Add Multiplex Cc Number with airSlate SignNow

Add multiplex cc number

greetings everyone and welcome to another genetic engineering and biotechnology news webinar our presentation today is entitled multiplexing PCR design made easy I'm Jeff ballistas technical editor for gen and now I'll be the moderator for today's webinar presentation walk into any molecular biology lab these days and on any given day and invariably there'll be a PCR reaction on going while PCR has been put to good use over the past 40 years and it's ubiquitous nature makes it indispensible towards daily laboratory operations the technique is anything but mundane as researchers and companies are always finding new ways to improve the reaction recently investigators have begun to capitalize on the fundamental nature of PCR through multiplexing reactions saving them valuable time and money though multiplexing reactions can be a benefit to lab productivity they're not without their drawbacks the added parameters of multiplexed PCR reactions can dramatically increase the complexity of experimental design and validation to frustrating levels for many researchers thankfully a host of computer software has been created to ease molecular biologists multiplexing design woes let's meet our speakers for this webinar will tell us how panel Plex software has allowed them to quickly and efficiently design panels for various disease identification and molecular diagnostic purposes our zhenka CB is the founder and CEO in celexa headquartered in Sunnyvale California dr. hiss EB will tell us how panel Plex has helped in celexa develop a screening platform for upper respiratory pathogens Addie Thea rajagopal is founder and CEO of chroma code located in Carlsbad California dr. Rajagopal will tell us how panel Plex was critical in the development of a multi viral molecular diagnostic tool finally John Santa Lucia is the founder and CEO of DNA software located in beautiful Ann Arbor Michigan dr. san lucia will tell us about some of the philosophy science and pitfalls behind PCR primer design and how panel Plex has symp find the process before our speakers get started I want to encourage the audience to submit questions for a Q&A session at the end of the presentations we'll try to answer as many questions as we can so simply type your question into the box below the presentation screen and hit submit ok with that let's begin our webinar in here what our speakers have to say RJ you have our attention thank you Jeff for the introduction and hello everyone my name is Erica Sevilla a CEO and founder of in silica a startup company in the area of local diagnostic here in Silicon Valley California our company has been working with the any software John and the rest of their team members for approximately right now two years I mean I've been using their tools extensively almost all of their tools both in our R&D projects and also for the product development I guess that is why they have kindly invited me to this webinar to share my experience I show you some of the results that we have got with their tools specifically you know with the tools that are relevant to the PCR primer design like Aniplex and etc we have been using any software tools for design of many different panels however in this presentation I will focus on showing you the results from a relatively large you know panel that has detection genotyping and also quantification for respiratory infections first let's discuss quickly what we're trying to build that in celexa and why is it relevant to the tools that DNA software has been developing here I'm showing you one of our products is called the Hydra platform which is a sample to answer nucleic acid amplification testing system as shown over here a sample from a patient is taken put into a cartridge disposable cartridge and this cartridge is put into a random access instrument and in about 90 minutes you can get the results of the test and in this in the hydro platform we can do a very high multiplex nucleic acid amplification testing we can quantify many different sequences in genotype we can do all of them and the level of multiplexing can go up to 1024 and it's all enabled by the CMOS bio chip that is embedded in the cartridge which does not only the heating and cooling but also genotyping and so forth what is important is actually the method that we use in the cartridges and the method that we have is shown over here I'm going to spend a few minutes on it and you will see how does it relate to panel flex it's called the solid phase multiplex quantitative rt-pcr which also includes melt analysis in all-in-one reaction chamber so it's a close to or homogeneous assay if you look at the the left of the slide you will see what essentially we have in the in the cartridges you know when the total nucleic acid is extracted essentially what you have is a pool of DNA and RNA that you're trying to see if specific regions of interest are present to identify organisms and in many cases those reason your interest that you have might be you know location within the genome of the organisms that there are clinically relevant mutations that you wanted to detect but when we have this in the reaction chamber we also put in the reaction chamber oligonucleotides array with capture probes with unique sequences that we have and these are really you know basically embedded on a sensor array which can detect the capturing of the nucleic acid or the hybridization of the target with the nucleic acid in real time when the solution is present so if you look at the middle essentially what the system does is as follows when I have the pool of the more you know microbial nucleic acid and all of them in a sample we perform multiplex nucleic acid to amplify the region of interest which are over there so there are many forward and reverse primers over there you know up to fifty or hundred they might be you know concurrently present we amplify all those regions of interest we amplify them asymmetrically that means the concentration of the forward is much more than the reverse department to create you know single-stranded more single-stranded after the plateau of PCR so is an asymmetric PCR and after this is done in the reaction we basically the probes that you have on the surface can capture the single strands and we have 1,024 probe location that we can you know populate they can you know essentially probe the applicants any you know you can check if there's very specific sequences are present or not they not only do the capturing but also we we can do the solid phase melt that means that when the you know sort of targets are hybridized we can increase the temperature after that and see at what temperature at one point they detergent but it basically by doing this we can we can better identify differentiate between the sequences and using a look at poly form more polymorphism and other things over there the whole the whole system is close to you there's more reaction chamber there is no washing nothing is necessary and the paper which describe this is going to come pretty soon in nature by technologies except that you know the months are so really this one this will come out the panel that we intended to actually implement on Hydra it was one of them was this comprehensive our upper respiratory panel or u RP panel is a fairly large one it has 27 viruses three bacterial panels to controlled including