14 RAQs about digital PCR

Have you ever spent hours googling that weird vibration your car makes? The itchy skin on your left hip that comes only at night every five weeks or so? How come the signal of clusters from your dPCR assays looks strange?

Sometimes, your question seems so rare and specific that you can’t find it on any FAQ lists, educational pages or Reddit forums.

We can't tell you if you need to check the air pressure in your front tires or which ointment will clear up your skin condition, but when it comes to dPCR, we got you covered. Our dPCR expert, Dr. Jan Rohde, took extra care to answer 14 of these rare head-scratchers during our “Ask me anything: Digital PCR edition” webinar. See if your unique dilemma made this list of rarely asked questions (RAQs).

1. Is it possible to perform touchdown PCR on digital PCR?

Theoretically, you can run a touchdown PCR as the cycling parameters can be flexibly adjusted on the QIAcuity® Digital PCR System. I haven’t seen a complete touchdown (with different annealing temperature for every cycle). Still, we do have customers who run a certain number of cycles at higher/lower annealing temperature first and then change the temperature for additional cycles. 

Different Tm per well is not possible, as the QIAcuity usually runs with a single temperature across the whole block. We have an essential gradient function, but it’s not flexible enough to accommodate zones of different Tm. 

Different Tm per cycle works when you program the software to treat the temperature changes as different subsequent cycling steps. 

2. In some dPCR experiments, I see the signal generated as a tail instead of a cluster of signals. It usually happens more with HEX. How could I avoid this and get a better and cleaner analysis?

Generally, a tail indicates some level of PCR inhibition or other issues with PCR efficiency with the reaction. It’s difficult to assess without looking at the details, but I have seen improvements in some cases when users switch to three-step PCR (separating the annealing and extension steps) or by modifying the annealing temperature. Another thing I recently observed was an improvement in the data upon using restriction enzymes to cut down the overall DNA size, even for samples with relatively low average DNA size, such as plasma.

3. How can I use dPCR to detect species in eDNA samples with high background noise?

When detecting a low number of target molecules from eDNA in high background noise, it’s a good idea to use positive and negative controls to characterize the parameters of the assay, such as the limit of blank and limit of detection (LOD). The LOD and limit of quantification (LOQ) can be improved by ensuring that the PCR assay works with high efficiency. We’ve observed that using the OneStep Advanced Probe Kit or Q-Solution Kit can help improve PCR efficiency, specifically for eDNA.

4. How can I use dPCR in the screening of transfusion-transmitted viruses?

When it comes to transfusion-mediated viral infections, you’re likely looking at low levels of the virus. I imagine the detection is similar to detecting cancer DNA in serum, meaning you’re looking at rare events. I was involved in a demo where we detected trace amounts of malaria parasite in patient blood (four copies of the genome). I imagine your application should work similarly well.

5. How can you perform a plasmid stability test using dPCR?

Plasmid stability is tested similarly to viral vector integrity. Basically, you design two or more assays for targets on the same plasmid and then measure them simultaneously. If both assay targets are detected in the same partition (so on the same molecule), then the plasmid is intact. If only one is there, the plasmid is truncated. As plasmids are circular, going for the ends is not really possible (since they have none), so I would shape out the assays somewhere along the plasmid or linearize the plasmid first and then go for the created ends (linearized plasmids also generally perform better in all PCR types, including digital PCR).

6. To analyze mRNA transcripts, is it better to perform the retrotranscription directly in the plate?

On the QIAcuity, you can do the reverse transcription off-plate (two-step) or directly on the plate using a QIAcuity QIAcuity OneStep Advanced EG Kit or a QIAcuity OneStep Advanced Kit for probes. There are advantages and disadvantages to both approaches. The OneStep mainly saves time when screening many samples at once. It’s also easier to count back to your original sample as you’re measuring the RNA molecules directly rather than first converting them to cDNA in another reaction and then having to calculate potential dilutions, etc. 

We’ve also seen examples of reasonably crude cell lysates working well with the QIAcuity, eliminating the need for sample purification and saving even more time. The downside is that RNA is less stable than cDNA. Hence, the two-step approach is better for safekeeping, allowing you to measure additional targets later. 

I usually recommend the OneStep process if you have a high number of samples and a finite number of target genes, whereas the two-step approach is better for lower sample numbers but a higher number of genes of interest.

7. If I’m using the same concentration of ssDNA as dsDNA, how would this impact the expected results?

This is a tricky question as it depends on what specific ssDNA or dsDNA you’re using. But let’s assume both are smaller strands of approximately 5 kbp. In this case, the dsDNA would transition into the partitions and both strands end up in the same partition to be counted as one copy. The same goes for the ssDNA. 
The two strands will separate if we denature the dsDNA before applying it to the nanoplate. If we manage to keep them separated until the partitioning is done, where the chance of re-annealing depends on the DNA size and sequence, the strands will move to separate partitions so that each copy of the dsDNA generates an output of 2 copies/µl, one per strand.

