 I've got a cup of tea over here because I'm British, so, you know, we've all got our means. Good, good, good, good. Getting too late for tea. So I'm for beer in Germany, right? Yep, exactly. Yeah, I just did beer hour for the Institute, so I'm the director of the Institute, so we had beer hour, moved it forward for a day. Are you asking people to supply their own beer? No. Yeah, so actually what I'm going to do is I'm going to have later, because it's a tradition with my best friend to break a fast lent. And so we normally get together, so this time I already picked out a bottle of wine, I'm going to put half of it in a decanter, leave it here. I'm going to take the other half to his house, put it on the garden fence, go back to my house, and then we're going to drink from the same bottle. Corona makes creative. Say what? Corona makes creative. Yeah, come up with new solutions for all problems. Yeah, exactly. Miranda, can you let me know if we're live on YouTube yet? I think I'm just struggling. We're live. Okay. All right, over to you, Detlef. Hello, everybody. Welcome to today's open research webinar hosted by Elife once more. The series, as I'm sure you've all read, aims to give early career researchers an online platform to continue to share their research as an alternative to the in-person gatherings, which we currently don't have anymore. My name is Detlef Weigl. I'm a deputy editor at Elife. I've been with Elife from the beginning, 2011. Early career researchers have been very important to us from the very beginning. We always wanted Elife to be more than just a journal. We wanted it to support researchers and careers in many different ways and this is really a great way that we are able to do this. I myself am a plant geneticist and mostly study plant evolution these days. So today we have again three talks. We'll first hear from Rob. Rob taught at Creighton University in the US and he'll tell us about expandable and reversible copy number amplifications driving rapid adaptation to antifungal drugs. And Rob is a recent Elife travel award winner and he is presenting today the talk that he otherwise would have given at the Microbiology Society meeting on Candida and Candidiasis. And of course Elife will support. Rob was an award for travel to the next conference. He has a talk as well. And then next up will be Natalie Clark from Iowa State that aims also in the United States. She'll tell us about generating multi-scale predictive networks on northern corn leaf flight resistance. Really excited to hear this about Natalie because I think about disease resistance a lot. And then finally, we have somebody who is a little bit closer to my home to tubing and Christian Minch from Goethe University in Frankfurt. And I have a very, very timely presentation on proteomics of SARS coronavirus to infected host cells and how this could lead to the identification of potential therapy targets. So after each 10 minute talk and just to remind the three of you after a minute I'll, you know, make noises or raise my arm whatever. After each 10 minute talk will have five minutes for questions for the speaker before we move on to the next talk. And to ask a question you can type it into the chat box and zoom or directly onto the into the Google document which is linked there. And this is also in the chat, the Google doc. And we are joined today by Miranda and Anja and Naomi from Eli, who are working in the background to support us. And they will also help me line up your questions. If you are able to ask. So, I will invite you by name to ask your question and team will unmute you so you can do that in person. And otherwise we can also just read your questions out loud and include your name if you wish. So this document is also a place for you to contribute shared public notes and we welcome you to do so, and to list yourself as a contributor in the list about speakers for today's webinar. Thank you for doing this and participating in today's talks. Finally, should let you know that we are recording the webinar and also live streaming it on YouTube and this is what you're seeing here. So, just as a reminder during this live webinar, please be respectful honest inclusive accommodating appreciative and open to learning from everyone else. Do not attack the mean disrupt harass or threaten others on courage such behavior. If you feel uncomfortable or unwelcome on any of these webinars, please contact us by email events at a lot of sciences.org inboxes monitored by any stars at Eli. And we of course reserve the right to ask anyone to leave and or to deny access to a subsequent webinar on zoom if there's ill behavior. If you help send a chat message on zoom to host or directly to Miranda now we are on. Yeah. All right, so, first we'll hear from Rob, and over to you Rob. Awesome thank you so much I really appreciate it. All right, can you see my screen and hear me. Perfect. Great. Well, thank you all for being here today and thank you to the organizers for putting this on. As you have said my name is Rob Todd I am from the cell Mecchi lab at the University of Minnesota and I'm a PhD student. Right as I begin I want to begin by saying that genome plastic is incredibly common within eukaryotic organisms. This, this can manifest as either shifts and ployty increases or decreases and the genome copy number, or an annu ployty and abnormal chromosome number. For example, here I'm showing you on the left a you ployd human karyotype in healthy non cancerous cells, where each autosome is present in two copies per genome. In contrast, if we look at the right panel, an annu ployd human cancer karyotype, we see some chromosomes are present in two copies such as from some one summer and three and some are even present and more copies than that. And besides just in human cancers, and you ployd is incredibly common within human fungal pathogens. In fact, 50% of all anti fungal drug resistant isolates of the human fungal pathogen Canada albicans are and you ployd. Canada albicans is mostly a commensal organism. It's found within the GI tract it's found within mucosal membranes of healthy individuals shown here as the red sectors of the pie chart. However, in immunocompromised individuals, Canada albicans can become pathogenic and lead to severe systemic infections. This led the CDC and this past year to release a report that said that drug resistant Canada species are a prime concern and are a serious concern for the public health. Now, before the dawn of the genomics age it was already well appreciated that this the genome of this organism is quite plastic. Unlike the karyotype that I showed you with the cancer cells previously and see albicans we use pulse field gel electrophoresis to identify karyotype differences. Here on the left I'm showing you seven different clinical isolates again same species just different clinical isolates. I'm showing you the karyotypes of their large chromosomes chromosome one through four and are, as well as the smaller chromosomes five, six and seven. And what I hope you can appreciate from this is that there's a large amount of diversity between clinical isolates of the same species. Besides just differences in chromosome sizes, loss of heterozygosm is also common within this organism, especially during passage through a host, and when treated with anti fungal drugs. This loss of heterozygosm can be due to non disjunction events which home as I go to an entire chromosome, or short range or long range LOH events that occur via recombination. The driving questions that I've had during my PhD is how does genome plasticity impact pathogenesis and the acquisition of drug resistance. And so, besides just chromosome size differences and the acquisition of LOH, I'm keenly interested in how copy number variations or CNVs arise within a population to study this I use whole genome sequencing technology. And to visualize that technology here I'm showing you the eight chromosomes in the SEAL against genome chromosomes one through seven and chromosome are each one of these boxes is a schematic of the chromosome. And on the x axis the left side represents the first base pair, and on the right side of each box represents the end base pair for that chromosome. On each chromosome there is a notch that represents the position of the central mirror. The difference of this plot is chromosome copy number from one till four. You'll notice that there's a black bar that runs along each one of these chromosomes at copy number two, indicating that each of these chromosomes is present within two copies in the sequence genome. This is our lab wild type strain. However, when we grow the strain in the presence of anti fungal drug. We often observe a wide array of any employees and CNVs, including trisomies of entire chromosomes such as the trisomy of chromosome platelets, as well as CNVs that amplify the left arm of chromosome five, and a recently identified CNV that amplifies the right arm of chromosome four. These are incredibly common CNVs during anti fungal