 Well, welcome everybody. Thanks for coming back for the second session of our information gathering session on industry advances in mass spectrometry and separations technology. I'm Trisha to whole skate on one of the associate program officers on this project and then Marcus halfner one of our consensus study committee members will be moderating today. But before that, we will go through just a few housekeeping items. So just an overview of this committee activities that this is a meeting that is an information gathering activity for the consensus study on sequence sequencing and mapping of RNA modifications. And the study and the meeting are sponsored by the Warren Albert Foundation and the National Institutes of Health. What's shown here is our consensus study committee and they helped us to plan this information gathering session. And you can also see our national Academy staff team here. And then, just a few reminders, you can make comments and ask questions by using the chat or raising your hand in zoom. If you have any technical questions or issues you can contact Nam and his email shown here but you can also private message him and zoom. My ideas made during this meeting should be attributed to individual speakers and not their organizations unless otherwise stated. Thoughts shared during the meeting should not be interpreted as the opinion of the National Academies of Sciences, Engineering, and Medicine, or the committee conducting the study. The meeting will be recorded for use by the committee and harassment and bullying will not be tolerated. So, if everyone could just be respectful of all fellow participants and speakers that would be great. And now I'm going to pass it over to our moderator Marcus Haffner. Hello everybody. So my name is Marcus Haffner. I'm an investigator at the National Institutes of Health and a member of this committee. That is tasked to develop a roadmap for achieving the direct sequencing of RNA so basically getting a comprehensive overview of everything that is an RNA structure modifications, etc. In the course of our information gathering will be looking at all sorts of aspects related to developing their final goal. And for the purpose of today, these are pretty much gaining insights in the scientific needs and particularly the methodologies and their limitations as related to the sequencing of RNA, and of course will be excited to hear anything about potential new technologies. So, our meeting goals are to understand the current capabilities and limitations of RNA separation and mass spectrometry. And we hope to hear something about ongoing or planned research and development activities by our company speakers. So, we have three speakers today. So, Martin Gillar from Waters, Mike Grieg from Brooker, and Ben Dunstead from Agilent Technologies. And so, we hope that the talks that we're hearing will be something like 15 minutes with five to 10 minutes of discussion, 10 minutes maximum. And with that, I would let's start and welcome Martin Gillar, who is a scientific fellow working at separations research and development at Waters. Welcome. I think I was muted. Yes, thank you very much for introduction. Time is short, I will get right into it. Next slide please. I'm going to mention a few separation methods, most importantly, unpaired reverse phase chromatography and some other potentials, and then talk about LCMS and some aspects of it that I see that we have solved successfully in the industry and some parts that are still at works. So let's move on. Unpaired reverse phase chromatography, in my opinion, it's, it's the best method currently available for LC and MS of oligonucleotides. And when I say so, I mean the oligonucleotide based drug. So, mostly those 20, 30 male oligonucleotides. Here I show the example 21 male beautifully resolved from the failed sequences, each of those failed sequences and small ones assigned by mass spec mass, and we can sort of deduce what the sequence is. If there are some peaks in between that we don't know what they are we can use LC MS MS. And because this is sort of sample unlimited application in an acrylic acid drugs, you can load a little bit more and to get the sensitivity where we need it. So so far so good. Next slide please. We are of course now much longer oligonucleotides that we deal with SG RNA upper left corner. It's like typically 100 Mary so let's have a look at these unfair reverse phase method. They can deliver the separation baseline separation, let's say up to 6070 Mary's and then it gets the selectivity diminishes. It's good. But it's about as far as we can go I am about to publish some paper with 150 Mary's boy a tail and RNA analysis. And yes we can get some resolution there but that's really really about as far as I think we can go. The next slide. And if we are talking about longer pieces of RNA, including some RNA size exclusion chromatography can do something very interesting here. Yes, it is not the high resolution technique. It doesn't give us an N minus one separations but it gives base on the poor size of those size exclusion columns. There is a new method that even have some potential for LC MS, but it's a method to resolve big pieces from short pieces, and perhaps could be used as a one of the sample prep techniques. Next slide. I think that when we deal with long polygonal clotides and long nuclear gas it's the LC no longer we have a resolution of an N minus ones. That's natural. Also the mass spec starts to lose its ability to analyze it with a decent sensitivity. I will later on show some examples why it is so difficult, but the solution to this, if we cannot do it in one piece of something that has a half million dolphins or one million Dalton. We can then do the regular trick we chop this large pieces of RNA into small and more manageable oligonal clotides. And now we can do this so called M RNA mapping it has been published in literature. It is used for SG RNA characterization maybe the RNA characterization, and mostly for M RNA. What I show here on the upper right corner, the red mass spectrum at