Video Transcript
I’m Dominic Meir. I’m a field application scientist here at Pcrassays.com, and this is the first of what we hope will be many educational podcasts about PCR design in general. We’re hoping to cover good versus bad multiplex design, false positives, false negatives, and multiplex PCR assays, and the challenges that COVID-19 brought about for PCR assay manufacturers. I’m here with our CEO and founder John Santalucia.
Hi Dom.
How you doing? Today’s episode, we’re going to go over good and bad multiplex PCR design. So I think we should just jump right into it unless you want to do a little introduction.
Sure. I’m John Santalucia, the founder of DNA software. We’ve been in business now for 23 years. I’m very proud of the fact that not only have we served some of the major leaders in the field, but now we’re also manufacturing the kits ourselves, and validating them, and shipping them around the world.
Let’s just jump right into it. So let’s start with the common problems that arise when designing multiplex PCR assays. Let’s just go over kind of an overview of those.
Well, the major problems have to do with false positives and false negatives. In any assay you need to design it to have a minimum number of false positives and false negatives. And what causes those? Well, the major causes of false negatives are things like the folding of the target DNA and RNA. So if you design your primer unknowingly to a region that’s highly folded, that primer is not going to bind very well. In particular, if the organism mutates and now a primer that barely worked before, now it has one mutation, it’s not going to bind. So that can cause false negatives.
False positives can be caused by off-target hybridization, and some other effects. One in particular I think is important, is anything that leads to consumption of a primer will lead to a false negative or at least a reduced fluorescence in the assay. So for example, if the primer formed primer dimers, or if it formed primer probe interactions, or primer amplicon interactions, or primer interactions with something in the background like the human genome, then any of those processes will lead to the primer being consumed. And if the primer is consumed doing something you don’t want it to do, then that means the primer is going to be not available to amplify the desired amplicon. And that’s a major cause of the failure of PCR. Also, just inability of many users to multiplex the PCRs. So combining them to all, get them to play well together, is some of the major causes of false negatives and false positives.
With all those things considered with the traditional methods of designing multiplex PCR assays, how can we go about that and really try to improve those things?
Right. I think the traditional method of trying to develop a multiplexed assay is to first try to make singleplexes that work well on their own. And then to try to combine those singleplexes into larger and larger multiplexes until you get something that works. That’s the usual method. It’s highly empirical and highly limited in its scope, and usually it fails, particularly with larger multiplexes. So the reason it fails is that you’re trying a very small part of the space of possible multiplexes. So for example, let’s say you wanted to make a twentyplex reaction, so with 20 PCR reactions all in the same tube. Well, what can happen there is that if you have 10 candidate singleplexes for each member of those twentyplex, then the number of possible multiplexes is 10 to the 20th power. So that’s Avogadro’s number, a huge number, right?
And experimentalists in the lab can try combinations. Maybe they try candidate one of all the assays and they see that hey, two and three don’t work. So they replace two and three with the second choices on their list, and then now those work, but now four doesn’t work. And you replace four, now one doesn’t work. So you get a whack-a-mole kind of situation. And that problem gets even worse the larger the multiplex is. So you really need an algorithm that is able to search the entire space, and that requires a twenty-first century algorithm.
And we’ve been developing those for our twenty-three years of business. We’ve developed an algorithm called the Depth-First search. It’s a twenty-first century algorithm that allows you… Our algorithms can search the entire space of possible multiplexes to find the ones that have minimal primer dimers, minimal primer-amplicon interactions, minimal primer off-target interactions, et cetera, and have all of the things we talked about with regard to just getting everything to play well together. And also dealing with all the variants, that’s another aspect of the problem, dealing with the inclusivity, all the known variants of a given organism that you want to detect pathogen, and all of the near neighbors and the exclusivity and the background like the human genome and stuff. All those things are quite challenging to make software and it’s something that we’re world experts at.
So if you had to summarize the keys to good multiplex PCR design, what would you say those keys would be?
I would say first of all, that the singleplexes that most people develop are usually fragile assays. They can barely work. Often they’ll have to optimize the magnesium or optimize some conditions, temperature cycling conditions to get them to work. And that’s a symptom that they’re fragile. And it usually means they didn’t take into account something like folding of the target. So we have this state-of-the-art algorithm called the Multi-State Coupled Equilibrium Bottle. That’s a mouthful. But it just means we take into account the folding and the competition between folding and hybridization in our algorithm. So that’s one thing you need to do. Collecting high quality inclusivity database. So this is one of the unique aspects of our software approach is that it combines the science of hybridization… Which is what my academic lab was known for. Combines that with the database approach, the modern availability of all the sequences of variants.
So collecting a high quality inclusivity database for the design, accumulating a good exclusivity database for near neighbors, and then accumulating the background database. All those things are important. And then, as I mentioned, actually detecting all of the primer dimers and off-target effects, those require individual algorithms for each of those. In fact, for primer dimers, for example, DNA software has acquired a large reservoir of knowledge about primer dimers that we’ve actually discovered four new mechanisms. So beyond the mechanism you see in the literature for just the pairing of the three primers and two primers and then a polymerase extending those, there are four additional mechanisms that are completely different that result in a product that looks like it’s from two primers.
It involves multiple steps to get to those products, but those often get very bad when you do a multiplex. So those are the keys. Minimize primer dimers, use state-of-the-art software that takes into account the full spectrum of variants. And you need high throughput computing to do that, by the way. You can’t do that on a desktop software. You can’t use some sort of… Freeware is just not up to the task of doing that sort of thing.
It’s very impressive. I would say, have you published any literature related to documenting the process of this higher level multiplex PCR design?
Yes. We published a significant paper with a variety of government agencies. So it was called the SPADA Working Group, which was a group of scientists from the FDA, the CDC, the Department of Defense, Los Alamos National Labs. And we were the only private sector company that was involved in that effort. But I was actually the leader of that group. And we wrote a paper on that that is publicly available, it’s on our website. And anyone who contacts us, we’d be happy to send you a copy of that. And we also have a number of white papers and other webinars that we’ve done that sort of explain our gospel of the world about how to do design and how to do things.
Awesome. Well, thank you for your time today addressing all of this. I think we all learned a lot, me included. This is really, like we said, this is kind of the first of many of these educational webinars we want to put together. Next, I think we’re going to talk about trying to limit those false positives and false negatives that you discussed. And especially considering the third episode, we’re looking at talking about COVID-19, and you mentioned the variance and how hard it can be to design around those. So I look forward to getting into those topics as well.
All right, super. Thank you everyone for attending.
Thank you.
Bye-Bye.