10x-pert Workshop | How to Single Nuclei RNA-seq

10x-pert Workshop | How to Single Nuclei RNA-seq



hello everyone and welcome to our 10 excellent workshop for single nuclei RNA seek thanks for joining us I'm Shana the tennis community manager and your moderator for today before we get started let's go over some quick housekeeping items our 10 experts will introduce in a minute we'll be giving a presentation followed by an open Q&A session we're recording this session we'll make the recording available on the tennis community shortly please note that all attendees are me and if you'd like to ask a question please select the Q&A option on the bottom of your screen selecting the Q&A option will open the Q&A box where you can type your questions please note that you can ask your question then automatically by checking the send anonymously option and finally if you're interested in tweeting during the session please use the hashtag 10 expert okay now let's meet your 10 experts for today up first is Sharmila Chatterjee Sharma is a scientist at 10x and is currently a member of the sample prep group where she is dedicated to developing new demonstrated protocols for the chromium single cell 3 prime solution she held the PhD in physiology and has worked on different aspects of cell biology including using the mouse as a model system to study developmental biology a fun fact about Sharmila is that she developed the nuclei isolation protocol that we'll be discussing today joining sure Miller will be polymer skin causes staff scientist at 10x and leads the development of 10x of single-cell RNA analysis package cell Ranger he has a page gene computational biology and previously worked on RNA biology research in real analysis software for non-invasive prenatal testing and liquid biopsy tests the funny fact about Paul is that he was an intern at a nuclear power plant and now I'm going to hand it over to Sharmila so she can get started with her presentation good morning everybody thank you for joining us this morning the agenda for this morning is to go over why we're interested in single nuclei RNA seek I'll go over an overview of the nuclei isolation protocol and the best practices that we recommend when following this protocol overview of the single nuclear RNA seek analysis with using the loop cell browser demo and a new key idea set and finally we'll leave plenty of time for some Q&A so please use the function to submit your questions so why nuclei using nuclei allows us to sell Atlas without really using actual cells there are certain cell types that are fragile and are difficult to work with and when I see difficult it means these cells are difficult to bring into a clean single cell suspension in such instances using the nuclear protocol might be a way to analyze or study these samples it also expands applications of droplet base single-cell genomics it's a possible solution for archived samples which tend to be damaged also tissue samples that are difficult to dissociate tissue dissociations requires a lot of optimization and each tissue is unique such that there has to be different dissociation protocols for them for all of these instances by deleting nuclei and running them through our connect a 10x chromium single cell three prime reaches solution is an option it's also useful to profile complex tissues and organs where nuclei can be isolated but not whole intact cells and so in in cases where tissue dissociation eals cells that are damaged or a sample that results in a lot of debris and ambient RNA it would be better to use the nuclei relation protocol and get a cleaner better quality sample to get the high quality data that the user expects it's also a future potential for clinical samples and it can also reveal molecular genetic regulatory mechanisms that are specific to the nucleus which can which cannot be obtained from using whole cells now to run single nuclei we have not made any changes in our system we are still using our existing chromium instrument and the existing single-cell reagent kit have internally seen that using the chromium solution enabled scaling of nuclei to ten thousand nuclei per sample and it also removes the need to offset optimize the drop seek protocol the protocol that we have the demonstrated protocol that we have generated is quite specific for our system to generate a clean good-quality sample that is compatible with our single cell three file solution it also helps to resolve biological samples these nuclei samples are equivalent to but slightly different than single-cell when we compare single-cell RNA seek data to single nuclear RNA seek data we see similar cellular heterogeneity however we do observe more intronic and antisense RNA mapping and also there is limited nuclear polyadenylation which means in some samples we might see fewer genes detected and this is ok this does not mean the sample is bad or it is or the data is not useful if in some cases where either leading cells might be difficult and it becomes impossible to get a sample that is clean and debris free going through this nuclear isolation route might be the way let's go over the key points or the major steps on how to isolate nuclei and what are our best practices that we recommend when using this protocol so we have developed a series of three protocols for isolating nuclei and these three protocols are from three different sample types we've used cultured cells fresh and brown embryonic mouths neural tissue and fresh adult mouth neural tissue we have tested your cats which is a human cell lines it's a suspension