the mass spectrum is the decolonial spectrum of the poly A tail. So we can do these, like 100 127 merlongs. You can see that there is a micro heterogeneity. All of these peaks is one of the species and it, the delta mass in between those peaks is the mass of one mononucleotide a so we can do those applications. We cannot get the mass directly for the intact M RNA, especially if it if it has the micro heterogeneity. The next slide. You can appreciate that this digestion. Here I show you one example of Helik hydrophilic interaction chromatography. It's an alternative it's LC MS compatible method. It is now being more widely used it's a new thing so everybody's excited. It perhaps doesn't have a as high resolution as I'm very very space chromatography but it has a orthogonal selectivity in a sense. If, and this is what I show here. If you just look at the shore all you're going to close this blue chromatogram 20 Mer, you can chop it by RNA Steve one digest onto smaller pieces. You could use it as a characterization method. You can see that the separation order is 49 and seven Mer, and that is weird, and that is because that nine mer doesn't contain phosphate group attached to the three prime and hence it has a different overall hydrophilicity. And that is the method that is really sensitive to the hydrophilic or hydrophilic groups attached to the oligonucleotide and that altered the selectivity and it really is kind of neat separation. The next slide. I know the markers. You mentioned that you are also interested in discovering the modification so when I say we are in a decent shape with LC and LC MS or oligonucleotides, and that we can do MS MS sequencing that is the confirmatory sequence we know the sequence up front. It's good if we have completely unknown sequence that is an order of magnitude more complex task, and you would need really a good software if, however, you know the roughly the sequence and you're looking just modifications for modifications in that sequence. It's somewhat in complexity lies it somewhere in between. Nevertheless, it gets tough and if you don't know where what is one of the method here shown is that we can take RNA we can digest it to nucleotide or nucleosides. And then just look at those small molecules for main constituents of RNA, plus all of its potential methylated and other modifications. And it works. It's published, but you lose the sequence information you don't know where the modification happens. You also need as the variable standards to do the quantitation so some limitations but this is definitely one of the good metal. Next slide. Maybe when I talk about LC MS and MS MS. And now I start to talk about longer oligonucleotide something 20 mirrors are easy to sequence completely for the mirror. It's not easy and you may not a complete, you may not get a complete sequence read through the sequence. If it's longer than we are in trouble. Definitely this is a research subject. If you go well beyond those scales of lens and I show here at the papers published recently the SG RNA work. What typically we observe for long. On the right side is the adaptation sodium potassium edaks are very significant much more significant than for shorter oligonucleotides. That's the left panel. The sample has to be purified otherwise you don't get a really useful spectrum on the right side. It's not as striking. The edaks are everywhere every charge states has edaks and the longer look like as it is the more abundant adaptation and difficult to handle this. It is also now when you lose the sensitivity the complexity is there. There are multiple charge states as I show here that could be easily for a 50 charge states it splits the signal you lose the sensitivity. And last, the software we would need additional software to do interpretations of these long LC MS and MS MS analysis. Next slide. I know I'm going fast through this, but I want to get to the this point for example so one of the improvements that we did at waters was that we recognize. Could you please click again. We recognize that. And customers have seen this, no click assets are very negatively charged molecules, and it has been argued that there is some sample loss. If you check first sample on a virgin column, that's the right chromatogram you may lose the signal, it gets better upon injection number two, and then the signal improves. This is confusing country to it if, and it has been argued that this is the non specific absorption to the Sorbent of chromatographic Sorbent on the column. Because it's mostly the non specific absorption to the metal surfaces, that means hardware LC hardware, needle, electrospray needle even a mass pack. Fritz are the biggest corporate. And we also know it gets much much worse in in acidic mobile phases, because now the metal oxide surfaces are now charged positively that they eagerly absorb these nucleic acid so it could be partially corrected with column preservation, special mobile phases at high pH but next slide, the solution that we came up with is, we really take our columns hardware, and we modify it chemically so that all the metal surfaces, especially in the fritz. And even the LC system itself could be more or is modified with the hybrid silica technology. So now there is a new possibility for sample interaction with metal oxide surfaces and the next slide. It has the effect that you see here. In the left panel already, there is some loss, when we inject the sample and that that doesn't happen when we have these modified columns. And again, this is the same column back with the same Sorbent except the metal hardware has been chemically modified so upon injection one, you have a full recovery of the sample. Is definitely something next slide that will enable us now to do high sensitivity analysis. When I mentioned on specific absorption. If you have hundreds of people malls or none of malls of the sample. This effect is negligible. Once you are going after tens of phantom walls. This non specific absorption effect is enormous. And it needs to be dealt with. So, the last part, and just