drug treatment with the as all family of anti fungal drugs. These CNVs all share a common breakpoint, and that is the center mirror. The center mirror of C albicans is inherited epigenetically through some a binding core shown here as the gray shaded boxes down below. And they are flanked by long inverted repeat sequences shown as red arrows. Importantly, only chromosome four and chromosome five have central mirrors that are flanked directly by these long inverted repeat sequences. This led us to ask the question are other long repeat sequences associated with genomic changes within this organism. And to address this question, we use the bioinformatic mummer suite to identify all long range sequences, repeat sequences within the C albicans genome. Through this analysis we identified nearly 2000 long repeat sequences that ranged in size from 65 bases all the way up to 6.5 kbs in length. We also shared similarities between the repeat matches between 80% and 100% identical. These repeats were found either on the same chromosome or intra chromosomal repeats or on different chromosomes for instance chromosome one chromosome three, that would be an inter chromosome or repeat. Here on the left I'm showing you a diagram of the eight C albicans chromosomes, and I'm color coded the repeats so that they all match their sequence identity. For instance, there are three light blue or rabbits at blue lines present on chromosome to this represents a very similar sequence being present not on one button three copies on that chromosome. Importantly, in the study and a survey of human clinical isolates isolates that were pulled from a mouse model, and through in vitro evolution experiments during exposure to antifungal drug, we identified multiple copy number variations. Copy number variations had a copy number breakpoint associated with one of these long repeat sequence sequences. In previously published data sets we identified that over 50% of the long range loh events also coincided with these long repeat sequences and bioinformatically determined sequence inversions also occurred at these long repeat sequences, indicating that these sequences really drive genome plasticity within this human fungal pathogen. So, I also was really interested in developing a program to understand how the CNVs arise within a population. To do this, I use in vitro evolution, where we expose progenitor isolates shown here is for both progenitor and progenitor be that are susceptible to antifungal drug to antifungal drug for 100 generations. After 100 generations of time I sequence the resulting population and identify if there are any CNVs that appear. Here I'm just showing you two examples from diverse genetic backgrounds that we identified a CNV in. And surprisingly, what we saw is that we saw massive amplifications of the genome that ranged from amplifications of single copies all the way to over 12 copies per genome. Importantly, none of these CNVs were found occurring at the centromere like the CNVs I previously described. So we were really interested in determining what was the mechanism and structure of these CNVs. So to do that we use a combination of read depth technology as well as allele analysis to determine where the copy number break points occurred. Here I'm showing you read depth chart where the X axis represents the chromosome position of chromosome four in this case, and the Y axis is the relative read depth. So on the right side, if we zoom in on that copy number break point, we see not one, but two copy number break points, leading to a stair step amplification to the maximum CNV height. Importantly, each one of these break points was associated with one of the long repeat sequences, known as repeat 148. So to zoom into the right side of the CNV, we again see a stair step amplification, except this time the copy number break points occur at a separate distinct inverted repeat, indicating that these CNVs occur at this occur between two distinct inverted repeat sequences. The next one is to address if these CNVs were beneficial in the presence of antifungal drug. To do this we took both a progenitor isolate shown in black, and the evolved isolate that has the chromosome four CNV that we just discussed shown in blue, and grew it in rich medium shown on the left and in rich medium plus one microgram per mil of fluconazole, the physiological concentration. What we identified was that there was no significant difference in either the max slope or OD, when these two strains were grown in rich medium. However, when grown in the presence of fluconazole the strain that contained the CNV had a fitness benefit. And importantly, what we noticed is that all CNV isolates have an increased fitness in the presence of antifungal drug. And then we began to ask the question of why are these beneficial. And so in this slide I'm showing you one representative image of a CNV that we identified on the right arm of chromosome three. Again, chromosome position by read depth. In this CNV, again flanked by two distinct inverted repeats forming a stair step amplification. Amplification of the gene MRR one by over 12 copies per genome. MRR one is the transcriptional regulator of the multi drug ethox pump and DR one, which has been associated with antifungal drug resistance. Importantly, this was not the only strain that we saw this type of amplification in. When we sequence other samples from diverse genetic backgrounds, we did not identify just an extra, but two additional strains that also amplified MRR one using the exact same long repeat sequences, but amplified it to varying degrees from four copies per genome, all the way up to 12 copies per genome, indicating that this is a conserved mechanism for amplifying important drug resistance genes. And so in summary, we've identified novel CNVs within the Canada albicans genomes that are flanked by two distinct long inverted repeat sequences. These CNVs are adaptive and amplify important genes involved in antifungal drug resistance. Now, I don't have time to go through today, but we have discovered a mechanism for how these CNVs are formed, and we believe that they are generated via Dicentric chromosome intermediate and the you and the break intrusion bridge cycle. So with that, I would like to again to thank the organizers and as well as my lab members, Dr. Anna Selmeki, my mentor and advisor, as well as the members of the Selmeki lab who've made this such an amazing four years for me to do my work. And so thank you all and be happy to take any questions. Awesome Rob. Thank you very much. Thank you. Nice. And also thank you for being so precisely on time. So the first question we have is from Ashley Hall. So Ashley, please unmute yourself and ask your question. Hi, um, so I actually don't have the text that I wrote right in front of me, but my question was, where are these new annuplities mapping to are they mapping right next to the native location of the sequences being duplicated or triplicated or however amplified or do you not know where they're mapping. Yeah, that's a great question something that we've been working on figuring out so what we have done is obviously the CNVs that I showed you that's just mapping back to the reference genome. So we needed to figure out whether this was an inline amplification. I'm not sure if you work in yeast but extra chromosomal circles are also really commonly found in yeast. So what we have done is we've developed a bunch of assays that use restriction digest and use that pulse field technology. And what we're seeing is that we are amplifying larger and larger fragments of the genome, which indicate that it's actually amplifying on a single molecule. So that chromosome itself is expanding and almost like an accordion like a ray. And how we verified that was the data that I showed you was all conducted using Illumina next gen sequencing. We've also done long read sequencing with the Oxford Danipur Minion. We've identified single long reads that actually enter in one part of the genome and recombine between those two long repeats, creating a full back mechanism that allows for the formation of a nice center. So what we believe we're seeing is a very large tandem accordion like inverted repeat structure. Does that answer your question? Yes, thank you. Rob, I was wondering what is known about these long repeats and different strains of Canada. Are they are they conserved or are they wildly different? Would you expect very different patterns of amplification if you started with different isolates? Yeah, that's a great question. So the repeat identification that we did was all conducted in the background of wild type strains. So we took the reference genome and basically aligned it to itself to identify these long repeats. So we haven't done that in other clinical isolates. That is something that we're looking at doing as well as with other fungal pathogens. Some of actually a majority of these repeats were previously known. They were known to be present in multiple copies, LTRs, TRNAs, those