HRM analysis, equipment requirements (View forum version)


Posted 08 February 2012 - 07:46 AM

Hey guys!! I am new here and I would really appreciate your help in this issue. I want to study some methylation patterns with melting curve analysis. So we are gonna buy a real time instrument for that purpose. I would like to know if any real time equipment would work for it, or if it needs some special features. To be more specific, we are interested in this particular one do you think it will work for HRM analysis? Thanks in advance!!! :)


Posted 08 February 2012 - 01:58 PM

Methylation HRM? Certainly NOT any real-time cycler will do. Many claim the ability to do HRM, but results can be disappointing.
Carusel-based Rotor-Gene is considered the best.
Check this page.


Posted 09 February 2012 - 06:44 AM

Thanks for the answer Trof...but in that case do you think a "normal" real-time cycler is able to do some melt curve analysis, although not in "high resolution"? Could it be useful to study methylation patterns anyway?


Posted 09 February 2012 - 12:29 PM

HRM depends a lot on the uniformity of the temperature across all wells. I'm not 100% in the picture now, but few years ago only the best companies were able to make a block-based cycler, that had the required uniformity. That's why the air-operated cyclers like RotorGene have unparallelled resolution. But this is all on that page.
The required resolution varies depending on desired use. Bigged deletions for instance create big temerature changes, that don't require such high resolution. Single base changes are more tricky, there are 4 types of them and some of these are sometimes difiicult to spot even on the very good block-based cyclers.

Bisulfite treated methylation is a SNP type 1, with big (around 0.5 deg) temperature change, but it depends what you will be detecting. If it's a single clone then you will have homozygous sequences. Homozygotes differ only in the temperature shift, not the shape of the curve, which is more prominent. In variable block heating, this could be hard to distinguish. There is a way to overcome this, by spiking the sample 50:50 with unmethylated sequence, then you create a heterozygote, which usually have very distinctive shape. This could be enough for even average cycler to distinguish. But it's hard to say, I'm not sure I would risk it if HRM is my primary goal. At least look for block homogeneity, LightCycler 480 claims to have around 0.2 degrees, and even with that it's not 100% accurate in determining homozygotes in some SNP HRM we do, if we don't spike.

If you don't clone and analyse the real treated DNA, there of course you would need standards of different methylation percentage, that will create heteroduplexes, but there you would face the problem of telling 20 % from 50 % and so on. And you can't really overcome this. In this case I would care even more about the selection of instrument.