human host control to check if the sample is actually taken correctly as you can see in this list or RNA viruses or DNA viruses or bacteria on all sorts of things Thanks one important thing is you know a lot of them are essentially detection you just have to know their presence or not but some of them like the flu antiviral resistance is basically genotyping a platform you know you have mutations across the different segments of the organism that you have to check and if they're present then you know that you know this strain for example is time or flu resistance and everything one important thing is you know beside the size of it you know we have to deal with a fairly large database that we down from GenBank and as you can see on the number of strains that we had that you know we had to detect dissatisfied inclusivity and so forth the numbers were you know tens of thousands you know twelve thousands but it's not the organisms like they're you know bacteria because they were less valiant they were smaller over that what we had to also include all of them in the design that we have although our panel or dual sorry are a biochip and the hydro system was highly multiplex you know dealing with this inclusivity and exclusivity and designing primers which are that general was it was a true channel without interfering with each other okay here is what actually we put together which we call it our assay development pipeline it has three components going from left first we have to set up you know in order to design the assay and the tasks are pretty straightforward first you have to know what exactly you wanted it to take from a genome point of view so that's a specification and requirements you know inclusivity exclusivity clearly you have to know all those things based on the strain you have to create a sequence database that is can be cumbersome you know there's a lot of repeated sequences there's a lot of issues with what you download from NCBI and all of them but essentially your setup is before actually use any of the tools like the final place you have to do but then you go actually to the you know bioinformatics part which in our case was automated primer design mainly done by panel plagues we had rigorous review that we did using the other tools of DNA software we did a probe design as well remember you know in our system we had surface probes or solid phase melt probes that we have to design - you know probe in your applicants and also we have when we get to the you know multiplex PCR we had to do prediction we have to debug in and all of them fall into the bioinformatic part of it the experiment and part of the the work as far as our pipeline work hand-in-hand with wire formatic board initially if you look at the right side of it initially we did the single phase primary qualification which is I get a lot of primer candidates and we have to screen them we rank them and down select them - the ones that perfectly we liked then after that we characterized number rigorously to the enth degree we had to check them for viability tolerance to variability of the sequence and the template we had to check the asymmetric PCR efficiency the yields that they had and also the PCR efficiency now also did the guard band testing basically changing the temperature of salt and all of those things and see if the you know sort of panel is actually robust that we use the VNA software tool to predict and then you know validated with the tools that we have and we did also the protesting and after you know the farmers were qualified we obviously did their multiplex testing and assembly and debugging them and after all of these things are done and all their requirements were satisfied that was essentially when we call the project with person now here is what we did and why our partnership with the any software mattered we enabled all of our by automatic pipeline by using the tools that DNA software app accept you know some of the other tools like you know the normal glass and things like that the automated private design was done by panel flex the reviews were mainly done by the thermal blast of de essentially on PE is it was the bread and butter of that part of it for the probe design we not only use the term of blast and off de but we also use another tool that I don't know if you guys are familiar with it's the only gold kinetic simulator because we have to look at the high res Asian kinetic and some of those things and when we were trying to you know debug and you know check the multiplex PCR you know sort of assembly we also use the term el personal de and there was one tool that you will go beyond all of these things and that was vision alarm is one of the my own favorite tools because to come to communicate between the scientists and you know between one formatic roof and the other ones I mean you just need a tool that anybody can use and they can have on there under desktop and everything and that was videoed all along so that that's how we essentially put everything together here the results or the highlights the first one highlights was whatever we got out of panel plagues if you put the right inputs in it or awesome primers if you get anything out of it that means that if your constraints were feasible you would get extremely good panels and we would get you know between four to five at least set of primers for each of the panels in the first run and when we tested them about 75% of them I actually checked this number before this presentation 75% of the primers generated by panel plagues meet the requirement in the first run which was extremely high that means they were working and then it was sometimes hard to decide which one do you want to keep which one you wanted you don't want to keep a knot and these are this is an example of the the results that we get out of the screening for example we did a dilution series and this is I think adenovirus B we call our ad v3 I think the acronym was that and this is a dilution series that we had the top a figure of it as you can see the green parts are for each dilution one to ten and the red is the NTC so the NDC and there's human background in the NTC as well so it's not just just your water that you have so it should you should look good and when we did their efficiency calculation and you could see that the efficiency was around forty nine nine five across that pretty linear and this was a typical results that we get out of these primers which are which was extremely extremely encouraging if you know where what you're what you're dealing with another important highlight was when we put everything together so obviously you know computationally initially we check that there's no private libraries and none of those things happen and that was the easy part of it when you generally put the multiplex together some of the panels you know they would not get annihilated but you know the efficiency would change a little bit some of them you know 10% 15% yields will go down and everything so you would lose you know the pristine single effects quality of it and then go to the level that you try to see okay what's going on are there any in the interaction or not the common interactions that people you know consider like primer dimer or Affleck on primary interactions and everything those are easily in the tools you can find out but there are another six of them that will come out and here for example