type transposons. However, we were quite surprised to see that many of these repeats contain open reading frames, entire open reading frames, or groups of open reading frames. So for example, we have one repeat actually located on the left arm of chromosome 3 where we see a CNV occur that contains four open reading frames that is 99.9% sequence identical to a region 11,000 base pairs away that was not known to be in multiple copies within the genome. So we do have, we do see multiple different types of genomic structures in these. But whether we see different CNVs in different organisms, all the data that I just showed was from very diverged genetic background. So if you remember one of the first slides when I showed you the C-albicans' karyotype where all those chromosomes were a little bit different. Those are actually the progenitor isolates that I was using for this study. So we do have diverged CNVs coming up that seem to be pretty common. All right. Thank you, Rob. Next question is from Marisa, too, and Naomi has gone and unmute her so she can ask a question ourselves. Marie. Can you hear me? Yeah, yes. Oh, that's great. So a great talk. It's pretty interesting. So I have some questions. First one is, are two of the four gene gain rather than gene loss? Yeah. So the data that I showed you was all gene gain. However, we have identified gene loss. And that was actually part of the mechanism that I didn't have time to go in today that led us to the idea that we were getting disinterchromosome formation. A lot of times what we see is that when these are forming, because the CNVs that I showed you were all stable within the genome, when they're forming, we believe that they are actually unstable. And it's due to a distal break. So a break near the telomere that then leads to the chromosome truncation. So you actually see a loss of part of a chromosome before it then recombines and forms the CNV. So we do see both. It's just at different times temporally. I see. So you think it's a combination, the result is combination of adaptation and some mechanistic force to form a copy of a variant. Yeah. So the second one is sorry. Yeah, no, no, no, just finished my question. Yeah, second one is, did you observe more polymorphic CNV rather than fixed copy of a change? I'm sorry, can you repeat that? Yeah, so we actually see other CNVs that occur at the same position that have different applications. We have one CNV that's present on the left arm of chromosome three that in one isolate amplifies it, just one extra copy per genome, and then three extra copies for genome in a different isolate, six extra copies and another isolate and then all the way up to 12. So we do see those kind of polymorphic expansions. Yeah, the last one is that our CNV is more functional. I think we need to move on to Natalie, but really thank you for asking these questions and thank you to Rob for again, great talk being on time and entertaining these questions and now we move on to Natalie who's going to tell us about northern corn leaf blight over to you, Natalie. Thank you. All right, let me just share my screen. Now we can see it. It's shared. There we go. Oh yeah, there it is. Okay, you should be able to see my slides and hopefully everyone can hear me. Great. So I want to thank the E life committee again for giving us all the opportunity to present our talks in this very interesting time, really generous of them to offer this to us. So I'm Natalie Clark, I've been a postdoc and Dr Justin Wally's lab at Iowa State University for a little bit over a year now. And I'm really excited to share with you my work today using a systems biology approach to investigate the genes that mediate northern corn leaf blight or NLB resistance in masts. So NLB is a major disease in corn is leading up to a loss of around $7 billion per year in North America. It's caused by a necrotropic fungal pathogen and here on the right I'm showing you leaves of corn that have been infected with NLB. And this kind of necrosis tends to happen within a week of infection so the progression is very fast. So given the debilitation of this disease on corn we're really interested in uncovering novel genes that are involved in LB resistance so that we can increase corn resistance to this disease and improve food and crop growth. So in order to do this we're using two complementary approaches. The first is a quantitative genetics approach in collaboration with Nick Latter at Iowa State University, his former PhD student Mercy and his technician Miriam. So Nick has a collection of 134 intermated B73, Mo17, recombinant, and bred double haploid lines. So I am not a quantitative geneticist so when I joined this lab I was like what are these. So here's just a simple schematic of what these lines actually are and why they're so useful. So these lines were generated by breeding the B73 and Mo17 parents they were crossed to form an F1 generation. Then these F1s were randomly intermated 10 times to form a genetically diverse population. So you'll notice that while these lines are genetically diverse they're not homozygous and if you use traditional breeding getting homozygous you can take a really long time. So what I think is really clever that they did is they actually use double haploidings they could get homozygous individuals weight faster. So this is a genetically diverse panel of corn that we can then use to investigate how genetic diversity impacts LB. So what we can do is kind of obviously is looking at disease phenotypes. So here are data showing that the net diversity of these 134 lines using disease severity measurements that were taken every week for six weeks and you can see we have a lot of variation. We have some lines that are very resistant to NLB and some lines that are very susceptible to NLB. So if we're looking at some quantitative trait loci or QTL mapping on this population, we should see some genetic intervals that could be conferring resistance or susceptibility to NLB. In addition to this we were also interested in looking at the molecular QTL. So by molecular I mean things like the transcriptome, proteome, phosphoproteome as I'll talk about. And the concept of this is very similar to disease QTL. So with we have genotype information so we can tell which part of the genome comes from B73 or Mo17 at each of these lines. And then we can try to correlate that with a change in this molecular level like transcript or protein. So in Justin's lab we have really been working to introduce this integrative omics approach when looking at molecular data. So a lot of people have probably done some sort of transcript analysis like RNA seek to look at transcript levels or try to seek to look at finding. But if we all remember our central dogma of biology, it's important that this transcript essentially is going to get translated into protein and that protein can then be post translationally modified through mechanisms such as phosphorylation. And it's been shown that in Mays if you look at the transcript and the protein levels for the same genes are only modestly correlated. So if we were to only look at the transcript level we would not get a complete picture of all the molecular changes that are happening in response to NLB. So what we decided to do was we decided to look at transcript data collected by QANSEQ and protein abundance and phosphorylation intensity data collected using MS. Now I don't have a lot of time to talk about today but we do have a pipeline in our lab that allows you to run many samples at once and get a lot of protein groups and phosphocytes out as I'll show you. So I greatly encourage you if you have questions on this to ask me at the end because I can give a lot more information. So using these methods we were able to detect around 35,000 transcripts, around 11,500 protein groups, and around 42,500 phosphocytes. And just to show the power of our proteomics pipeline, this is a comparison with some other large scale PQTL reports that happened to mostly in animals and you can see that we are detecting way more proteins than they do. So we're going to be able to get a much more complete picture of how the proteome landscape is changing in response to NLB. In addition so far there have been no published large scale phospho-QTL reports so we're really excited to see how this additional data adds to our understanding of NLB. So now on to the results. So first I just wanted to show a summary of the QTL we have identified from each data type. So as a reminder EQTL is transcript, PQTL is protein abundance and phospho-QTL is phosphocyte intensity. So you can see here that we have a moderate overlap in QTL, mostly between the transcript of the protein. So around a quarter of the protein QTL overlap with the transcript QTL. And based on some other studies this is a pretty good ballpark for overlap. So that gives us confidence that our analysis was done pretty well. I then further split these up by cis and trans QTL. So cis QTL are QTL that act locally in the range of the gene