MV emPCR issue

[ajthomas]

I've been frustrated with emPCR for a while now because cpb ratios never seemed to correlate well between emPCR sizes. In my experience I found that a cpb ratio that yielded the desired 8-10% enrichment at the SV scale would give me anywhere from 25-50% at the MV scale. In addition, I seemed to get poorer data when using MV empCR kits. I could never figure it out until just recently.

I did a MV emPCR that looked somewhat broken when I took it out of the thermocycler. I've never had a broken emulsion before, so I was a bit surprised. I wasn't sure it was broken, either, because I had never seen one. There was just a thin layer of the aqueous phase at the bottom of the well with a few beads in it. To be sure it was broken I looked at a bunch of 1ul aliquots under a microscope. Sure enough, it was broken, so I threw it away. To make a long story short, I paid closer attention to the next couple of MV emPCRs, and that thin layer was there, but usually was only noticeable if you looked for it. Looking at aliquots from that bottom of the well versus the middle of the well showed that the emulsion was only partially broken; at the bottom of the well were vesicles with many beads and the rest of the well was filled with single-bead vesicles. I eventually figured out that the emulsion wasn't breaking, but instead was not being formed properly in the first place. Using the standard protocol there were a lot of vesicles in the emulsion with anywhere from 2 to several dozen beads. Icreasing the time didn't help, only increasing the speed of the last shaking step did. This last time I increased the shaking speed of the last shake to 20 Hz rather than 12 Hz and there were many fewer of the many-bead vesicles, and those that persisted were much smaller, but there were still some. The sequencing results were not great, either.

Has anyone else noticed inconsistent enrichment or sequencing results when doing emPCR at the MV size? If so, have you looked at the emulsion under a microscope and can you verify my experience? I'm trying to figure out if this is something unique to me, or if it is more wide-spread.

[RCJK]

I've always thought the MV rxns were less reliable and more prone to broken emulsions than SV or LV. They seem to give more variable enrichments as well.

Recently I've seen the same thing as you. Set up a bunch of MV rxns and the next day all wells had broken emulsions. A pellet of beads was visible at the bottom, then a clear layer, then a normal looking emulsion layer. I discarded all of these and set up some new ones for a couple of the samples with a different lot of oil. These were mostly fine, with a few broken wells here and there. I then set up some more MV rxns for my remaining samples, but this time I set up 2x MV per sample on the same plate. 1 rxn used the same lot of oil as the first set (lot 93834420), and the other rxn used the same lot as the 2nd set of rxns (lot 93869520). Within minutes of dispensing the emulsions, before even putting the plate onto the thermalcycler, I could see the emulsions made with oil lot 93834420 already separating and bead pellets forming. I did not see this with the rxns using the other lot of oil. Needless to say, by the next morning, all of the wells from one of the oils had broken emulsions while the other set was ok.

Roche says there are no known issues with this lot, but may be willing to replace the kit. I'm still waiting for a confirmation of this.