a large snippet of information I mentioned that we need software for LC MS and MS MS interpretation, and even when we do the mRNA analysis mRNA mapping. We need a software that basically predicts what we expected masses should be in a sample. There are tens of hundreds of peaks manual interpretation is increasingly difficult so all the manufacturers now have developed some kind of software we have a software is called intact mass. And here I show example for just a simple synthetic oligonucleotides 21 mer. It's heavily modified in many of those mononucleotides they have a certain modifications that could be also oxidation, the amination, there could be truncation and so on. The interpretation manual interpretation is quite tedious. Now with the software, it actually will assign all these LC MS peaks with the expected mass. And if there is a match it will call it and it will tell you yes you are truncated yes you are deaminated. Yes you appear to be the losing software and it's oxidated form of phosphorotioate and so on. In the case, and there will be cases where simple MS will not be enough to make the call. We need to go to a next next level and that's the confirm sequence software that will perform LC MS MS and based on the fragmentation will make the interpretation of the sequence. I am talking about confirmation sequence. It's not a denouvel sequencing. It's a confirmatory sequencing. So the next slide I think is my last one. In summary, as an industry, we think we are in a decent shape with LC separation methods. Even for long oligonucleotides now we are developing the MS is sensitive enough for many applications. The software is getting there but the caveat here. We mostly focus on this nucleic acid based therapeutics and not necessarily into biological samples extraction of small amounts of mRNAs from biological matrices and these more complex applications that Marcus, you may have in mind. That's that would be my conclusion off of this talk. Thank you very much, and we can open the discussion. Thanks a lot. Martin. So, yeah, let's start the discussion. Questions from the committee. Oh, Brenda use. Thank you, Martin that was great and so this is not my expertise but I understood that when you were completely digesting the, the oligonucleotide you could get good quantitative data. I'm wondering about quantitation in the when you are just looking at fragments or longer pieces. And I, my particular question would be something like okay if you have a tRNA, and you think it's tRNA fee, but actually at some positions. It's modified, not at all and some, it is modified a lot. I'm wondering from that sample, can you tell how much heterogene, heterogeneity you have in that sample. So the tRNAs obviously heavily modified. There could be various versions to micro heterogeneity. We could do some size exclusion first capture and isolate the tRNA range from all other different sizes. We can send it to a second dimension chromatography LC MS where you can get partial or actually quite a nice separation of all the micro heterogeneity. And the LC MS will give you additional separation dimension so you will get mass. So the first simplest application you will get an mass. This is good enough to infer what modifications are on your tRNA. If you know roughly the sequence that it is, if you don't know what the sequence, hence the mass, it should be, you will end up with list of masses in relative abundance that's that's about as much as you can get. So if you really need to look into what the heck the sequence is and where the modifications occurred. Then you probably would chop it into smaller pieces and do LC MS MS on those pieces. And that, as I said, is an experiment that is not one afternoon that could take you a month. If you are a very proficient with the metal. Thank you. And so, so what you're saying I think is is what you would hope is you could deal with that heterogeneity in the separation method. If you could, because mass spec it's going to be harder. Nowadays, I would say it's about equal. It is hard to get a good MS spectrum. But we now have a good mass spectrometers and good methods. I think it is possible to get it. Thank you. Any questions. So just a naive question from my side so how complex can a mixture be for you to resolve on your on your columns in the, in the ideal case. Okay, as I said, you digest it, it has a 4,000 nucleotides, when you digest it with certain enzymes are nice T one is a good example. You get your 150 peaks or more, you will resolve partially or completely about 50 to maybe 100 of those. If you have any solutions. You will have to go into MS level to really figure out what each of those peak is, and even MS will not give you a complete answer because, well, you have some scrambled six measures, there is a good chance there will be five of the isobaric. So I just scrambled sequence right that it exists in that mRNA, and you will not know simply by MS, which one is which. And now, can you confirm based on those data, the sequence, if you don't know. So you will have to go to MS MS. If you really want to know the separation power is quite amazing, but never will you resolve everything that is there. All of those peaks may hide something minor some minor components that are real, and you may not know what they are. More questions. Yeah, we're in time, then. Thank you, Martin, and then let's move on. Next speaker will be Mike Greek, who is the executive director of pharma and biopharma broker. Everybody. Let me get my slides to share. Is this in full screen mode now. It is not. Okay. So thanks everybody for inviting me. What happened. I got to make one change here. The slides were hidden from me previous. Sorry, technical difficulty here. Okay, sky this. All right. There we go. Just a bit on my background. I basically started my career doing all of those in RNA, many years ago. So it's pretty exciting to see that they are coming back into play. My job was that I just pharmaceuticals, which is now I own us. And basically all my early research was all the goes in RNA, joining brook