of interest, whereas trans QTL act distally. And again just as we would expect you can see that most of our overlap is happening in cis QTL. So based on what we know about cis and trans regulation this is a result that also makes sense. So these are just good quality checks that give us confidence that our QTL analysis was good. The next thing I did was I decided to overlap our molecular QTL with disease QTL. So in addition to the QTL I identified through my analysis I also incorporated two other sources of the NLB QTL. And you can see that around almost 50% of our molecular QTL overlap with at least one disease QTL. So this gives us confidence that the molecular changes we are seeing at the transcript protein and phospho protein level are regulating NLB resistance. So finally I'm going to kind of give you an idea of where the systems biology approach comes in, right, because up till this point has been very straight systems genetics. So as a systems biologist my main expertise is network inference. So I like to build these networks and try to predict what genes might be important for a function. So here I'm going to give an example on how I've done that with respects to a gene called benzozaxenoid biosynthesis 9 which I'm just going to call BX9 because that's an helpful. So BXs are metabolites that are involved in insect and microbial defense. With respect to NLB, HTN1, which is known to confer NLB resistance, it's been shown that their mutants have lower expressiveness of BXs. Additionally, knockdowns and knockouts of multiple members of the BX family have increased resistance to NLB. So we thought this would be a good candidate just to initially see how this network inference approach could help us identify nobilities. So first I'm going to show you the QTL that we found for BX9. So first I'm showing a zoom in of the transcript QTL. So you can see we found two QTL for BX9 on chromosome one. One is cis and one is trans. If we look at the PQTL, which was inferred using the protein abundance data, you can see that we also have that same cis and trans QTL. We have an additional trans QTL. So we have some overlapping QTL between the transcript and the protein. So given what we know about this we would hypothesize that the transcripts and the protein levels are probably correlated with BX9. So I just made a correlation plot. So on this plot we have transcripts on the X axis and protein abundance on the Y axis for BX9. And the colors are showing the genotype at the genetic marker closest to BX9. And interestingly you can see we actually have this negative correlation. So in the B73 we have higher transcript and lower protein and then opposite in the Mo17. So I haven't investigated this yet, but I think it's a very interesting result because normally we would expect these things to be positively correlated. So these are the kind of conclusions we can make just looking at the QTL data. So next I wanted to try and predict, since BX9 has an EQTL, maybe there's some transcription factors upstream of BX9 that might be regulating its expression levels. So to do this, I constructed a pipeline called Scion, which stands for spatial temporal clustering and inference of omics networks, is on GitHub, if you would like to check it out, it's an R-based pipeline. And how this pipeline works is we basically treat our QTLs as clusters. We infer a network for each QTL and then we just put them all together. So that's one novel aspect of Scion. The other novel aspect of Scion is that we can actually integrate different types of data. So for example we can look at if we use TF abundance as the protein level or the phosphorylation state and see how that changes the predictions of the network. So this is what happens if I build a protein abundance network for all of the proteins that we detect. And here I am coloring the genes based on what chromosome they're in. So you can see that all the chromosomes are kind of very nicely clustered and there's not a lot of cross communication. However, if you look at the phosphorylate network, totally different story. You can see a lot more cross communication between these genes. So we think that perhaps at the abundance level we're seeing more cis-regulation whereas at the phosphocyte level we might be seeing more of that trans-regulation. Further, if we extend this to BX9, we actually see in both networks we have different TFs predicted to regulate it. So in the abundance network we have this mid-family transcription factor, whereas in the phosphorylation network we have this homeobox and this BHLI transcription factor. And interestingly, these two transcription factors in the phosphorylation network, we did detect them at the abundance level. So it seems that their phosphorylation might be required in order to regulate downstream genes like BX9. So this is where we're going in the future with our project. So I hope I've been able to show you how we can use this kind of systems biology, quantitative genetics, integrative omics approach to address the question like NLB resistance. And right now I'm working on selecting transcription factors as well as kinases that may mediate NLB resistance and biologically validating that. And with that, I would just like to thank the Wally Lab, our collaborators, Nick Lauder, and Sean Christensen, who I was not able to mention, but we are working on the Tableau QTL data with him. And again, Eli, for giving me the opportunity to present my talk. And with that, I'm happy to take any questions. Thank you very much, Natalie. Great. So in issue, breeding for resistance is always yield drag. Has somebody done something similar to what you have done with yield so that you could ask whether the genes, the low side that you come up with, to make sure that they don't negatively impact yield? That's a really good question. I don't personally know because as I stated, I've only been working on this for a year. So my expertise with the field is kind of limited. But we could say maybe someone has performed a similar study on this that's been published. You could perhaps build complementary networks like you're saying, using yield as a trade, and then you could perhaps integrate the network. So perhaps, you know, you're looking for transcription factors that aren't as important in the yield network because that shows you that they don't impact yield. And then maybe that can help you create a subset. So that's one nice thing about this approach is that if you have other data sets of interest, you can integrate them and you can make more informed conclusions that way. Absolutely. Mm hmm. Cool. Great. Okay, so we have Frank Menke. Hi, Frank. Hi, Detlef and Hi, Natalie. Very interesting. Very nice talk. I really enjoyed it. I have a question. So I'm assuming you use label free quantitation to quantify both protein and fossil peptide levels. Correct. So you're actually using TMT labeling, and we're using 11 plex system. I think that was the other alternative I considered. But so, do you have any idea of what kind of dynamic range do you get? And are you able to detect proteins consistently at the lower end of the dynamic range? Yes, so before we did the QTL analysis, what I actually did was I, so we had 14 different TMT 11 plex runs. So each run contained 10 of the lines and a pulled reference. And what I did before doing the mapping was I only took the proteins that were detected in 50% of the runs. And that still ended up being a majority of proteins. So we do see pretty consistently that you can still detect the proteins. The main issue that we see with TMT relative to label free is the ratio compression. So if you look at the full change of proteins, they tend to be a lot closer to one that you would expect biologically like from an RNA seek experiment. However, some good news on that front is that they just released a TMT pro, which is a 16 plex. We've used it in a couple experiments of the lab and the ratio compression is much improved in that system. So I think that's something that they're actively working on to improve. The follow up question is, were you able to see proteins or phospho proteins on the peptide phospho peptide level that you can associate with either pattern induced immunity or even something like a factor triggered immunity in these samples. I haven't fully done that analysis yet. But it's something that we could definitely look into and see if we see any trends with that. Absolutely. Right. Thank you very much. Thank you. Thank you, Frank. And we have a question from Hossain. Hossain Fazolinia. Are you still there, Hossain? Otherwise I can just ask it. So, so Hossain is asking in your Venn plots, were these data on individual phosphorylation sites or was it on entire phosphorylated protein since a single protein can have multiple phosphorylation sites. Our phosphorylation data is site-specific. So in these phosphorylation networks what you're seeing is basically each circle represents an individual phosphocyte. So if you have a protein that has multiple phosphorylation