you know two examples that we see so when we put the panels together we saw that the h1n1 panel but somehow in the presence of adenovirus each panel it would just a efficiency would go down and we look at it there was no angular and there was nothing that would happen and then when we went deeper and we look at with thermal blast and you know abdi even deeper we found out the situation like this in this situation this shows the sir that the duplex of the two primers within these two panel one of them the top one is the excess primer for ad noe which is 300 nano molar and the other one is a limiting primer h1n1 which is a first 50 nano molar remember this is a symmetric VCO the likelihood of this happening in the PCR temperatures is extremely low this is about I think point O one or some ridiculous low percentage of happening and the TM is about 28 degrees Delta G's minus 2.9 kilocalories per more but if you look at it carefully you will see that this is not an exponential growth but there is a chance that the limiting primer my extent and get essentially you know deactivated because then you know it cannot work with the right template and everything and we associated this with that reduction in the yield of h1n1 and we quickly fix it you know you know adding I don't remember what we did we added it you know intentional mutation or you know choose another primer that be huh in example 2 below that one of our you know PCR [Music] panels shuttle or yield and there was no reason for it was not exactly on the percent and everything and when we look at the structure and this is the entry virus d68 panel you would see that you know 23 percent of the limiting point verse where in this kind of you know hairpin structure and you can easily see that and you can increase the concentration through the level that you want and everything would actually go well so these things are actually really important when you wanna you know polish your your multiplex and these were all what we did with any software tools the third highlight which is essentially my conclusion based on this project or this partnership that we had was our development time was really fast and the main reason was the amount of redo that we had to do on the primer design was just did not exist it was just you know most of our problem was you know which primer to use and which one would play better so the iterations were very limited the number of iteration limited so we quickly finish the project and essentially there was no there was you know the number of full-time people that be needed on this one especially on the experimental part was extremely low so we did this whole project in about four months with three people which is which is unheard of so in conclusion what I want to say why why essentially we use DNA software tools and this is completely from a technical point of view maybe about you know ten years ago if you ask me that you know can you have a fully automated multiplex primer design pipeline I was saying yeah probably not but right now it's possible and after going through this project I will see the models from DNA software thermodynamic models are excellent they can actually create a real robust assets and you can do the guard ring you know sort of in a way that you know no matter what the panel will work the other thing is the the tools that the DNA software has I mean we use them every day you know you don't need to be for all of them you don't need to be a sophisticated part informatician I mean specifically visual lamp and all of them is just a must-have so they have been a right partner for us I don't know for others and their technical expertise is extremely good they have been really helpful you know in the design and this is not just under validation but also on making the message that we have much better so the result was you know we were worried about how much the the design of these panels would take time when we're actually pushing our products and you know a sort of limited budget in a start-up and the result was yeah we didn't we didn't need to worry about it that much it was just a straightforward it was a lot of work where he was straightforward and we didn't need to do you know I can mindless coding around primer three or deal with any of those three words that I tree that they're okay but there are they're they're not gonna get you into a robust system at thank you thanks very much charging that was a really great presentation to start off this webinar before we move on to the next presentation I want to remind the audience to submit questions for our Q&A session at the end of the talks we'll try to get to as many questions as we can so simply type your question into the box below the presentation screen and hit submit all right with that out of the way addy Thea floor is yours Jeff thank you for the introduction Joe and John thank you for having me join you for this webinar I'm excited to talk to you about what we've been up to at Karma code we're enabling mid pecs Diagnostics on existing TPC our instruments as you guys know PCR was a seminal discovery Kary Mullis came up with a concept while on a memorable road trip of the northern california and it's fundamentally changed the way we perform microbiology in fact PCR has conquered the world because it's fast cheap and ubiquitous in traditional qpcr instruments and assays you can measure four to six things per reaction one per color channel although the instrumentation is ubiquitous it is critically limited in the number of things that can measure at once at chroma code we are a bunch of biochemists engineers and machine learning experts and we've reimagined multiplexing in existing qpcr instruments with high definition pcr we can enable you to run a panel of tests in a single PCR reaction markers for oncology for infectious diseases for gene expression we've designed a set of multiplexed chemistry's that allow us to engineer pcr curves now with that same PCR instrument we can read out many targets in each detection channel and with our cloud software we can interpret these curves to map back to combinations of targets that were present in a reaction one of the aspects we engineer is the color intensity of each curve with HD PCR each target that's measured in a single color channel has a unique curved intensity for example if targets a is present it generates a curve with a brightness of 50 if target B is present it generates a curve with a brightness of a hundred and if target's see is present it generates a curve with the brightness of 200 however every combination of targets also has a unique brightness let's take another look at how this works consider a reaction where target a is present in this single color channel target A's presence generates a PCR curve with a brightness of 50 if instead target B were present in the reaction the curve would have an endpoint intensity of 100 target C similarly would have a signal intensity of 200 indeed each one of these targets if they were present alone would have a unique curve intensity but with HT PCR each combination of targets also has a unique curve intensity so for example if we have a reaction that has both targets a and B present the resulting curve would be the addition of the curve for target a alone and the curve for target be alone with HD PCR we can fit many targets into a single qpcr curve in fact it is the only method that allows