or five years ago. And we've started a lot more focused on this. I've shown a lot of data. I can show you all at notes if you want or we can discuss some more. I just wanted to kind of briefly go over what we have and we've room for discussion. I think intact mass overall is, you know, we're in pretty good shape for intact mass up to 120 years. And same, you know, for software and other things interpreting that. We also have moldy. And, and when I talked about intact mass and talking about with isotopic resolution. That means we can get the PPM mass accuracy level which a lot of people demand. So with ESI we've done 120 more plus with this type of resolution. And so that's serving most of the needs, you know, the even tRNA some poly a tails guide RNA things like that. But I think overall the technology is in pretty good shape there. Maldi top we have a lot of customers using Maldi when they're just looking at oligos. Maldi can be quite good, even with isotopic resolution of the 30 murders. And so for people, you know, producing a lot of oligos different things, they can use Maldi and it's very high throughput. And so, in currently software platforms, whether it's ours or any of the other vendors, I think, for intact mass, I think overall most of our customers saying we're in pretty good shape there, where we need to improve is the sequencing. So right now what we do for sequencing for electric spray, it's all CID. Maldi can also be used for sequencing using ISD, combined with CS, CID, Maldi sequencing, what we have found first with electric spray using CID is we can generally sequence almost anything we've done with up to 70 and 100 Merce intact full sequence of CID. As long as you have good enough fidelity in the data and software you can you can pick that out and this is even modified oligos up to about 100 Merce. With Maldi ISD, it's really more kind of a for confirmation, the way ISD works, it's not predictable. And so we don't know when we're going to get, you know, 20% coverage versus 80% coverage. So as I said, some people use that just as a quick verification. Okay, we have part of the sequence, we have the right mask. But this is something that can be developed in the future. Our software is called Biopharmacompus oligoplast is the portion for the DNA and RNA oligo sequencing, and that's still evolving. Clearly, improved top down sequencing software is needed is the data, especially for as you get bigger and bigger, it's very, very complex. One thing I don't have on here existing college existing technologies is any of the upfront platforms and from photography, we need that to our friends at Waters and Agilent because they're experts at it and that's not exactly our specialty. So, I don't discuss that much here. As far as developing technologies, we are looking at things on the front end. I started my career working with oligos and in our lab, nobody was allowed to have french fries for lunch because oligos and RNA and DNA will suck sodium and potassium out of the atmosphere. And as some of the slides from Martin showed so we really need good ways to efficiently desalt these samples. So we're working with a company called Integrated Protein Technologies, they make a device called a sample stream. It's a membrane based microfluidics, and we found that works really well to be solved. And we've also done an mRNA digest on this. This is something we just presented at this here. Basically, the ideas is you can take an impact mRNA. The way this device works is you have a membrane so we took a 10 K membrane. We just took an RNA so that was big enough to say on top of the membrane. And the nice thing is is once you start to digest and once you get about a 20 mer it actually comes through the membrane, and then we can sequence those to do the mass mapping quite easily. So this is the type of technology in the front end that we are developing. We also just bought a company called Phasmatec, they have something called an Omnitrap. This will help us more for sequencing and enables MS to end all the electron type association techniques and UVPV. And even though we find that a CID we can generally sequence almost everything with the proper software, we have found specific regions in mRNAs, which I probably have some type of structured region where we're going to have to do some collisionally activated unfolding or something else before we can get the full sequence out. So we know we do need to expand on the different types of techniques for MS MS. We have trap-diamobility on our instruments. It works great for anything from small molecules to peptide levels. And so we think we can use this to really improve the separation and analysis of nucleotides, especially having modified nucleotides or with stereo centers. And we also have quite good AI prediction of CCS values for lipids, for peptides, and we believe we can do this for nucleotides too as there's already small molecule prediction. So we think collisional procession is something that can at the nucleotide level can be really important in the future to be confident in how you're identifying these things. Another thing that at Gruker we do is we do a lot of moldy imaging of tissues. We get asked about imaging of oligos or RNA and the level of tissues is too low to do, but now we're working with a company that makes probes, basically antibody probes to attach to oligos and that's something that we're developing now to see if we can look at distribution and tissues of oligos. This technique works great for eyes and a lot of other tissues. And so it might be a good way to identify not only how your oligo drugs distributing but potentially for larger type of molecules. Sensitivity for PKPD where it was mentioned by Martin, you know, we need better ways to get higher sensitivity to actually look at PKPD. We can look at a lot of these molecules when they're prepared and purified, but to actually look at them out of plasma and tissue samples, we're really going to need boosted sensitivity and also the front end sample prep, how do we get those things ready? And