sites, the network will use those individual intensities separately. So sometimes you see if the phosphocytes are very close, they tend to co-regulate the same things. But if you have phosphocytes on different regions of the protein, they could actually appear in totally different modules of the network. But yes, our data are on a site-based level. Cool. Good. All right. I have one last question here. Where did I make a note to myself? Sorry. Now I confuse myself. I scroll down too far. So it seems that you have a lot more power to detect protein groups than what people have been able to do before. What's your secret sauce, Natalie? So what I'll do is, in the Google note doc, I'll actually post the paper. So this main method of doing this TMT 11 flux and getting all these protein groups is really spearheaded by Song, who is the other postdoc in Justin's lab. He's done a lot of work optimizing our mass spec system and optimizing the sample prep so that we can get that large amount of protein groups. And he has published a very nice methods paper with really great step-by-step instructions so that if you want to try similar processes in your lab, you can. So what I will do is I will post that paper to the Google doc so that other people can find that protocol. But that is why it is all the magic of Song, as Justin would say. Wow. Nice. Okay. Excellent. Well, thank you again, Natalie. Awesome. Really, really like that. And now finally, we have Christian, who is going to tell us where we stay with proteomics. And we'll hear a bit about SARS-CoV-2, very, very timely. Christian, take it away. And Naomi or somebody, can you just unmute Christian? Thank you. So thank you very much to Eli for organizing this and thank you all for joining me here today. And yes, I will give you a bit of our new insight into COVID-19 or SARS-CoV-2. And really for me, where this is all starting and coming from, is that we're interested in cells and stress responses in cells. And this can really be in different areas, whether it's stresses in cellular compartments in the cytosol. We also care a lot about protein aggregation in the context of neurodegenerative diseases, but stress can also happen from externally. And I think really the important part for us is what we want to do is to get kind of systems-wide, cell-wide overview of what's going on and to understand the dynamics. And in the case of stress, what that usually means is that there are signaling responses that have effects on transcription and translation and modulate a lot of cellular effects such as metabolism and cell cycle. And one of the things that can cause stress from externally is an infection can be a pathogen just like viruses. And with several projects going on on this, and with really a main focus on looking at translation that has been our main interest recently. And while we've been working on this, and this year happened. So on February 1, there was an evacuation flight from Wuhan to Frankfurt. This is where I'm located. And out of those 150 roughly people who got evacuated, two turned out to have SARS coronavirus too. And I think what's interesting from a science side from us till then, but it was also meant that in two months now we haven't really slept because we tried to kind of transfer our systems to that and see what's going on. I just need to take a very short detour and one of these methods that we have recently established that we're working on. And this is something we call MEPROT. And the idea there really is, and this is all spearheaded by a really brilliant PhD student in my lab, Kevin Klang, and there we wanted to use proteomics to look at translation. And essentially we used very traditional approaches using pulse xylec to label eulosynthesized proteins, and you can then look at the incorporation over time. The problem is there if you want to look at dynamic effects, it has to be a short labeling that you really know what's acutely happening. And if you do this kind of labeling for let's say two hours, it's only a very small fraction of your proteins that will be labeled, and they will be lost in the soup of all the old proteins. So this is kind of shown here, this would be your signal for this eulosynthesized proteins and very often that ends up being below some detection limits. And this technology what it really was is we call it multiplexed enhanced protein dynamics proteomics, where we add a booster signal that essentially boosts the signal up to an area where we can determine it. And just as in the previous talk, this is largely based on adding a team T to that, so it's multiplexed, which means that in this one sample now we can quantify from nine individual samples what their translation rates relatively were at that time. And that works pretty well. In cells, here's an example from Hila cells, from as little as 15 minutes labeling to get a very robust and accurate quantification and it's linear, depending on how long you label for. And the great advantages of proteomics they really are it's relatively cheap and the throughput can be high so we can do experiments like this, where we look at 27 individual samples look at translation rates for more than 5,000 potency or to analyze that. So we took this technology and brought it now to a SAS coronavirus too. And this is really been a fantastic collaboration. She with these guys here it's a interest in adults lab at the biology department here in Frankfurt at the University Clinics Denise and his lab and also Benjamin Koch, and really what happened in February 1 they got the virus and I think we were very lucky and that the interest had already been working on SARS 15 years ago. And he was one of the first to manage to set up a cell system for the infection, so that you can add virus to cells, look at what happens in the host cell and amplify the virus to keep on doing experiments and he transferred that to the coronavirus to now we managed to have, as you can see here an infection system or mock, mock infected cells would look like this, and upon infection 24 hours later you have a huge amount of cell death and release of the virus so this is a functional infection system what we use for that is cacotu cells so these are epithelial colon carcinoma cells, which tried a lot of different cell lines it's not that easy with viruses because cells really need to be permissive. I think they everywhere lucky because these cacotu cells really work very well for SARS back then already. I think with that we quickly got a very robust system to work with that's now allowed us to look at the host cell responses and more kind of quality control here in fact these cells with the virus to wash us and monitor the supernatant of the two six 10 and 24 hours can see that we do get an accumulation of virus in the supernatant over time so the system is working. And then we apply this technologies that I mentioned before especially this meepot proteomics to infect cells and then at two, six, 10 and 24 hours post infection we measure the translatome in the cells and the proteome to really get kind of a temporal resolved idea of what happens in the host cell after infection. So we have the proteomics with it and then leading to the analysis. I just want to mention this point. All this data is now available this preprint so I can go and check it out. I'm also put a section on my website where you can find more information be for example, there's a link to all the wall files which are deposited on pride, so all the proteomics data is there. And we also made an interactive tool where you can go on there type in the name of the gene protein you're interested in and see whether we detect it and whether it changed. So with that, now we get to the data and this is now the translation data this is for this five viral proteins that we detected. And I think that was encouraging over time we see that these viral proteins are being produced. I think one thing also to point out is that here this this is not an accumulation of the protein it's really approaching translation rate over time so this still increasing means that the amount that's being produced at any given time is actually increasing over time so there must be additional things going on in the cells that don't just produce a concentrate but increase it. And then I think it gets quite surprising for us when we looked at the global host cell translation over time and this is shown here it's just the and the averages medians. And I think you might be able to appreciate it doesn't change that much and this is not really what we were expecting. I mean, here after 10 hours you have about 23% decrease. It's a little bit unfortunate and that from SARS and a lot of the other coronaviruses we don't actually have with information what effect the real virus infection has on the cell. But from the data we had that at least hinted in that direction that there should be a stronger effect on the host cell translation. So we're wondering where that comes from what that means for the entire viral response