cell line we have tested two nine three T's which is human adherence align NIH 3d trees that a mouse adherence align normal PBMCs human cells and we've tested the embryonic ei teen Mouse combined cortex hippocampus and ventricular zone as a tissue type and also the adult Mouse combined cortex hippocampus and vegetarian if you have seen the protocol you must have noticed that the lysis protocol for all of these samples that we have put up in our demonstrated protocol is similar with minor variations in them and we expect that the lysis protocol will be compatible with many but not all cells or tissue types all cells and tissues are unique and they all require their own individual optimization and so even in this instance we expect that additional optimization may be required for the preparation of the cell or tissue types that are particularly sensitive to suspension composition also based on handling techniques or the way the tissue has been previously stored of cryopreserved further optimization might be required so what are the important considerations that we need to keep in mind when using the nuclei isolation protocol we have suggested a few best practices in our demonstrated protocol and these are best practices are in respect to cell lysis washing debris removal counting and concentrating the nuclei from both single cell suspension as well as neural tissue in preparation for use in our 10x genomics single cell 3 prime v 2 solution the aim for these best practices is to minimize the presence of nuclear aggregates dead cells cellular debris cytoplasmic nucleic acids and potential inhibitors of reverse transcription all of these dead cells cellular debris and aggregates prevent the generation of a good quality sample and prevents good data being generated out of this panic single cell platform as I mentioned earlier all samples are different and each sample requires its own optimizations similarly the protocols described here are expected to be compatible with many but not all cell or tissue types an additional optimization may be required for the preparation of cell atisha types that are particularly sensitive also preparation of single cells or isolation of nuclei directly from solid tissues or cryopreserved samples will require additional optimization steps based on the tissue type as well as based on the method of tissue storage and preservation you I would like to now go over the major workflow steps that are involved in nuclear isolation from these different sample types there are three major workflow steps the first being the lysis and the nuclear isolation this for single cell suspensions requires around 35 minutes this will also this also includes washing steps post lysis however there's further additional steps that we recommend using after lysis such as the myelin removal as well as the density gradient centrifugation in case when we use samples such as the embryonic mouths brain tissue we used an additional step the myelin removal step the embryonic mouse brain has myelin and post slices the removal of this myelin helps to clean the sample up making debris free that generates higher quality data in the case of an adult mouse brain tissue we added another step and that was the density gradient centrifugation step and this helps to clean the samples up further all these steps require additional time myelin removal adds around 45 minutes to the entire protocol and adding the density gradient centrifugation add another 60 minutes when we dip the myelin removal step we used melt any biotechs myelin removal beads to remove myelin and we also used a sigma density gradient centrifugation we use Sigma reagents to perform the density gradient centrifugation though we did this on our tables of ultracentrifuge it can also be done on a swinging bucket representing speech so why do we suggest the incorporation of all of these additional steps post licensed we have consistently seen that additional cleaning using myelin removal and densities interrogation generates a higher quality sample that yields in better data as can be seen on this slide the first the top panel are images of nuclei isolated and we're seeing countless images showing as nuclear the first image shows a lot of debris and background and this can be matched with the barcode plot rank plot where we see that the knee is not properly developed and it's a sloped plant wherein the algorithm is unable to differentiate between single cells and the background when subsamples go through a myelin removal the sample starts to clean up a lot can be seen in these countless images also the barcode rank plot that has started to show single cell behavior where it's able to actually differentiate between single cells and the background performing density gradient centrifugation improves the quality of the sample even more it also shown in this barcode rank plot where there is a proper knee with very good sensation between single cells and the background and adding these steps also translates to the other metrics that we see when running the single cell 3 prime beta solution so when we look at fraction of reads themselves we do see that the fraction of reads improves when we have incorporated myelin removal as well as density gradient centrifugation so it improved from 45 percent to 70 percent with incorporating myelin removal as well as density gradient centrifugation on the other hand we do observe that there is a decrease in a reads map confidently to transcriptome as well as mapping to the sonic regions and this is expected with Nick Ferris samples but