so that is something that we have some instrumentation when we think we'll work well, it's just something to develop more. What's really needed in my opinion is improved enzymes for RNA sequencing. Right now the choices are a little bit limited. So potentially using this membrane type approach will help us because then it doesn't really matter what the membrane or the enzyme is and so that might solve some of the problems. The role of ion mobility, especially as you get to larger and larger oligos or structure to RNA, how much can this help us and what can we do as far as predicting potential collusion of cross-section. And then of course CDMS, charge detection mass spec, it's available now, but it's not quite to the accuracy needed to be to look at intact mRNAs, especially if they're inside a liquid nanoparticle. And that's really all I had here at a couple other slides, we want to look at pictures, but I thought this should be more about the discussion or other questions. Thank you so much, Mike. Susan. So Mike, I thought that was great. But I had some questions because my spec is in my area, although I've used mass spec with with great respect. When you were talking about high fidelity data, because you said the software wasn't was okay as long as you had high fidelity data and I wasn't sure what high fidelity meant. Does that mean a large amount of sample or lack of heterogeneity or what does that mean? A good question. So let me get my laser pointers that you can have it on here. I included this slide because it should be so what I mean by high fidelity data is that the actual quality signal to noise in shape of the signal is is very good. And so, whoops, sorry, this is, I think this is animated for anyway, if you look at the red is the predicted isotopic envelope and the blue is what we measured. So high fidelity data in our term means that you what you predict it should be matches very well with the measured amount. And when it matches well, you can get a very, very accurate mass and parts per million. A toss are really, really good at this trapping mass spectrometers are not good at it. And so if you don't have this fidelity of your isotopic envelope, say one of these peaks here one of these peaks here is not where it should be you cannot predict the mass. And so you really need good ion statistics and good use of all your eyes. So that's hopefully that was clear enough for a non mass spec person. Thank you. Did that make sense. Brenda. Yes, thank you for that nice talk. I'm, I'm interested in your comment about. Sometimes there's problems with messenger RNAs and you were speculating that maybe there were structured regions, if I understand stood correctly and I'm trying to talk about that in comparison to t RNAs which have a huge amount of structure, which, and maybe you're talking about okay t RNAs you chop them up so much and so the structure is not a problem but I wonder if you would put that in context for me. Yeah, I would love to do that more but we're just not quite sure to be honest. So we've actually sequenced t RNA top down almost completely. And that's what Dan for Bruce he's actually the one who wrote the original algorithm for our software. So we partner with him a lot. And so t RNA doesn't really seem to be a problem with the m RNA we looked at it was about 1300 nucleotides and when we did this using different enzymes and also our membrane method. We're in the same region in, in every way we've done this, and we're not sure if it's, it's hard to tell if it's a structured issue if generally we don't have an issue with modified nucleotides, but maybe there's a certain sequence of them. There's a certain biology that sometimes that when we can't sequence by CID biology also can't deal with it. And so it's, it's, we don't know. Yeah, thanks that that makes sense. And so when you talk about RNA enzymes here are you talking about nucleosis. Yes. Okay. Sorry. Another question Susan. Another question if I may. And this just reflects a defect in my note taking. You said the software was was good enough in one situation but then you said you needed better software for another situation just reiterate what that different situation was. Yeah, so basically, when people are looking at intact mass and want to look at purity so basically if you look at like say this one here. This is the LC chromatogram and this is the UV signal, but the mass back total iron chromatogram look the same. And so, you know, waters Agilent others have done a really great job with chromatography, and getting better and so the software is pretty good at taking those of say, you know, shorter piece shorter oligos or pieces of RNA get good separation and we can quantify quite well, what these various intact deficiencies can be when it comes to MS MS in doing sequence confirmation and analysis or especially to know those sequencing. The software is not fully automated. So you can't take a chunk of RNA, throw it into a system and have the whole process from front end to back and automated where you start with sample prep to full sequence coverage in an automated fashion. At least not that I'm aware of. Okay, I get it. Yeah, that's good. Thanks. Okay. Thanks again, Mike. That was really good informative great. So the last speaker of the session is Dr. Ben Dunstead from who is a senior scientist at Agilent. And yeah, welcome. Let's see where we'll get his slides up. Yeah, I'm working on it. We're getting there. Okay, are you seeing the normal mode as well. Is correct. That's correct. Okay. Okay, so thank you very much for inviting me to do the meeting. I'd like to thank the organizers for the opportunity to speak. And I think it's clear from the context of this meeting that, you know, and LCMF can play a critical role in the characterization of nucleic acid, both as natural components and biological systems, as well as synthetic products for the use in emerging technologies. The ability to effectively couple high performance liquid chromatography separations with high resolution electrospray ionization based mass spec has really