and so on but I think once we managed to look at the host cell proteins and the rate at which they're being translated we found what we thought probably some of the answers that you can see over here this is the there were a lot of things in there from ribosome translation elongation translation in the Asian and really kind of hint towards at this after viral infection there's actually more of the translation machinery produced to overall and keep up with the translation. So with everyone wondering whether this is something that is important for those cells whether if we target translation and also this increase in the translation machinery, whether with that we could prevent viral replication in the cells and that didn't come out of the blue. So they had been data in the past for other coronaviruses for different drugs that inhibits translation that they have beneficial effects towards inhibiting the replication. And up here is our data for two compounds like hexamide and emitting you may know it's very commonly used to proc translation. In case of emitting specifically also the cytosol and what we show here is inhibition of viral replication. So we will monitor to what extent we still have this viral replication here. So those responses you can see here that's with quite decent IC 50 so they're both my low micro molar, we get an inhibition of viral replication in the cell. So it was very nice, had some tools in our hands, then moved on to other data that we had to hear more specifically and changes in protein abundance upon infection and that's here on the right you can see again this many later time points that things start changing, especially after 24 hours. We won't be cluster of things going down. And there were a lot of things in cholesterol biosynthesis and so on and therefore it was not as interesting as cluster two, and I will point out two specific things there. And here again is an enrichment of what we found and for me it's two things sticking out and changes in the spliceosome, but also changes in carbon metabolism glycolysis TCA cycle and so on. And I will start on the spliceosome aspect of that because Joseph was really knew there hadn't been many links between viral infection and spliceosome and I think it's still a little bit of an undiscovered part of what happens after infection. However, there is actually data in there and we particularly found quite a bit of data from looking at interactions of overexpressed viral proteins in the cells what they actually interact with in the cell. And there were a lot of interactions with spliceosome components pointing towards a possible function of these viral proteins in modulating the activity of the spliceosome in the cell. So, who went back to our viral replication essays and used an inhibitor for the spliceosome for the analyte B, which is an inhibitor of the protein SF3B1. And here you can see that really very low nanomolar concentrations we prevent viral replications in the cells. And then I mentioned carbon metabolism before here it's a focus on glycolysis. We tested also their one inhibitor which is 2-deoxy glucose, many people use it in the lab it's inhibited sex of kinase and prevents glycolysis. And there too it's non-toxic concentrations even though they're higher, we could prevent viral replications in the cells and that was nice because for another virus, rhinovirus, they had been studies that showed that also their 2-deoxy glucose can prevent viral replication and they also showed that in animals. And I think for me what's rewarding as a basic scientist is you can do something that gets picked up and then taken to the next step. And just yesterday there were these news that the company molecularness now they have compounds that are very similar to 2-deoxy glucose and essentially enhance it in some ways that they're taking this now and trying to get that into the clinics against COVID-19 treatment. So summary, we set up an in vitro cellular system for viral replication to study. They carried out translatome and proteome analysis to define what is the host cell response to infection, identify key cellular pathways modulated upon infection, and ultimately tested inhibitors of these pathways, translation, splicing, glycolysis, nucleotide biosynthesis, and they all prevented SARS coronavirus to replication in cells. And again, down here is the link to the preprint. And with that, I would really like to thank my lab. This is a picture from happier days where we could all still meet outside and have fun. A special thanks to Kevin Glenn, who did the work on my lab side of it. I mentioned our collaboration of the interest in ESA, Benjamin and Sandra, the head of the biology institute. Also want to point out even Dickage was head of my department and has been a great mentor and thanks to different funding sources that made all this work possible. And thank you very much for being here this time. Great. Thank you. Thank you, Christian. I'd like to take the privilege as the first question. Yes, have you or has anybody else done something similar with either the original SARS, I guess, coronavirus one virus or any other coronavirus. I mean, in a similar system, no, I think mainly what has been done and that's what we know for SARS is that people have overexpressed this viral effect of proteins inside the cell and then monitored with 35 S methionine and what happens to global translation and they saw really severe reductions close to 100% in some cases. The question is how whether it refers to what happens with a real viral infection and then the simple answer is we don't know yet. This is something we're interested in and want to find out, but I think I mean one thing I mean it's a real technical problem in the past. What they do is write this on profiling and it's very expensive and you need a lot of material. That's not something you want to do with the kinetics. And you would require huge amounts of virus of course this is all done in a BSL free lab it's not so straightforward now we have a way with proteomics to do that. And I think it would be interested, interesting indeed to see what's going on in these other viruses because maybe we can also go the other way what we find now that works for SARS coronavirus to might also be let's call them older coronaviruses and for us of course the question is also that could be your source coronavirus three in the future or other similar one and I think now is the time really to prepare and understand as much as we can. So is it known which actually the so a what are the target cells in in the patients and then what is actually the best in vitro system to mimic those most important cells. Yeah, I mean, best known that's whatever discusses of course along and I think that's mainly because the patients needs needs to go to the ICU and get air from there. But it's really not just that there's a lot of reports there's a lot of people who have inflammation in the heart for example there's a lot of reports kidney failure stays a lot going on in the colon so I think that's we're actually not off. Yeah, but I think that's something that needs to be studied in more detail. I mean for us, of course, we're trying to also get to other cell lines is not that easy. Okay, cool. Burak Burak Tepin. All right, hi Burak. Hi. My question is like more generally about the the pathways that you guys are targeting. It seems to be very broad like pathways that are like important for cellular processes. Of course it's going to inhibit viral replication. My question is like, is it safe to the host cell that you inhibit these processes or like is there anything specific to SARS code to virus. Yeah, I mean, this is always a little bit the question of the context of your cells like all the IC 50s I showed they're non toxic for those cells. Of course the question is how does another cells in the body. Most of these things are not in the clinic some some have been tested and I think this is something we need to extend on. But nevertheless, viruses are different and even though these are brought and brought they go on cellular pathways they don't prevent replication of all the viruses out there it's not as simple as that but I think, nevertheless, you're you're right on one of the important questions here. Thank you. Frank Frank man key. Hi, Michael. Great talk very timely research. Amazing work. I was wondering if you infect cells in the presence of the inhibitors that you identify that result in lower infection rates. Do you then see a change in transcription of those genes that normally go up with the infection. And are there other genes that coming up or have you been able to try that yet. The truth is, we haven't done it yet. I mean there's something we want to do but I mean, okay we're two months in now but it's really still all very much sort of kind of get started on. I think it's important and I think we need to see what's going on there. I still think we also need more specific inhibitors here or there right to see whether there is something that