us to map a set of targets to a unique curve signature we call these curve signatures levels and you can see here how the curves themselves fall cleanly into the level bands we've got a lot about chemical interactions in the PCR itself and we've engineered the key PCR chemistry and developed mathematics and signal processing to drive uniform performance within an instrument here you can see a number of cue PCR assays that have been run by different operators on different days on a same instrument this consistent intra instrument performance drives consistent inter instrument performance the same assay performs reproducibly and robustly across a number of key pcr instruments now I'd like to take you through a few examples using our respiratory viral panel platform demonstrator assay the graph that you see in the background is a screen shot from our cloud software in this particular reaction we had three targets that were present that were detected in a single color channel we had 10,000 copies of RSV a 10,000 copies of human Rhino virus and 10,000 copies of RSV be had the reaction only had RSV a the qpcr curve would have terminated in the band that you see highlighted in black if instead there had only been human Rhino virus the curve would have terminated in the band that you see highlighted here in blue similarly if RSV B had only been present you would have had a curve that would have terminated and the band you see highlighted in red but because all three targets were present all three sets of primers and probes were amplified in this single color channel resulting in a curve that uniquely terminated in the band highlighted in green HD pcr works the same way with clinical samples in this particular sample influenza A HD was present in our assay we have primer and probe pairs for both the pan influenza a marker and the h3 subtype because both marker sets were amplified the resulting curve from this clinical sample terminated in the band highlighted in green HD PCR is engineered so that curves terminate and deterministic bands regardless of the concentration of the targets that are present in the reaction this is Croma code HD PCR it's a controllable chemistry engineered for robust performance across targets and concentrations it's mathematics and algorithms to interpret curve signatures and deliver consistent results on multiple platforms all of this enables mid flexing on qPCR instruments but wait there's one more thing we can not only enhance qPCR instrumentation but we can also enhance digital PCR instrumentation HD PCR lets you take a traditional two color digital PCR instrument platform and measure 10 to 20 targets per reaction this is only possible because we're able to model the complex interactions that take place in multiplex PCR and that's where DNA software comes in John and his team have done a lot of hard work to build a tool visual iymp that's very insightful and helpful an hour of multiplexed as a design in fact we're currently evaluating their panel Plex which is the next evolution in in silico design software we're looking forward to working with DNA software and the upcoming months to summarize we're chroma code we're bringing mid-flex panels to a PCR instrument near you thank you for your time today and and Joe thank you for hosting me for this webinar please feel free to reach out to us if you have any questions thanks Aditya that was a very detailed presentation before we move on to our final presentation I want to Ryan the audience once again to submit questions for a Q&A session after the last talk we'll try to get to as many questions as we can so simply type your question into the Q&A box below the presentation screen and hit submit all right let's move on to our final presentation John the audience is listening all right so I mentioned that I will be your presenter for the third part of this webinar Johnson launched eeeh the founder and CEO of DNA software the DNA software was founded to commercialize research started my lab and to extend upon that research and go further to deliver software for automated designs we've been in business for more than 17 years and we've helped numerous in organizations with their molecular diagnostics we pride ourselves in particular with solving some of the most challenging problems in oligonucleotide design and we also have used the knowledge of doing previous projects to distill it down and to software that is very easy to use all right so today I'll be telling you about the challenges of multiplex PCR and our unique approach to that problem all right so let's first talk about this you know PCR is a wide widely used method and has been engineered in numerous parts of its components but PCR one aspect of PCR that has been neglected is primer design if you think about instrumentation like the instrumentation offered by in celexa for example many companies have worked very hard to miniaturize PCR to introduce micro fluidics to improve the sample preparation techniques with kits and robotics micro fluidics they've done a lot with improving the master mixes go in there the buffers and additives and they've done a lot with the data analysis but one part of PCR that has been largely empirical has been the design of primers and this is despite the fact that PCR if you think about it is a relatively simple reaction you have your genomic DNA you may have some contamination in there but you really only have a polymerase and some salts and that is a something that should be subject to scientific investigation and we've done exactly that at DNA software to try to understand the causes of failures of PCR now many many researchers not knowing how to design PCR will put forth the effort to combine various free software into a PCR design pipeline which is shown here and we'll be showing you each one of the stages that typical design pipeline like this has have difficulties or problems with way they're set up one of the first problems is this funnel that I've shown here if you look at the funnel the idea here is that computation is is limited and so we should take a wide number of primer candidates and then as quickly as possible filter those down to a few candidates because the computational methods get more and more expensive that very philosophy of computation being scarce is something that leads to leads to problems with not finding the proper primers so for example some of the problems with multiple sequence alignment which is often used to find conserved regions of a genome is that the sequence is what we really care about for PCR is not sequence identity conservation but we care about how well will have primer bind to different members of a group of sequences and multiple sequence alignment really does not give that another aspect that folks will do they'll use a free software like primer 3 which does indeed do you do a wonderful job predicting two-state melting temperatures and predicting primer dimers even but it is neglecting the competing folding so they might use a separate program like m fold to predict the secondary structure of DNA but they don't combine those two to understand the competition and lastly they may use a program like blast to check the specificity of their primers and check to see if their primers for example bind to anything in the human genome but we'll see again that blast is really not the correct tool to use for that problem and lastly this