only existed since 1997, with the introduction of HFI P ion pairing approach. In this short talk, I want to present a retrospective review of the development, the techniques for characterizing oligos from Agilent's perspective as both the provider of synthetic DNA and RNA into the manufacturer of the systems used to do the analysis. And we'll also discuss optimization improvements and emerging trends and opportunities for the future. I spent a lot of time on the synthetic side of this so I'm very happy to talk to researchers, looking at native oligos and learn what they're looking for. So, here's the outline. I'll go, you know, a brief history then factors the effect. And we'll talk about separations and MS signals, and finally talk about, you know, similar to what the other speakers talked about, our tools for impurity analysis and sequence confirmation. So a brief history to start in the mid 90s, HPLC separation of oligonucleotides was a well established techniques. The dominant approach was based on the use of triethylammonium acetate as an at neutral pH as an ion pairing reagent in reverse phase chromatography. However, it's quickly found that coupling the separation with electrospray ionization was not trivial, sufficiently high concentrations of TEA to maintain retention and resolution led to high levels of signal suppression in the SMA mass spec. Using a concentration of TEA, the levels where ESI could generate useful signal resulted in low retention and poor resolution as shown in this figure. So you can see if TAB increases, you start to lose that retention, but that's where you start to get a signal in mass spec. In 1997, we developed and published the HFIP TEA ion pairing method that overcame this problem. The proposed mechanism of this method is that ion pairs and pH are maintained during the separation resulting in good retention and resolution. However, during the desolvation of the ESI droplets, the volatile HFIP is removed, and at least at the surface of the droplet, pH rises. The oligos are ionized and equilibrium favors the free charge of oligo, which can then undergo ion desorption into the gas phase. This time HFIP TEA method has been widely adopted and a number of studies have examined the mechanism as well as optimization of various parameters. The separation of itself is relatively complex and involves a number of variables, some of which are listed here. You know, and they're all subject to optimization. However, when you couple the requirements of the separation with often competing requirements of the mass spec is a difficult problem. More different problems often require different specific conditions, so there isn't one, you know, best condition. It is important to optimize the speed and resolution of separation to balance that with the sensitivity and spectral information for the mass spec produced. As a secondary but important factor LCMS grade high purity HFIP is expensive and there's a desire to reduce consumption this product. These have evaluated alternative flora alcohols, some of which are shown here. To date, none have shown broad improvement over HFIP. Furthermore, having emerged as a standard reagent for this application HFIP is commercially available at high purity and remains a go to reagent. However, the choice of alkylamine ion pairing reagent is a different matter. Longer more hydrophobic alkyl means resulted greater retention and separation in some cases. Depending on the target all go under study this can be exploited. Beyond their ability to affect selectivity most alkyl means shown here behave similarly, although adjustment of gradient conditions are required. In general, triethylamine, tripropylamine, hexalamine and octalamine are the most widely used. The pH of the separation also plays a key role. While the original publication used neutral pH it has since been shown that a basic pH between 8.5 and roughly 9.5 results in a good balance between signal retention and column lifetime. Data here show the best signal for this 130 former is at 8.85 with dramatic fall off on either side. In terms of spectral characteristics one of the challenges that previous speaker have alluded to is, and especially for longer oligos results from attic formation, particularly from sodium and potassium. This effect gets worse as all the length increase and these addicts can compete can complicate the spectral the spectra and hide impurities. We have found that desalting as well as low levels of edta can significantly reduce the attic formation and clean up both the raw and deconvoluted spectra. And so here's an example of this, you can see a lot of the sodium addicts here that have been greatly reduced. And again this gets more difficult the longer it is. Of course the column characteristics play a key role in separation and LCMS compatibility. There are a number of suitable columns of the market that meet the requirements for the application including Agilent's advanced bio oligonucleotide column. This 2.7 micron 120 angstrom for C18 is optimized for oligos and stable between pH 311. And it's good for thousands of injections. I've shown data here for range of temperatures but technically the operating limit is 65 degrees. The column also is available in wider diameters for prep and semi prep applications. I briefly want to turn away from HFI PTEA ion pairing method. And I'm sure that some of you are aware of growing concern over PFA's. These polyfloral alkyl substances are becoming significant environmental problem. Because they're widely used in a bunch of products. So we're a little concerned, especially in the EU that the European Chemical Agency is taking steps to eliminate PFA's. And the implication of this is that we may no longer be able to use them in the future for oligonucleotide separation. So one approach to eliminate this that was spoken about earlier is hydrophilic interaction chromatography. A hillock separation uses C-12 as a weak mobile phase and an aqueous buffer is a strong mobile phase. Recent improvements in stationary phase chemistry such as Agilent's hillock