as long as we are able to detect the genes and form the cell clusters this results in good data the kind of data that users would like to see so when we take a look at Luke cell browser we see that we are able to detect all these different markers and cluster cells based on the based on these markers we can detect excitatory neurons inhibitory neurons astrocytes oligodendrocytes and radial glial cells and like I mentioned before all of these additional cleaning steps such as myelin removal and density gradient centrifugation helps to improve the quality of the sample that's been generated resulting in higher quality data so additional points to consider when using this nuclei isolation protocol are there are a lot of different places where optimization can take place one of the most important places to optimize is the lightest time it is very important to monitor the lysis efficiency to determine optimal conditions as you might have seen that in the nuclei demonstrated protocol that we have put out we have changed lysis times for different samples there is a different life this time for cell suspensions and a different time for isolating nuclei from tissues similarly all cells and tissues are different and they all might require optimization of the lysis time to get optimal quality good quality data we are monitoring lysis is important so that we avoid overloading the cells which will result in the nuclei lysing also we would like to avoid under lysing the cells where the sample now becomes a mix of nuclei and life so this will contaminate the chimp and also increase ambient RNA in the background also if the user sees and that the nuclei ten are sticky or aggregating they can increase visa concentration from 1% to 2% and this helps with preventing the nuclei from clumping together other cases where optimization might be required our centrifugation times and speed different cells are happier in different sensation speeds because the different cells have different let up they're different level of fragility also we do not want to have very high centrifugation speeds in time because this might result in damaging the nuclei and causing these nuclei to burst releasing the ambient RNA into the solution along with generating just debris also we do not want to have very low speeds such that the nuclei do not effectively pellet down and there is loss of nuclei when performing beneath their isolation protocols when working with a new sample type it is recommended that you save the supernatant and count the nuclei in the supernatant so as to gauge if there has been a significant loss while centrifuging the samples also but the user can optimize the density gradient centrifugation we do recommend using the density gradient centrifugation to clean up the sample we in our protocol have recommended the sucrose gradient for all using the tables of ultracentrifuge however based on volume and values availability one can also use the swinging swinging bucket rotor centrifuge one can also use the octave prep if comfortable with it and has already used it and both of these methods are used to clean up nucleus we do not recommend mechanical disruption of cellular membranes since this is difficult to control for time and pressure between users and also between cells and different cells and tissue types also we do not recommend reducing the concentration of BSA in the nuclear washing resuspension buffer as you might have noticed that for the nuclear Regulation protocol we recommend using 1% PSA which is higher than the 0.04 percent BSA that's recommended for the single celled protocol this high higher percentage of BSA is required to prevent nuclei from aggregating therefore it is also not recommended that you replace the nuclear wash and resuspension buffer with the PBS and 0.04 percent BSA temperature is also a key factor it is important to keep the nuclei cool so maintaining samples on eyes during lysis and also using centrifuges that have been previously cool down help prevent nuclei from bursting as well as aggregating also it is not recommended to start with an improperly stored tissue improperly stored tissues have already started going through some amount of decay and degradation and using such a sample as starting material to isolate nuclei does not result in very high quality sample also when performing nuclear isolation on mixed species we suggest that you do not mix the cells and then isolate the nuclei because we have seen a higher percentage of high percentage of nuclear at nuclei aggregating it is recommended that one isolates the nuclei from the different species and then mixes the nuclei to perform mixed species when comparing the gene expression between neuronal cells and nuclei isolated isolated from fresh and Brannock mouse brain tissue we see that they follow a similar pattern as can be seen here the top two panels are replicas of neuronal cells and the bottom two panels are replicas of nuclei isolated from embryonic mouse brain tissue and one can see that all the Cluff the clusters seem to be similar with the exception of a few clusters that might not be there in the nuclei isolated samples we also see by performing differential gene expression that are different markers can be seen when comparing nuclei and in brining brain cells and all the different clusters can be can be isolated such as the excitatory neurons the inhibitory neuron the astrocyte the oligodendrocyte and the radial glia with that I would like to point you all towards the support n a genomics comm website where all the resources for single cell