Each strand of the dsDNA would be amplified in qPCR, which is why there can be discrepancies between dPCR and qPCR depending on whether you separate the strands in dPCR before partitioning. The same is true for multicopy genes that are close to one another. These tandem copies might end up in the same well in dPCR and be counted as one copy, as the dPCR cannot distinguish between the number of target copies per partition. Digital PCR can only differentiate between the presence and absence of the copies (the quantification is based on summing up that information from the thousands of PCR reactions run in the partitions).

8. Tips for using dPCR to validate a sequencing dataset?

You’ll require specific assays for all targets identified in the NGS screening. Then, you can check (with higher precision) whether you can find the differences between the populations you saw in NGS again.

9. Do you have any advice for designing competitive probes?

When designing competitive probes, I suggest using LNA-modified oligos in the positions that differ between the probes (the SNV position or the indel). This increases the annealing differences between the probes. Be aware that when using a target of very low abundance, it’s normal to have a high confidence interval (CI). This is caused by the high effect of random chance at low copy levels. For example, if you have a solution with 1 cp/µl and take 3 µl, it’s completely random whether you end up with 0,1, 2, 3, 4, 5 or even 6 molecules in the partition. 

Using a positive control of known quantity and higher abundance can help immensely in this case, as it gives you the certainty that you’re detecting your actual target. You could order synthetic oligos with the sequence or construct plasmids with the target to use as a control. 

In case of separation problems, please note that when using Cy5, you could experience a slightly higher background fluorescence. This can be problematic if the assay you’re using in the channel isn’t performing optimally, as separating positives and negatives becomes difficult. You could try to alleviate this by using an assay with better performance and a more abundant target in the Cy5 channel. 

10. Could you compare digital PCR to DNA metabarcoding when it comes to eDNA analysis?

In metabarcoding, you use primers to amplify conserved regions out of a large population and then run sequencing on them to distinguish the variable regions between the primers to determine species. In dPCR, you’re not sequencing, you’re always looking for known sequences. So, you would have to have specific assays for each species you’re trying to identify, which isn't very practical due to many potential species.

What dPCR can do better than the sequencing approach, though, is quantify the numbers of known organisms and detect smaller quantities with more sensitivity. I see dPCR as a tool for following up on the sequencing results. For example, after the barcoding finds interesting species, you can quantify the number of these species in the population and monitor if their abundance changes over time.

11. How can I find suitable primers for an EvaGreen-based dPCR assay used to examine the integrity of CRISPR-targeted alleles?

You can find the best practices for designing primers in the QIAcuity Application Guide. These practices follow the dPCR MIQE guidelines. We also sell many EG-based assays for gene expression,  microRNA detection and copy number variation. We’ll soon launch a custom copy number variation tool, which you could use to design assays for your CRISPR application (although these assays are probe-based to facilitate multiplexing).

There can be multiple explanations if you see discrepancies between the copies measured in the same sample for similarly designed PCR assays. One would be that, as you work with EvaGreen, the assays run on the same sample but in separate wells. Depending on how accurate your pipettes are, you would expect some differences in percentage here.

The PCR efficiency of the primers can also affect the measured concentration if it’s really low. For CRISPR, it’s also worth considering whether the integrity of your target sequences varies. For example, suppose you design an assay that covers a region that could be modified during your genomic modification. In that case, your primers won't bind properly after the modification, whereas primers for a more stable region might still work fine.

12. When and why would someone want to do radial and amplitude multiplexing?

Radial and amplitude multiplexing are secondary multiplexing strategies used to increase the number of targets that can be analyzed per well of a digital PCR reaction. With digital PCR platforms such as the QIAcuity , allowing up to five targets to be analysed per well, radial and amplitude multiplexing would be unnecessary and expensive if analysing less than five target per well. If more than five targets need to be considered, amplitude-based multiplexing can work on the QIAcuity, based on anecdotal customer examples, but not internal validations.

13. What’s the interaction between eDNA and PCR inhibitors and how can I interpret positive eDNA detection when inhibition is known?

Environmental DNA often contains PCR inhibitors. These can impact all PCR reactions, though dPCR is less affected by inhibitors due to its end-point detection and only looking for a positive/negative distinction for each of the thousands of reactions run. So, if the sample is known to contain inhibitors and comes up negative in qPCR but positive in dPCR, your sample might contain the target after all.

One way to check for the impact of PCR inhibitors on the dPCR is to use a spike-in of known amounts and an increasing amount of the sample matrix (1, 2, 10 µl...). Usually, inhibition increases when more sample matrix is added. Ideally, if available, one would use a spike-in of the same sequence as the target. Using a different synthetic spike-in isn’t 100% comparable to the actual targets (as the primers and their binding abilities differ). However, it can be more cost effective when looking for multiple targets and can still give a good indication of inhibition.

14. How can I define the appropriate input RNA amount when measuring a leukemic fusion RNA during the drug treatment course from at diagnosis where the fusion copy number is very high to the MRD phase where the target copy number is very low?

Run different sample dilutions if you’re unsure how much target RNA is in the sample. In my experience, using undiluted, 1:100 and 1:1000 usually has at least one of the three dilutions in the dynamic range of dPCR.

Was your question number 15 and didn’t make this list either? Don’t despair. Hit us up on social media, keep checking our webinars for future “Ask me anything” sessions on digital PCR, or follow the blog for simpler answers to fascinating questions in future posts.