Z column are making this a potential emerging solution. And so shown here are some separations that can be achieved using this technique, which is compatible with electrospray mass spec. And so these are just some small 40 mer DNA and 20 mer RNA. It's not at the level and not as good as the ion paired method that have been talked about previously, but it may become necessary. Finally, I want to highlight a couple of common workflows that are supported by the by Agilent and in particular our informatic solution bio confirm 12. The first is target plus impurities, which is used to identify and quantitate impurities of given given oligo. So the workflow shown here. You start up by, you know, set up your LCMS run, which use chromatography separate the target and any impurities as best as you can. And this has been fully covered by previous talk how important and useful it is to separate that out. The next step is to enter the nucleotide sequence. The system will generate a database based on target impurities, both stuff that we know about and, you know, things that the customer can put in as well. Then it uses a feature finding technique to find the actual oligonucleotide compounds in the data. And finally, it compares these futures features to calculated masses or isotopic signatures of the target impurities. And this is an example of the isotopic signature, but it was well described previously. This slide shows some impurities that can be identified using the software. This particular data is interesting because we're looking at a synthetic hundred 50 mer RNA that has paste modified bases that are shown here. These modifications are known to undergo and undergo an in source or the factual decarboxylation over by modifying source conditions we can we have substantially reduced the decarboxylation signal, allowing deep urination and deep urination to be identified. So that's here you can see this minus 44 for the three pace modifications that were in there. That was eliminated by adjusting the source conditions. And now the software can identify a different thing, some of which I've highlighted here, you know, and minus one and plus one and some deep urination, deep urination, perimeter nation signals. You can see the sold addicts. And this is an example you have 150 Mer's very hard to get rid of those. The second workflow, you know that all of us are working on is a bio confirm in bio confirm 12 is called sequence verification. This is currently as close as we get to sequencing a leg of nucleotides, but it does require an oligo sequence to the software knows what it's looking for, because as have been described the data is very complex. So the workflow here is you start up by setting LC MS run, which will generate iron fragments using MS MS for that target sequence. You also need to enter the nucleotide sequence and the system will automatically generate a theoretical list of fragments isotope patterns to look for. And here is the fragments, the McClecky fragments that are, you know, generated by this fragmentation process, as well as we do see some base fragments that are cutting off. So it'll generate this huge list. And then this will be seen in the sequence ladder, which I'll talk about later. But the final step is to compare all these theoretical fragments. And the software will look for the acquired MS MS spectra and annotate which fragments are matched. This slide shows the sequencing workflow window sequences can be input with any modification by adding the chemical formula both of the nucleotide and the nuclear base. So, yeah, here's what that screen looks like. The workflow setup is here where you enter all the mass spec conditions, etc. And then you get your TS, your tick, which shows how the separation worked. And then MS MS spectra, which can be quite complex, but the colors are indicate which fragments that it found. Over here is the fragment confirmation layer, which I'll talk about in a minute. And finally, the target oligonucleotide fragment ions that were identified, and these are scored based on how well it fits the predicted data. Okay, and so here is the fragment confirmation ladder of a 40 murder that we sequenced. And the dots indicate the fragment type, so at which point it cut. And, yeah, and so then it looks at the entire sequence and determines, you know, which fragments verify the sequence correctly. And this data or this slide is to show that you can combine multiple files to get a good overlay and the different colors indicates, you know, which fragments and what the final sequence is. So in this case, we got most of the sequence covered. But we're missing a couple spots. So we did a different one, for example, on a different charge state of the initial oligo to get complete sequence coverage. Okay, so in conclusion that I don't need to reiterate at the short talk. So, you know, basically went through the history, where we are today. And I think the improvements needed have been talked about in the past. So it's just expanding the software to be able to do de novo sequencing. And again, this is going to look a lot easier for smaller oligos, but it gets more and more difficult as they increase in length and complexity. So with that, I'll end and I'd be happy to take question about any of this stuff. Thank you very much. Any questions from the committee. So how are you. I didn't, I mean, in the end, you presented how your, how your software is dealing with the LCMS MS data so are you integrating modifications into that workflow identification of the of the stuff. Absolutely. So, I mean, we have a lot of the basic modifications in there, but you can absolutely add whatever modifications you want. And the way this is done is that you put in the chemical formula of the modification you're interested in, and you put in the chemical formula for the nucleotide as well as the nuclear base. So if you get, if you get a fragment that cuts here and you can also get a fragment that cuts at the base so you can, you can, you can find modifications that occur both at the sugar and at the