sample prep is available the isolation of nuclei is also available over there we recommend using these guidelines since they have resulted in good quality sample and good quality data for us and we think following these guidelines along with optimizations done by the user will help yield very good data on the Penix from you single cell 3 prime v 2 solution and now i would like to hand over to Paul who's going to be talking about the nuclear RNA seek data analysis so in addition to shipping and instrument reagents and workflows we also ship software with our products and an example of some of our analysis offer can be find not found on our support website if you want details on the types of output files it produces you can go and click on these various links and all of this falls under the single cell gene expression section which includes the analysis of both nuclei and single cells in the same documents so the name of the software that we provide that actually does the analysis of the raw sequencing data is called cell Ranger and the goal of this pipeline is to take your raw sequencing reads and generate a cell by gene expression matrix which is a standard analysis format that's used for downstream analysis whether that be dimensionality reduction or Weir clustering or cluster specific expression these are the most common types of analyses that are done for single cell rna-seq or single cell nuclei seek datasets some important points about the sub Ranger analysis software is that not only does it provide outputs in standard formats it runs on Linux and it's it's free for download so you don't have to pay to actually use this and you can either run it on a single machines or high performance compute clusters depending on what those modes is more convenient for the user examples in addition to the the gene expression matrix that I just mentioned you also get a file called the cell browser file this goes along with another tool we provide called Luke's cell browser this this tool is not specific for nuclei analysis this can be used for either single cell or single nuclei analysis and essentially cell Ranger doesn't doesn't actually distinguish between the two types of analysis so loops all browser is available for Windows and Mac and you can download it again it's free for download at our website it allows you to do a lot of things with a single cell or single nucleotide gene expression data interactively so the main point here is that you don't actually need to use our Python to do this analysis this type of analysis is available to people who don't have a programming background some of the examples of the things you can do what loop cell browser are finding significant genes either that distinguish a cluster from the rest of the sample or looking for genes that are differentially expressed between two nearby clusters or sample conditions it also allows you to identify functional subgroups of existing clusters through gene expression tools and once you've done all that you can create save import and export custom clusters and custom gene lists for analysis and other tools for example if you want to them go into art or Python into your own analysis you can start with loop and then go into those tools later and just to reiterate this tool is free for download from our website so given that I just mentioned that neither cell Ranger nor loop cell browser distinguished between cell or nuclei data there's one change that you might want to make when you're doing a class specific analysis and this relates to the fact that the nucleus is enriched for on the spliced or pre mRNA transcripts so the example of going I'm going to show you here was actually run with a slightly different reference from our existing downloadable standards ass reference in this case this is a Kree mRNA BRR so what that means is essentially instead of only requiring reads to map to exon to be counted towards your my accounts we can now take use that line to introns and also count those to review my accounts and essentially what we're doing here is we're treating the entire transcript including introns as the target sequence for counting you mines and there's actually a one line is one kind of code required to convert a standard reference into this type of reference and if you need help with doing this you can contact our support team another important point about the analysis I'm about to show a movie cell browser is that the brain cell biomarkers were obtained from this single cell adult mouse brain paper tastic at all in nature neuroscience in 2016 so now I'm going to take us over to the actual loop cell browser app so here's the type of a screen you get when you open up loose loops I'll browse around a data set in this case the data set is 2,000 adult brain nuclei from a mouse aligned against again the the pre-mrna reference and not the standard reference that only uses exons and what you get when you open this up typically it'll be white and not black and this this black background is a new feature that was added in MOOCs all browser 2.0 and that's essentially there to increase contrast in cases where otherwise the colors might get washed out let's switch back to the high contrast dark background mode and zoom out to a better view of the entire cellular heterogeneity that's happening here so when you open up a loop cell browser by default what you get are the standard clustering analysis results that come out of the cell Ranger pipeline in this case you can see that it defaults to the graph page clustering method which in this case found