design pipeline is really kind of designed around a single target at a time and multiplex PCR is about quite different than that you have many targets in the same reaction and you need to account for the system effects alright so because the current approaches for designing PCR don't work that well researchers are left with the empirical optimization approach and typically the idea here would be to start by optimizing individual single plexes and getting those to work and certainly if a PCR doesn't work in a single Plex it's not going to work or get better when you make it into a multiplex but then the idea is to try to combine those single plexes into larger and larger groups and when a single Plex doesn't work or interferes with one of the other members of the of the multiplex to fix that single Plex however the process of fixing that single Plexus causes other reactions to fail and it's not known why those other reactions to fail and so you typically takes a PhD level scientists three to six months to get a 10 Plex pcr to work if at all this results in a sort of you know whack-a-mole situation where you might have a seven Plex working you go to make it an eight Plex and boom a different primer doesn't work and then you have to try to react demising go to a scratch often this is sort of a whack-a-mole situation where different problems keep popping up and talking about the multiplex problem we the problem here is to try to get sets of primers that are mutually compatible we're compatible means that the primer sets amplify with similar efficiency don't cross hybridize and don't form false amplicons let's take for example of xxx plex much like the size of the reaction we did for in celexa so if you were to try to keep let's say in a 30 Plex you keep 10 primer pairs candidates for each of the targets each of the 30 targets so to start with you've got 300 forward primer candidates and 300 reverse primer candidates plus probes alright and you might ask well how many possible combinations are there well for each panel there's ten combinations of primers so as you add panels it multiplies ten times ten times ten in the case of a thirty Plex it'd be 10 to the 30th power possible multiplex solutions a number that is 10 million times larger than the Avogadro's number and as you go to even larger plexes you know consider like a hundred Plex 10 to the 100 power would be a truly astronomical number larger than a number of particles in the universe so we have a problem here we have an exponential explosion and each trial multiplex combination requires an extensive amount of computation for each one to check the specificity for example and also to check for cross reactivity and false amplicons and competing folding so it's such a huge space of possible possible multiplexes there's a need for a different approach so our goal at DNA software is to solve the multiplex PCR design problem all outright and our approach to solving the PCR problem design problem is a three-pronged we look at the science of the problem we investigate the causes of failures of BCR and implement those understandings into our software further we develop new algorithms for predicting the secondary structure that competes with primer hybridization for calculating off-target mis hybridization reactions that lead to false positives and false negatives as we'll see and also a comment for this multiplexing effect I talked about in the last slide and lastly we bring to bear massive cloud computing resources so that we can address the full challenge of the problem including modern large sequence databases all right so I'm gonna be giving you a brief overview of the four most common problems in multiplex design this is what I'm giving you today is an abbreviated version of a larger discussion that we've presented in two free white papers and a webinar that you can view at a different time they're available on our web site at DNA software or on YouTube and but I'll give a brief introduction to those today just to sort of set the stage so multiplex panel design challenges include the following for the first challenge is that targets need to be amplified at the same rate but often in real multiplex PCR reactions one of the amplicons can take over the reaction resulting in amplification at uneven rates and we'll talk about what the causes of that are another artifact can occur that can occur in multiplex PCR is due to cross hybridization where a primer from one target binds to the amplicon of a different target or you can get background amplification where primers amplify some contaminating DNA that's present in the reaction and we'll see that both of these problems are not addressed with using a blast search in addition there is a need in a 21st century to include large sequence databases as a part of your design and in particular design of a inclusivity exclusivity and a background list of genomes and lastly is the issue of getting everything to work well together in a multiplex PCR taking into account the fact that multiplex PCR is a complex system with many interacting variables alright so one of the first effects that we would like to discuss today is the issue of target secondary structure on the left hand side is the model that is most widely used for predicting the thermodynamics the melting temperature and Delta G for primer hybridization and in this model two random coil strands a and B come together to form an a/b complex the temperature at which you have half random coil and half duplex is the melting temperature and the amount of energy free energy released in this process is the Delta G of the reaction now that two-state model works well if your actual reaction really does involve just these two states random coil and duplex however the problem is that real DNA's are not like this real DNA's are folded molecules as shown on the right hand side so your target DNA is a molecule that is often highly folded and if the folding region shown in green here if their region where the primer is going to bind is folded then before the primer can bind that region must become unfolded and that reaction can be slow kinetically slow and it also can inhibit the equilibrium of the reaction so in addition to the target being folded the primer or probe itself can be folded and again before the primer can bind to the target the primer is to unfold and that is also a slow reaction alright so at DNA software we take this into account the competition between folding of single strands hybridization of the two strands and what we call the multi-state coupled equilibrium model all right so the multi-state coupled equilibrium model is one of the one of the strengths of our software and that is important for accounting for false negatives that occur in reactions all right this next slide is showing you about some of the causes of false amplification in PCR that can consume primers and cause the PCR reaction for some of the targets to shut down the first one is widely known the first mechanism involves primer dimerization in which the three prime end of each primer hybridizes to one another and such a hybridization event would lead a polymerase to extend that primer and thereby consume that primer and cause that particular reaction to get shut down that involves that primer this is a mechanism that is widely understood and accounted for in most software design programs less often accounted for