base, potentially, although it probably requires a lot more experimentation to to weed those out. But yeah, absolutely any modification is possible with the software. And again, this is stuff that has been done in the past, but this makes it a lot easier because it does a lot of that work for you now that you used to have to generate the stuff on your own. Did you call on me Marcus I didn't hear. Okay. Thank you Ben and I appreciate I particularly enjoyed hearing some of the history of the solvents that I didn't quite know about that and my question is, is actually directed to to all the speakers and as you know, our community is really interested in pushing or or what is the chance to push this technology forward for particularly for RNA based technology and, and I'm wondering how each of the speakers would describe the motivations or the drivers for their companies to be invested in and develop new technologies for RNA mass spec and yeah I that's that's I would appreciate any comments on that. Yeah, it's a good question I guess I can start, you know from my perspective. It's going to be market driven. And the way that normally looks for something like the software is that people request certain things like the sequencing technology and the more and the more that that happens. The more that we you know track that information and you know through our sales people are or general inquiries. And so, you know for a long time mass spec was very peptide focused and you know the markets changed more so you know with the therapeutic oligos coming online so that's the direction that this is driven so that's one good way to do it. Another way is to you know form collaboration. In, you know, fields that we think that there could be a market or interest you know we always want to be pushing the envelope. But yeah, it's we also are interested in what the customers want so I think collaboration's been what what do you mean partnerships with academics or what what exactly do you mean by that. Yeah, we do that so we'd like to partner with the leaders in the field to push that. And so it makes a lot easier because then we don't have to do some of the development work on the side so developing new technologies that it's a great way to do that. Thanks. Any other speakers have comments on that. I think it's, this is marking your similar. It's market, market driven. So, we now see the big upswing because of M RNA, because of fast development of SI RNA technology so we all are getting these requests from customers. And therefore it makes sense to pay more attention to it. I also have a history with nucleic acid analysis and so on so part of my drive is scientific interest. If you set that comment is interested to push the technology and maybe a repurpose of what is available there into your field. That's maybe something that we don't see that often. But if you actively approach the companies. After key contacts. We would definitely consider some like discussions collaborations. We co publish papers and so on and so on. Thanks. I guess there's one more person if they want to weigh in. Not a whole lot to add and it's all market demand driven. You know, as I said, my background's all the girls in RNA and so I'm excited to see it going. But it's, it's pretty obvious when you look at the market what's going on. People are talking about RNA, RNA, RNA seek things like that, just all the time. Good. And it's really kind of coming down to transcriptomics, all the way down through proteomics and every other type of omics data starting to be in high demand to build a multi platform analysis for discovery. Thanks. Yeah, I this is for all the speakers as well I was kind of wondering. It's more of a workflow question so I know, you know, a lot of comparisons have been made to proteomics and I'm kind of curious at this moment where the field stands for taking, you know, doing the same sort of workflow that proteomics, people do you know digesting sort of proteins into smaller peptide pieces separating those, and then analyzing them and then comparing the MSMS to a database of known protein sequences I'm wondering like, is there a common workflow that exists currently for RNA, or is that just not there yet is the separation combined with the MSMS combined with, you know, the limitations and software is there. Do you see it going in that direction, or is that sort of still a limitation because I'm seeing that most a lot of what I'm seeing here is very targeted, I don't own RNA sequences or molecules but I'm wondering what the discovery based method, like outlook looks like compared to what proteomics is doing currently. Yeah, I guess from my point of view. The technology for oligos and RNA hasn't really developed much in, since I started it until recently. So I think workflow wise for kind of work proteomics was 15 years ago. It's a little bit segmented, people still look at the MSMS data to make sure it looks good, or people in proteomics never do that anymore. So I think it's, it's moving that way but it's quite a ways behind as far as complete workflow that can be industrialized in that formula. The perspective is that there are similarities between proteomics and let's say genomics, including some of the mapping like peptide mapping or RNA mapping. All nucleic acid work is two reorders of magnitude less sensitive in mass spec. That's a big problem. Second, we can make up for it by some PCR methods we can amplify certainly pieces of DNA RNA and to be used that. But obviously it doesn't really solve the problem if we are now looking for some modifications you that amplification would basically eliminate that information. So it is there are some similarities but no the genomic is going to be slightly different different story. Yeah, but I agree with both Martin and Michael on that. It is behind it has been for a long time. Unfortunately, because I love Olicos as well. But yeah I think it'll follow a similar pattern, similar methodology but it just needs to be developed. If there are no more questions. I would like to thank all the speakers again this.