eight distinct clusters in the MIS nuclei dataset so the next question you might ask is you know how can we interpret these clusters what how can we take prior data from the literature and map these clusters to what we know our G markers are ready so an example of one of the ways you would do this is you would go into this gene expression interface here and what I've done here is I've pre-populated this list with four broad cell types that are based on markers from the Tasik a tall paper let's just walk through some of these so the the biggest set of cells here are the glutamate or lumen – motor check cells which are also known as the excitatory neurons you can see that makes up the vast majority of the nuclei in this dataset a smaller set of cells here belong to the GABAergic neurons and these are also called inhibitory neurons in the literature now even smaller set of cells would be the microglia you can see you only need one marker in this case which is CT SS 2 actually very very clearly pull them out so I'm going to walk us through an example of what happens when you want to so given that you know we painted those cells with those markers what I've done here is I've essentially added them to a category and so instead of using the cell Ranger clusters we can now color the the nuclei with these custom gene lists that I've created and so you might see that there's a cluster out here that's completely unassigned so one thing we might want to do is go through and add more markers to to assign a identity to this cluster of EBI so to go back to the gene expression interface I've kind of giving you guys a spoiler alert here it's going to be an astrocyte cluster let's walk through the process of actually adding these new markers that are specific to astrocytes and these are again obtained from the fantastic it all paper and the first one is this gene aquaporin 4 you can type it into this box either the gene name or in this case the unsound blue gene ID and here's another gene jz6 you can see as I'm typing it autocompletes and tries to help you find the correct gene ID so just with these two jeans you can already see that a lot of stuff lighting up in this cluster that was previously on colored so now if I want to go back and actually create a category from the steam list I can go here and actually filter nuclei on these markers I can go to the already existing category based on the tasket biomarkers and create a new group called astrocytes so if I say that it takes me back to the category list where you already have the the three clusters that existed plus the one that I've just added for astrocytes this is the process that you'd go through if you want to actually take the unbiased heterogeneity that you see from the the tease me pop which is what we're looking at and try to identify what the actual cell types were that these nuclei originated from so another really exciting thing you can do with this data type is given that you can label some of the cells with these known cell types you can see that even among the the excitatory glutamatergic neurons you can see that there's other clusters all painted with the same broad cell type so one question you might ask is what's actually driving the clustering of these subgroups of nuclei together versus all being pumped into one big cluster and presumably these might actually be subtypes of excitatory neurons so when we can try to answer this question is we can go back into the the default graph based clustering if we can just remember that for example clusters 6 & 8 both correspondent to the excitatory neuron groups so if we want to ask what genes are actually driving the difference between these two groups that might be unrelated to the markers we've already added into the cell browser as we can do what's called a differential or cluster specific gene expression analysis so let's start by just hiding all the clusters and focusing in on cluster 6 & 8 so I've selected these two clusters and the question might be what genes are differentially expressed between these two clusters so the way you can address that question is go to the significant genes button on the bottom of the app and then you have two choices here the first one is called the globally distinguishing gene expression analysis and essentially what this does is it takes each cluster and compares it to the rest of the sample the other option is the locally distinguishing differential gene expression analysis and in this case what we actually do is we just compare these selected clusters to each other directly and so we're not actually using the rest of the data here we're focusing in on the two clusters that are highlighted and only looking at differential expression within this set of epi so I've clicked on the local gene expression analysis and what we're looking at now are the results of the analysis and loop cell browser I'll give this to you in a few different forms the first of which is a heat map and so we can do here is you can highlight the genes that are thought to be differentially expressed between these two clusters and the other view is this gene expression table which at the top shows the most differentially expressed genes sorted by p-value so we can do is we can go back to looking at what's actually in these clusters so this gene on top you can see that it's clearly differentially expressed between what we're cluster is six and eight it's pretty much off and this in this top cluster of excitatory neurons so if you want to go further down the rabbit hole way you can do is go and look at these biomarkers