is primer cross reactivity so the issue here is that in a multiplex PCR you have many primers and many amplicons all in the same reaction and if the amplicon from one reaction say that you have an ample con from the Zika virus shown on top here and below here is a primer from the influenza virus say you're doing a 2 Plex that involves those two viruses if the influenza primer binds to the zika amplicon that cross reaction will consume the influenza primer and create a shorter amplicon thereby spoiling both reactions for both the influenza and Zika detection there's also a series of other possible amplification reactions that can occur that I will leave for the users to look at our previous webinar and lastly there's the concept step II hear about off target background amplification so if we were trying to make a diagnostic for viruses or bacteria would be concerned with the human genome as a possible background ne we wouldn't want those primers to amplify somewhere in the human genome that would deplete the primers causing also the multiplex PCR to not amplify uniformly and cause problems I'll return to this in one moment let's take a brief aside and talk about the problem of using blast to predict primer specificity of course a blast search is one of the most widely used methods in bioinformatics it's meant to determine sequence similarity and to infer evolutionary ancestry but it's really not appropriate for did for the use of predicting primer specificity so let's talk about why that might be the case so what's wrong with using blast to predict cross hybridization well the problem is that blast search is based on sequence similarity not complementarity so as a result users will perform a work around and that workaround is to take the complement of their oligonucleotide and then use blast to search for hits to the complement of they're all ago and that work around immediately produces problems the first problem is that it gives the wrong ranking of hits so blast is searching or ranking the hits based on the amount of similarity we'll see why that doesn't give their correct ranking in a moment a blast search misses about 80% of the thermodynamically stable hits because it's using the wrong scoring it also gives you many irrelevant hits you know a blast of your primers against the NT database a nucleotide database gives you many irrelevant hits that you don't care about also blast does not distinguish you against about hits that are extensible by a polymerase versus hits that are not extensible by a polymerase and not as problematic therefore it also doesn't attack detect amplicons that formed by pairwise combinations of primers alright let's look at that why ranked blasts gives the incorrect rank of hits and the fundamental problem here is that sequence similarity is not the same thing as thermodynamic stability of a complementary sequence so for example not all miss all matches are equal alright if you consider a GC base pair is not equal in stability to an 80 base pair so blast on the other hand because you just take the complement of your sequence and score all matches the same effectively it is scoring all G matches the same as all a matches which is equivalent to saying that a GC pair is the same as an 80 pair which is not correct blast also doesn't understand that different mismatches are unequal instability it instead is looking at an evolutionary model where a mismatch is an indication of a mutation that's occurred very different than what we're looking at with primer design where we're looking for thermodynamics of complementarity and so blast scores all mismatches to be equally penalized for example a GT mismatch it would score the same as a CC mismatch when in fact the thermodynamics of GT is a very stabilizing mismatch and a CC mismatch is the most destabilizing mismatch in thermodynamics of complementarity it also doesn't account for some several other effects such as the fact that bulges in blast are scored based on insertion and deletion frequency well that has nothing to do with hybridization bulges are thermodynamically very unfavorable for very different reasons then then the reasons scored by blast it also doesn't encode dangling ends blast doesn't include the position dependence of mismatches so mismatches at the end of a helix do not contribute the same as mismatches in the middle of the helix it all doesn't account for temperature and salt effects and importantly it doesn't account for whether it's extensible or not alright so now let's talk about an important problem for modern primer design which is the idea that we can use collections of genomes in our design so for example if we were going to design a diagnostic for the Zika virus we would want to make sure that that diagnostic detected all known variants of the Zika virus and in this case we did a search of GenBank and found 168 complete Zika genomes and certainly we would like any diagnostic we designed to detect all of those we call that the inclusivity list we'd also like to make sure that our diagnostic does not amplify genomes of near neighbors and give a false positive so in this case near neighbors we would put them into a list called the exclusivity list and this would be a collection of viruses in this case that are either phylogenetically related to the Zika virus or are so-called near neighbors or present similar symptoms in the clinic and thence and hence likely to be in a sample that would give a false positive and then lastly is a background list where we would want to include any genome of organisms that could cause a false positive so for example if we were doing a diagnostic for the Zika virus well the Zika virus would be tested from a sample derived from a human being and we were to make sure that the human genome would not present a background amplification that would deplete the primers and cause false negatives potentially all right well let's talk about an application of this now so a common artifact observed in PCR is shown on this slide this is a dilution series for a target that's been diluted tenfold at each one of these different concentrations and this is the way PCR should look it should be that if your target is in high concentration you get a high amplitude of fluorescence signal and it as you lower the concentration the amplitude of the PCR signal should and stay the same but in fact what is often observed is a set of curves like that shown on the right hand side here where we see a high fluorescence intensity for high concentrations of target and then as you get lower and lower concentrations of target you see that the fluorescence intensity gets decreased and this should not be happening and many users are not aware of what the causes of this so-called step-down effect is well it's due to the off target hybridization to the background DNA that we talked about in the previous slides so if your primers are let's say your Zika primers in the previous example if those primers are binding to human background genomic DNA well those primers will get depleted and if the primers are depleted then the amplitude of the PCR signal is going to decrease as shown here and this becomes more of a problem depending on which isn't a higher concentration the target DNA of interest Zika virus or the background DNA and as you get to lower and lower concentrations of the target DNA then that background DNA consumes more of the primers