HRM or seq it is the preference of the PCR reaction that define the outcome. Very often you will not see a linear relation. What I have encountered often was when I added a little bit of methylated DNA to unmethylated DNA (say about 25%), melting curve analysis jumped straight to 100% methylation.

pyromark: https://www.qiagen.com/us/shop/automated-solutions/sequencers/pyromark-pcr-kit/#orderinginformation

hrm: https://www.qiagen.com/us/shop/pcr/real-time-pcr-enzymes-and-kits/qpcr/type-it-hrm-pcr-kit/#orderinginformation