we can transfer more easily from from the clinics. And there's a lot of more tests necessary and especially these two deoxy glucose, for example now, and it would be important to understand what's going on there and how good our clinical angle here really is. Great. And another question is. So you're one of the first to look at the proteomic level at SARS, this virus sorry. Have you tried to look at patient material and see if you can detect the virus to what level you're able to detect it as compared to all the other proteins, the human proteins that are there. So we haven't done it yet. There was very recently a preprint that came out taking from the blood purifying cells from patient had been infected. And of course, in many ways it's easier to have a cell system as we do to get a more broad overview and we found a couple of things also towards interference response and so on. What was striking is that splicing and the splices so was also one of their top hits actually their geo it was the number one thing that was written down so I think that kind of nicely confirms each other. But I mean there's still a lot more biology we need to understand about the infection processes and the wholesale response to that. So that actually meant maybe I did phrase the question properly but would you be able to adjust the settings, the proteomics methodology to actually detect the virus in human samples. I think it would be possible. It's the question how what's the percentage of whatever material you get that is actually infected but we don't see all the viral proteins it's close to 30. But we see a good number and I think now with our data and also some other people's data we know now which peptides we can detect particularly well. So I think there would be this option also to go by targeted approaches to specifically look for specific viral peptides. And have the samples for that in a moment to try it out. Sorry, would you be able to share those peptides we're looking into actually assisting using proteomics and we're doing a lot of targeted proteomics and that would be great. And as I said really that was important to us. All the data in this paper is available if you go my website there all the raw files there's other preprints that came out already people who reanalyze the data. It's, it's there, you can also more specifically look out which proteins it was but I think for you use it's actually better if you want to set up targeted as is to go on the raw files right where you can see. And then the spectral behavior and MS to and so on it's, it's all there and if you have problems getting to it just let us know. Great. I think this is actually one of the great things that we are seeing with the current crisis that they really push for, you know, preprint, and so on. And hopefully many of you have seen this at elive you know we always had this is also also of course sharing open access and so on, but also how we look at papers that we look at at the papers in front of us and don't imagine a paper with all kinds of bells and whistles that are not there. So for us actually this crisis has been a reminder what our original exhaust is and we published a little thing and maybe some of you have already seen this and I've been really pleased to see the editors they've all taken this up so when we look at the paper we know people. You know very few people can currently do experiments and then if at all possible, if we essentially think a few more experiments would be, you know, nice to have but they are not dramatically going to change things. Tell the authors look reword perhaps some of the few things right in there to fully support this, this or that experiment would still be nice to have but go ahead and accept the paper or invite revision. We are only, you know, we're no new experiments are done so that's hopefully another thing that's going to stay with us after this crisis so so thanks to all three of you thanks Rob thanks and not only thanks Christian really you know enjoyed this tremendously thank you to the audience thank you to all of you with questions. This series is going to continue next Tuesday on April 14 at nine o'clock British summertime starting with Anand Salah Lakshmi from University of Bristol will talk and we'll talk about the role of Roger and private necked introvert genesis during and genesis. Then Stephanie Machoni from Louvain will talk about categorical representation from sound and sight and the vertical occipital temporal cortex of sighted and blind and then Yuri, who from Chapman University will talk about neuroscience of recognition for decisions that matter and this is going to be chaired by my colleague Tim Barron's also deputy editor from the University of Oxford. So for up dates follow us on Twitter at elive community and you'll also see the website they are so just Google elive webinar series that's that's easy to do. And so yeah really thank you again all of you stay online to chat I think this is going to be open for another 20 minutes or so so we can chat directly with the speakers so I already had a question to Rob because my question to him was a bit confusing. And if you'd like to stay please do so, and everyone else. Thanks for joining us. You're welcome to leave and we'll be ending the live stream now. Thanks. Alright, so I'll ask my question to Rob then can ask that actually to everybody. I just want to say if it's okay but actually going to continue the live stream on YouTube, because this is really interesting I think for the viewers over there. Is that okay. Cool yeah, sure. Totally. Thank you. I think my question I did I realized afterwards it didn't come quite correctly across. I was wondering these, these repeats. How much do they vary between different isolates so if basically so what I'm thinking about you know this this idea, can you be prepared for the future so if you know this can turn out to be favorable and certain environment, then you know maybe a little bit different to having these repeats and specific places in the genome so in and so I was, you know, I jumped forward from the isolates to the in vitro selection experiment that you had done and that's what I had in mind. So that's something that we're really keenly interested in is, so we're seeing these recurrent CNVs that are amplifying these genes that are important for antifungal drug resistance antifungal drug tolerance. So one of my hypotheses is that because we're not actually selecting on the repeats themselves we're selecting on the sequence between them that if we were to say stress the cells for hypoxia maybe we would amplify hip, or something like that if that located between repeats. So yeah that's something that we're actively looking at, looking into just because it does seem a way that this non myotic organism could rapidly allow for adaptation of its genome. So yeah, we do think that that could be that is a very strong hypothesis that we have that we would like to test. I'm now diversity between different isolates the data that I showed you today was all from divergent genetic background so I sense that we're diverged by 3.3% of their genome so nearly 100,000 bases different. So they do seem to be pretty well conserved throughout the Canada species. Cool, cool. And Natalie, I also wanted to ask you another question. This, the chemical that you mentioned that Ben so sorry, I don't know. So is that only effective against northern corn leaf blight or is that also effective against a bunch of other parasites and pathogens. So is that only effective against other diseases to I haven't looked super into it because I was mostly just seeing what if they have been linked to NLB at all but it is well known to have a role in immune response. So I would assume that it's also affecting other diseases and that brings up another interesting point. Yeah, for example, other sites have an NLB QTL mapping and maybe some related to seasons like southern leaf blight or leaf rot there tends actually to be a lot of overlapping QTL. So it's likely some of the genes that we find will be responsible for not only NLB resistance but other disease resistance. Yeah, that's one of the, one of the big challenges breeding how do you breed for as many resistance as possible but you don't kill the yield of the plants. And that's, that's, that's one of the things that I, you know, as an evolutionary ball and just think, think about a lot because that applies to wild plants as, as well how do they square the circle be as resistant as possible without, you know, giving up fitness in terms of making seeds. So, Christian coming back to this question that was asked earlier this so you said that you apply these drugs at a level that actually doesn't has no strong effects on the proliferation of the of the cells so that's really quite quite remarkable. And it's, it's essential. I mean, if you do call the CPA essays where you look at psychopathic factor of the compounds that you compare at the same time whether the drug itself is toxic to those cells. At