and is able to compete in a reaction so in a single Plex PCR this step-down effect is is somewhat of just a nuisance but in a multiplex PCR this effect is one of the major causes of failure of the multiplex PCR it results in some of the amplicons being amplified in an unequal rate and that in turn leads to false negatives in a multiplex PCR alright lastly is the problem of getting everything in the reaction to play well together as I mentioned earlier multiplex PCR is a multi-dimensional landscape that is has a huge number of possibilities of combinations of multiplex interactions and the iterative empirical approach is suboptimal or fails the leads to the multi whack Amole I discussed earlier even an iterative computation approach will not work even though computation it can drive billions of possibilities billions are nowhere near the 10 to the 30th power combinations that we saw so multiplex is a complex system and it requires a 21st century approach and we have implemented such a 21st century approach into our algorithm so we call it the multi picking algorithm which tests all permutations of single Plex candidates and it uses a brilliant depth-first search with pruning algorithm to solve that combinatorial explosion in a tractable fashion and it guarantees that we find the top-end best solutions multiplex solutions out of the whole space of 10 to the 30th this is something that is completely different approach than the empirical approach all right so we've encapsulated all of the concepts that I've talked about today into our multiplex software called panel Plex and we like to think that this makes multiplex design made easy this takes away the whack-a-mole and it's encapsulated into our products the panel Plex MDX and panel Plex NGS and for more information you can look at the information at the bottom here for contacting us at DNA software thank you for coming to our webinar thanks John that was a really great presentation to finish this webinar and you provided the audience with some really great insight into primer design as well as showed us the power of panel Plex so we thank you for that before we start our Q&A session I want mine everyone once again this is their final chance to submit questions for speakers so hurry up and alright looks like we've got some really great questions have come in already let's begin the Q&A and try to get to as many questions as we can give us a few moments on our side to get all the logistics squared away and we'll begin the Q&A presently all right guys so let's begin the Q&A session um guys the first question actually is going to be I think one for both of you you might want to weigh in on this one of our audience members would like to know how many targets are there in the multiplex how let either one of you answer first whoever decides is the best person to answer John do you want to go first oh sure all right so the size of multiplex that we can handle depends on which product you're using of ours so for molecular diagnostics we did for our Jiang's company in celexa we did a 32 Plex but we've done much larger plexus in other contexts so for example for cancer panels for targets in the human genome we've been doing plexes above 50 Plex 100 Plex in that neighborhood and we're current working working on a panel Plex to extend its capability to above 200 Plex so for most applications for like next-generation sequencing and for molecular diagnostics to flex levels that we're hitting are the ones that most of our customers are telling us that they need our Jean yes so from a from our platform point of view generally what limits the level of multiplexing is not the computation and the primary design part of it is how much you product you need to create in local diagnostic assays generally similar to what we have in order to do the detection you need you need minimum level of you know amplicon generated in a if you want to drive the high relation and detect them on the array you need about 10 animal of purflex so if you run the numbers where the capacity of PCR and how much product you can create you should create we can create after 40 50 cycles 40 cycles generally that ends up being more than 50 plagues or in almost a 70 plagues would not give you that much targets that's a limitation of PCR you go for the larger panels as the John mentioned prion for sequencing and everything obviously this number can go extremely large so it depends really on what you want to do and if you want to do a quantitative analysis or you're using it as a you know preamp you need just tenfold increase from the background to do the sequencing those numbers can be then extremely lucky all right great thanks guys our jegging stick with me we got another question for you one of our December's would like to know what other options did in selects that consider besides DNA software that's a good question I think we spend about you know six to nine months and we were looking around for how we're going to do these laws and multiplex panels and we were we were heavily engineering and by formatic bias so you know at least we thought we knew what we were dealing with we look on all the kinds of wrappers and all the things as you you can do around primer 3 and all of them we were aware of the models that John have is you know as usual off on everything but at the end of the day we looked at it and check the algorithm of John and game be any softer a lot of pain and suffering and finally we decided to go that I think just that decision was the right one the results were extremely good I mean and the rigor was there so I'm right thank your Jane John that a question for you what is the limit of the Plex size panel Plex can design ahh that's related to the question we just had basically you know the NGS version of the software right now can handle up to 100 Plex and we're working to make it go higher than that the version 4 MDX actually we usually work with summers when we're combining panels like we did for in celexa we do that in a concierge mode where we will do that in-house to handle the larger panels and oftentimes customers have very personalized needs in terms of how those are set up so they require more hand-holding and more of a sort of consulting mode but software itself can handle you know quite large flexes as I said all right thank John our Jane is a question for you what did you mean when you said that Dini software is the right partner for us do you mean to imply that it's not right for everyone yeah you know this is sort of my assertion I would say yes is the reason is you have to have you know a level of sophistication and you want to have a level of sophistication in your primary design and also have the team and the team that is receptive to it because at points it's very much you know complex and competently intensive so under consumers yeah you know you know anybody can do it you know you can work with DNA software and they will they will design and give you the ass and probably they know more that what what you even think but if you want to actually in-house for the constant message develop it you just have to have the right bioinformatics tools and expertise and be receptive to it sometimes people who are experimental you know they'd rather change a base over here and there to you know optimize their multiplex and you know there's some guidelines you know that they have and it just can become