Microfluidic PCR Issues, Master Mix and Consumable Variations? - (May/01/2014 )

I am a mechanical engineer on a project team that is develping a microfudic PCR devce.  We analyze the PCR products using an end point fluourescence with either probe based primers or EVAgreen dye.  The reagent mix is force flowed through a series of microfluidic channels on a polypropylene microplate, and then partitioned off in small volumes on the surface of the microplate.  These partitions are then thermally cycled to completion (30-40 cycles).  We are finding that the reactions are losing efficiency, or totally eliminating amplification, in the partitions that are far from the source feed.  This is not a totally repeatable process, even though all of the volumes are consistent, and velocities are fairly consistent, etc.  We have tried to use different PP resin stocks and some seem to perform better than others.  Also, some master mixes perform better than others, while some do not amplify at all.  We use the same reagents in a standard QPCR reaction and are able to get results that are as expected, but when we use our microfluidic platform we have the previously described issue. 

Our working hypothesis is that some part of the reaction mixture is sticking to the PP surface and is thus being used up by the time the fluid reaches the further partitions.  We have added BSA and it does seem to help (but not eliminate, regardless of concentration), theoretically acting as a sacrificial molecule to stick in place of the critical reagents. 

We are using a custom designed injection molded, polypropylene microplate consumable.  The plate has carbon black masterbatch added to eliminate the background flourescence of the PP. 

I am basically looking for help from people who know more about biochemistry or materials science than I (or my team) do. 

I am aware that this is a difficult process to describe in a few sentences, so may be difficult to visualize.  I think the important variation from standard vial based PCR is the extremely high surface area to volume ratio of our device.  We have tried many things over the last few months, so I will answer any questions that you have.

Thanks in advance.

-Stephen Y.

PS.  I was not sure what forum to post this in, so it is also in the molecular biology section.

Edit - It's alright S. Young! I went ahead and deleted the duplicate post in the Molecular Biology forum. I think it is better suited in this subforum (maybe not) and I am sure someone will view/answer your question shortly. -JerryShelly1

Are you sure that the partitions far from the source feed are getting the same volume as the ones closer?  If not - you could have an evaporation problem, either during the cycling or during the distribution.  It could be that that part of the plate is near a heat source (motor/pump? light source?) that is heating the reagents and evaporating or damaging the reagents.

Mechanical shear forces can damage long DNA - how big an amplicon are you expecting?

Hello, Thank you for the response.

We are sure that the partions all receive the same total volume (but not sure that everything is at the same concentrations).  We use a ROX referrence dye to make sure that volume is the same, and to normalize our results, using the fuourescence of the dye.  The ROX does not vary in intensity so it, at least, is at constant concentration for all partitions. 

Each partition is individually, and completely sealed as a single reaction. They are liquid only, with no room to evaporate unless the vapor pressure overcomes the seal or the fluid boils, but this has not been an issue yet.  We also can visually check he sealing of the partitions using a microscope, and have done so, and all partitions are remaining sealed throughout the process.  We are also careful about the handling of the plate so there should not be an issue as suggested with some portions of the plate receiving anything the other parts do not. 

The mechanical shear thing is interesting.  We are aware of this to a small degree, but do not know how to test this.  We do tend to get better results with genomic DNA, and our typical amplicon length varies but can be assumed to be in the range of 80-200 base pairs.  The geographic locations of the poor amplification are slightly predictable, with areas of the plate that are more likely to fail.  One area that was failing quite regularly was in an area where we thought there to be more shear forces, although the other areas are not in high shear zones. 

What does it take to shear DNA?  Is our typical amplicon length at risk?  The reagents do flow through small cross sectional area passages, and at a realatively high velocity to area ratio.

Thanks again.

-Stephen Y.

Also, We are do have the ability to run tests if any are suggested that may lead us closer to an answer (if we haven't already tried it).  So if there are any test suggestions.........

Short amplicons aren't a problem usually for shear forces, it is usually the target DNA that it is the problem here - it gets damaged and won't amplify (unsurprisingly), but short amplicons are often not affected because the chance of the DNA being sheared in the spot that is amplified is relatively low.

I don't know much about the mechanical strength of DNA, but it is common in labs to take special measures to protect DNA when extracting (cutting tips so that they are wide bore in attempt to reduce shear), you can also shear chromosomal DNA by passing through hypodermic needles.

You may be able to test the integrity of the DNA by passing through a gel on a chip sort of thing - but it would depend on the detection method and how much DNA you are using.  Increase the amount and test??

Have you thought about siliconizing the surface?