and try to understand from the literature what if anything other groups have said about these biomarkers and a few different ways to do this are to either just search for the genes through literature which is the most straightforward approach or you can go and do certain things like pathway and else move sella browser won't do this for you there are ways to export these genes and use them and other tools some examples of ways you can do this would be an are there a lot of gene ontology based analysis packages they're also some web tools like the venerable NIH David tool for example is one where you can basically just paste in a gene list and it will tell you which pathways are associated with those genes so given that I've mentioned that you can export this list of differentially expressed genes let's actually talk about how you would do that so before you export you have some options here that we're newly added in loops all browser 2.0 2.0 so the most important one is probably this which you know how to filter the number of genes that you would actually get so let's actually export all the genes once you've done that you can click on this arrow export table to CSV and you'll get a CSV file containing this table with the the gene names the clusters the p-values and the log to fold changes of the putative lis differentially expressed genes alright so that was how you would do an analysis on either cells or nuclei and loop cell browser that I think it's time for us to move on to the Q&A session it's not I'll hand it over to Shana to handle the Q&A all right thanks Paul up thank you sure Mel and Paul for a great presentation we're going to start with our Q&A session please remember to select the Q&A box on the bottom of the screen submit your questions we also have already have a lot of questions submitted so we're gonna do our best to get through as many questions as possible and I'm going to answer them in the order which they are splitted so we're gonna get started our first question is why do nuclei any different protocols depending on tissue just like cell nuclei need different protocols specifically for lysing as well as washing because different cells you're isolating nuclear from different cells and different cells or tissue types are different so the time required to lies each of them might be different the strength of the detergent required to lies each of these cell types to get nuclear might be different and therefore the time two lies will be different strength of detergent could be different the amount of debris that is generated host lysis could be different and so that correlates with what additional steps one has to use post lysis to obtain a clean good-quality sample so taking all of these considerations each cell type will require optimization to obtain nuclei all right thank you then our next question if you isolate from tissues how do you know the origin cell types of your nuclei well presumably you should know the tissue that CI came from first of all and that should kind of inform you as to the sets of markers that you would obtain from the literature typically the way people do this they'll build up kind of an internal database of markers they expect to be associated with each cell or nucleus subtype and that would basically just come from the literature so if you don't have that prior knowledge it's kind of hard to actually say what the cell types are all right next question what about disease related aberrations and processing of RNA that possibly happens in the soma will you not miss a lot of information by only sequencing the nuclear RNA and it's the example yes the most data you will get out is off a whole cell there is no doubt about that but and so if that's the kind of data you're looking for which is in the cytoplasm and you're not going to get from the nuclei the best way or the best route to go is using whole cell however we recommend the nuclear protocol in instances where it's difficult to get whole intact cells and that's where the nuclei because it has an expression pattern similar to the whole cell might be useful right next question why exactly are nuclei that are to isolate from tissues and hotels to nuclei and not break you could isolate nuclei from tissue as well as whole cells if you have obtained a very good quality whole cells you don't really need to do nuclei unless you want to study the nuclei specifically also if you're trying to isolate nuclear from tissues you could just bypass the extra step of isolating cells and directly isolate nuclei from the tissue so it's just you know bypassing one step to get to your final sample if you can do both this is in reference to the Y nuclei slide you shows the beginning of the presentation could you please we explain the last point under about revealing molecular genetic regulatory mechanisms specific to the nucleus by using single nuclei RNA because there are processes that you think are localized specifically to eat our nuclear rnas or rnase transcripts that are expressed solely in a nucleus that's where you would get more power because you'd be focusing on the nuclear rnas versus just looking at the mixture of the nuclear and cytoplasmic and I think you would expect that most of the transcript I saw would actually be cytoplasmic so in principle given the lower sensitivity of droplet based methods it's possible that by focusing in on the nucleus you have more power to detect things that are specific to the nucleus whether those be regulatory mechanisms or transcripts that simply are enriched in the nucleus versus in