really hairy if you if you do those things and you're not doing it systematically because you can just get out of half the multiplex and get out of half again I just tweak it manually if you will so that's that's what I make that was right for us you know I'll follow up on that you know the multiplexing empirical approach even with the sort of limited bioinformatics approaches that many companies will take are often lead to that whack-a-mole situation I discussed in my that's my part of the webinar where folks get into this loop of design and redesign and reoptimize ation and without knowing what's going on and not really taking an account of system effects so that's really the essence of what we've spent a lot of time figuring out all of those causes that make you know multiplexing not work we're in like primers cross hybridize or primers are forming false amplicons with the background DNA those issues are not something that typical users have the expertise on hand to do make the proper software to be able to do that ahead of time so unfortunately many companies that end up coming to us are ones that have spent a lot of time you know more than a year trying to get a multiplex to work after great pain then they'll come to us and realize that we can help them tremendously and go much much faster in just a tell a story we had one company come to us with a sexually transmitted disease panel that they had been working on for well over a year they had two different groups of scientists work on that and they both sets of scientists were unsuccessful they came to us with about a month left in the project and said hey can you make this 30 Plex multiplex STD panel for us and we were successful and got it to work on the first try so those kinds of stories happen a lot at DNA software and a lot of users don't have the awareness that they need in order to realize that they need help in that DNA software is already invested 15 years of research and development more than 15 million dollars making the software that we have the panel Plex software so that they don't have to reproduce you know something that's already been discovered namely I just wanted John to the what you mentioned you know I think is left in presentation and that's you also have to know your enzyme and you know your master mix you know some of the myths that people have that one of them is better than the other one or you know I like this kind of enzyme mix because you know I have been using it laughs for a long time that is also something that you need to be careful about it I mean you know if you have C prime X so in the glaze or not and these have a tremendous effect on the multiplex when you get over there and I think that's that's something that also when you look at the infraction of the primers and everything you have to be careful about and sometimes people forget about it and see you know they just stick with one that they have which you have to be careful right yeah very often we we need to educate our customers about you know things best practices in the field and and that can actually be one of the most valuable services we provide sometimes is getting people on the right track in terms of you know best practices in terms of enzyme they use a thermal cycling process they use you know how to properly get the databases for inclusivity exclusivity and background those sorts of things are not a sort of different way of thinking about the problem and oftentimes we bring that to the table of really helping people solve their problems all right thank you guys very good discussion John we have another question for you since you're on the line um one of our arts members like to know how long does it take to run a design and panel plex depends how big your Plex is if you are and it depends how big your inclusivity and exclusivity and background panels are so a typical like let's say you had like I said mention the Zika virus during my talk with 168 targets and then it had the background had 5,000 different viruses in the exclusivity and then we have the human genome in the background so that um panel took around 8 hours to run using our 33 CPUs dedicated 36 CPUs dedicated to their job so we're doing a massive massive calculation you know these are not something you could run on a laptop computer so that's one of the advantages of using our our cloud-based computing so each panel you know it should you know so it depends on the size that was a sort of large panel there sometimes they don't scale as much as you might think so some of the larger panels in our Jiang's just talked that he mentioned had you know 7,000 to 10,000 sequences in there inclusivity and exclusivity those we munch right through those some of the rate limiting steps are more like how long is the target you're having us analyze and how many you know like so for example a virus target is a lot faster to run than a bacterial target for example so it can be complicated answer to that question the the larger plexes that we do for next-gen sequencing those panels like a 50 Plex something like that 50 different targets in the human genome making them multiplex against that runs at about 20 hours on our system right now so that's just sort of to benchmark times alright great thank you John arguing good question for you what's your estimation of efficiency of multiplex strategy for paper-based amplification or lab-on-a-chip operation I cannot comment on the paper race I don't know exactly what the question is referring to well not in the lab on a chip based you know because I've at least what we have experience with generally the transfer that we did from a benchtop instrument to the lab on the chip was was pretty successful the only thing that we had to be very careful about it was obviously any kind of inhibition that happens in a you know sort of the lab in the chip platform one does generally materials on surface or surface to volume ratio but another thing which was very important was the temperature profile that you have insult the assays depending on what kind of structure your applicants have is very important on how the direct effect on the efficiency so meaning that you know your ramp rates and exact temperatures there should be for the denaturing and all of those things so if you keep it actually the same as what the instrument is your vegetable instrument is which by the way it eventually instruments do not have it perfect you know sort of uniformity and every you know by far but you know just if you keep it within this region I mean you know within those rooms I think the transfer and the efficiency would be you know similar to what you have in benchtop you know 10% if anything that's not necessarily what we all deserve alrighty thank you RJ folks with that we've come to the end of our webinar so let's remind everyone that the webinar will be archived on our site for about six months at WGN news comm so if you missed any parts of it you can watch it again or if you know somebody else that might be interested in it you can feel free to forward link to your friends and colleagues I like to thank our Jiang Aditya and John for their informative presentations and I like to thank you the audience for your attention and your thoughtful questions and a very special thanks to DNA software for sponsoring this webinar so hopefully we'll see you again at another gen webinar in the near future good bye for now