the same time, though, very often the viral infection changes the cells to an extent that they might get sometimes less sensitive to the drug in terms of the cell itself, but it can prevent the virus. Yeah, very cool. So, so, so, so you weren't working on on Corona viruses before right did I get that right. Not on Corona viruses. So I think for us, I mean, the interest of my lab is those stress responses and years have been doing a lot of proteomics is generally speaking I think what I care about is these dynamic things because, particularly, I mean, one of my main focuses is unfairly protein responses, right and I think the one thing we've learned is that those start, happen and finish within less than 12 hours so we really have to come up with more sensitive ways in the future to get this temporal profile sorted. And I think what we've learned now that we have these tools to look at translation it actually opens the door for a lot of new things and more specifically of course if you want to have a response on proteins you need to look at proteins. But the time it takes and the amount of change to really be able to to measure protein abundance differences. This is quite massive. And this doesn't have the time scale we need. But translation, this is super quick it can change very, very acutely. And now we have these tools where in the short times we can label we can now actually determine the cell that changes way before we can see them on a protein level. What is the amount of material you need for for your methods. It depends. I mean, most of the experiments. I mean, it's a bit different in our cell. Yeah, in my lab if you just do a regular experiments but really not limited at all we usually work with about 100,000 cells. Okay. This is mammalian cells in the tissue culture healer or hex cells. It's still not that much 100,000 cells. We can work with a 12 well blade or six well plates and I think this is really what was also very important here because it takes some time to get to a point to create enough virus. And especially also if you want to move into more primary cell system that are usually very limited in cell numbers you can get right so you have to bring it down. We did ribosome profiling and I think it still addresses a different question as an absolutely important technique. But in our hands, one of the main issues is we have to start with 15 centimeter dishes and usually several ones of them to get enough material. And for us as we don't do it that often we have to use kits I mean those 10 samples cost us for 5000 euros and for the omics very 10th of that. So it's a young lab. That's really something to take into consideration right. Yeah. How does that that sense and not only how much do you use for your proteomics. You mean from like, how much materials. Well, normally we start with like 100 milligrams of tissue that's been highly optimized. Yeah. And for some organisms, that's looking at like Arabidopsis or corn for some organisms you might need more like we've been collaborating with Stacy Harmer on sunflower and I know for that experiment. So it's a low amount of tissue they have from the sunflower because they only cut like a very specific part of the stem. It's been a little more challenging to get enough material. Now when we is telling me I'm talking too much so. Specific. I was wondering whether you had any questions for each other. I thought all your talks were like, fantastic and I'm just curious if you're interested in what each other said. I mean, to some extent Natalie may ask because I know I think QTL studies generally are very interesting and there's, there's much more differentiation and just cis and trans QTLs right and and I think, especially with the depth that you have you can get a lot of kind of secondary regulation probably out of your data and I'm just wondering also but it's really interesting tool that you showed us and to what extent you can also get to those effects and distinguish from your data. Yeah, absolutely. So you could for example, you know, follow down the line of the network grade because you probably saw even though we restricted the network to inferring for each QTL a lot of QTLs are actually influenced each other so you have this very big network where things basically cross communicating so you could basically, you know, start from the top and see how that signal kind of travels throughout the network and how many like you said secondary or tertiary even later treats. It's ready later. Definitely. Very nice. And for me, if I may, the other question. I know that's always a little bit to the tricky part with these QTL studies is how big the sequencing on sequences in your genome, how they get shuffled around so how well you can position back the effects on your genome in the genes and they are done on the plants so well. How many genes do you vary to usually have in your bullseye. How well can you tell without additional work. What was actually the important gene. Yeah, so what I've seen is, it doesn't just differ by species but it also differs by the data type of you so when we look at the transcript and I think part of this is because of the TNT labeling so I mentioned there's this kind of ratio compression ratio compression you're going to have less variance between your samples right so that basically decreases the overall power of your analysis. So on the transcript level are low size going to be much smaller. Some can have as few as like 10 genes. Some of the larger ones might have 100 200 on the protein level the intervals get much larger so like the smaller intervals might have say 25 to 50. You have a really big interval could be like 300 400. So Natalie was there so the eqtl and so you have cis and trans qtl and most of the qtl assist qtl and they are you would normally get to the gene because it's a regulatory variation in the gene and that's very different for for proteome. So they are you normally don't get it so so there is much more trans variation than than cis variation. So I think if you look at the at the numbers we still have a slight majority of cis variation in the pqtl but you're correct that we have much more trans than pretty eqtl and same for the possibility as well. Yeah, and then that's the general I should know this I don't sorry, but that's the general thing with proteome versus transcript variation that proteome, you have more trans versus cis. Okay, yes, that's at least been shown with the animal pqtl studies. Okay, cool. Hmm. So, so question, question for Rob, since you are working with, since you're working with a human pathogen. Is that, is that an S2 classified pathogen. So generally speaking, no, however we do treat it like it is. So, so what I'm getting at. So many of us are thinking about how can we contribute to, you know, practical stuff with coronavirus like and then one issue is of course whether you know you are set up for BSL two or not so the extraction. So in theory, I actually don't know about the new set up since we changed universities, but I know previously, we did actually apply for the BSL to level. So those are still also only classified as BSL one. However, we do think they're, they will be shifted to be a cell to here in the future. So Christian, your work is all BSL two or It's all BSL three. Yes, BSL three event. Yes. So it's it's isolates from those patients. And yeah, I think so that it's three because you're amplifying the virus or Yeah, and I think also I mean, I mean, I'm close to Frankfurt Airport, right. So the good and bad thing is we get things first where you usually because things are right by the airport and I think to quite some extent it wasn't also so clear what the regulations would be. And I think it was also just to be on the safe side right because we didn't know how bad the first would be and I mean luckily that is all them biology here in the Institute with the English to nettle and Sunday to say they have a BSL three lab. Yeah. And that's where we do things for us. I mean, in my lab, once we have a life set it's, it's zero right. It's good. Like the biology so yeah cool. The rest is tedious right if you work in them. These are free lab. Cool. Good. All right. Well, thank you very, very much again. I think all of us have earned now that weekend is almost there. So hopefully, except for the except Christian, Christian said it. No weekends anymore. The rest of us. Well, thank you very much again, the three of you. So and thank you to the elive team. They seem to be still online. Thank you Naomi. I'm Miranda Daniel. Thank you them as well. Thank you all for your talks. They were wonderful. I was listening intently. All right. Thank you. Thank you. Great sharing. Yeah, it definitely was easier the second time around. I was not nearly as stress and I do know about this kind of stuff a little bit more than the systems neurobiology that I had to chair the last time. So you'll be excited to know we plan to continue and with some research themed ones. So get ready for the invite. Thank you. Thank you.