the cytoplasm alright and on the next question these are a couple questions and I'm going to group into one how do we count nuclei we have your bio-rad if the guys taken up by the new nuclei then what does it reflect and then there's another two questions about using track sorting as well for better clean up so we in-house use the countess to count nuclei we're using trypan blue it stains the nuclei blue and become them when we are when we do this counting we are looking for the the entire sample is stained blue and there are no life so that's one thing and so it's just regular counting can also use the hemocytometer to account using any of these dyes all work whatever is whatever you're comfortable with or is if your convenience why do we use backs for better cleanup facts can be used for better cleanup we have tried to put out a protocol that requires not require fax and in case fax is not readily available in the lab also it's a protocol that that might not take as long as it takes to run a run flow cytometry so the longest we have used for cleaning samples is the density gradient centrifugation which is a 45 minutes spin and and it seems the sample of facts requires people to be back strained and have the knowledge to sort but facts can be used and it can be used to clean samples up as well next question if I'm going to use Mac's sorting like the impact protocol which allows enrichment of cells but leaving beads that are used to pull down so can I load the cells with the beads retain will it affect the efficiency of connect single cell isolation no it will not affect anything in the 10x platform or you can load cells with the beat with the max beat on them we have tested this internally with other samples and there is no problem how you look at sspe samples we haven't yet lived at FFPE samples we are working on flash-frozen samples to have lit nuclei from that but we haven't done anything with formalin-fixed or paraffin embedded samples and we will be working on that in future hopefully we can get you some protocol in the near future okay the next question and so what evidence do you have that mechanical disruption is more variable than the trituration that you use in the protocol so we have tested mechanical disruption in this instance we have tested it I used a bypassing cells through syringes of different gauges to lies the cells up and obtain nuclei and the result was not consistent one it did not yield a hundred percent nuclear isolation in every every time so there would be fractions of cells that would stay back as whole cells or we also observed that the amount of background was much higher with a lot more debris and the barcode ranked plots where and as clean as we expect or we would like it to be so that's the only mechanical disruption that we have tested in house to know that it was not consistent and did not yield very high-quality data or high quality samples to run through our single-cell three-prime liter solution next question do you perform the myelin removal and density centrifugation steps at four degrees or room temperature so the density centrifugation step is at four degrees so we use a centrifuge that's previously cooled to for C and then on the samples on it like I mentioned we try to keep the nuclei on ice as whenever possible or at 4°c the myelin removal we use as per the milk of milk antibiotic protocol and the column for removing the myelin is run at room temperature okay great next question do I need any special reagents for running nuclei Otacon the 10x instrument no the nuclei nuclei can be run on the 10x system just like single cells it's the same chromium instrument as well as the same single cell three prime V to solution the reagents that we offer there is no change alright next question what is the density gradient centrifugation step remove the nylon yes it would remove myelin however we've noticed that especially in the adult brain the amount of myelin debris is really high and so just one so it helps to split the cleaning up steps into myelin removals first and then go to debris removal with density gradient just doing it all in one density gradient step does not effectively clean the sample up one could do multiple rounds of density gradient centrifugation to clean it better if they want to avoid myelin removal however the density gradient centrifugation time is 45 minutes whereas the mass removal is much shorter so we've effectively removed myelin using the beads and then remove the debris or any remnant myelin using density gradient centrifugation step thank you next question and look at what percentage of mitochondrial genes are picked up is this influenced by debris removing stuff we don't have that number on hand we can get back to you on that I will say that we have two nuclei data sets available on our website so if you want to look at the fractions of mitochondrial genes in those data sets you can actually get those gene expression matrices and and do that unfortunately we don't have that number puffer head sorry I don't recall it being massively different from the same number as it as you typically see in single cell RNA seek in our system which is roughly five percent at most typically my next question we're looking at the gene expression heat maps on the PC clock tick Satori and inhibitory neurons seem to be Koch Luster's also sells game to coax press workers of both populations how can this be avoided using single nuclei RNA seek the phenomenon you just mentioned is not specific to nuclei or single cell it's just a it's a question of how specific them workers are that you've chosen there are a couple of reasons why you might see markers lighting up in multiple clusters where you don't expect it first of all the most likely reason is that the marker just isn't 100% specific you know there's going to be some biological noise that causes some transcripts to be expressed in unexpected places the second reason is that you might have doublets we we know what to expect in terms of doublet rates shouldn't be terribly different for nuclei versus single cell in our system which is roughly 0.8 percent per 1000 cells I think that's contributing to some of the some of the stray points that you see lighting up and the tease me pot another thing is ambient RNA which we do see varies with handling of nuclei or and the same thing applies to cells and as you saw from the slides earlier adding more and more clean up steps to the cell handling and Prue's decreases the amount of ambient RNA that you had which should reduce the number of times you see a marker and unexpected places next question these are two questions about different sample type the first one is have you worked or trying to optimize other tissues especially mouse adipose or liver the second one have you ever tried none a million cells for example plants for nuclei purification no we haven't tried plant for new TI isolation as far as other cell types like the other mammalian cell types like the adipose tissue and the limitation we haven't tested them in house though we paying that this protocol with minor optimizations should be applicable to isolate nuclei from them as well but we haven't tested either of those samples in house all right and the next question you tried using frozen adult tissues can you comment on the general data quality between frozen and fresh tissue or cell lines so we've used frozen cell lines we we saw the cells and when we count them and we see that they are they're good viable cells and we have used that to isolate nuclei have we are working on generating a protocol to isolate nuclei from frozen tissues so this is we're trying to generate a protocol to get nuclear from either frozen and – CD or flash frozen tissue and hopefully we'll have something for you in the near future but we have used frozen cells where we just saw froze themselves as for meta cell manufacturer instructions and we count them under the counters to see that they're viable and healthy and then proceed with nuclear isolation from yourself complete in the next question during density certification and your experience do you select for a specific size nuclei no in during density centrifugation the nuclei pellet to the bottom of the tube we also don't when we are removing the supernatant post density centrifugation we tend to leave behind around 100 microliters and then we so that we have captured all the nuclei or capture all the pelleted nuclei that we can we think we're getting all this because when we compare it to our single cells data for similar tissue types we see a strong correlation great next question what algorithm is used for clustering so the default clustering algorithm that cell Ranger uses is based on lose a module area modularity optimization it's similar to what's available in the our package sûreté we also produce k-means based clusters some people are interested in those it's a bit easier to understand although the clustering quality tends to not be as good with k-means versus the graph based methods great next question does the pipeline take customized genome or say genome of model organisms other than human and mice yes we provide a tool called make reference that allows you to specify a custom genome fastest sequence and a custom gene annotation GTF file so if as long as you have those two files in hand for a given species you can generate reference that's compatible with cell Ranger currently on our website we provide for download human Mouse human Mouse mixture and I think Mouse by itself as well but we provide a tool that lets you build anything from those input files so yet you should be able to do that for any well annotated well characterized species great and the next question how are multiple gene markers summarized into one number to overlay on the plot in loop cell browser there are a few different options for aggregating genes on gene list the what I was displaying there with max which is kind of the most sensitive technique to use it'll take you know it's the maximum among maximum expression among all the markers there's also some I believe minimum is also an option and I think probably average as well was in there yeah so there's various different options for doing that type of analysis great we have time for one more question how do you try this new seek method with your me DJ assays T and or B film not yet but you're interested in once we have tried something in-house we will definitely have information put on our 10x support website my great thank you and about wraps up for today thank you to our ten experts Sharmila and Paul in to our attendees for some great questions we'll be posting a recording of this session on the tonight's community shortly and since we had a lot of questions come in I apologize we couldn't make it to them all if we didn't get your question or the follow up question we urge you to post it on the 10x community in our single-cell format community that 10x genomics comm and one of our 10 experts can take a look at it there so thank you very much and have a great rest of your day

Leave a Reply

Your email address will not be published. Required fields are marked *