Add Multiplex Validated with airSlate SignNow

Add multiplex validated

so I asked Sarah to come give us a primer on that Sarah is actually post like in my husband's lab so and I know that she does things meticulously her experiments are amazing in Germany and she worked on chromatin dynamics and gene regulation and developed a very strong interest in the emerging field at that time at that time combining wet lab and bioinformatics which i think is a very clever use of time I didn't bite somebody and about a year ago she joined her husband's lab as I mentioned and she started working on long no not coding RNAs which David black had been working on for several years and found some very cool things about them as opposed to what some other kinds may have heard this week from who we won't mention so a long time long non-coding RNAs do play a role in disease and other types of developmental processes so she's now looking at the non-coding transcripts own in a breast in several breast cancer models math models and of course this is the focus there is on disturbing new pragmatic and therapeutic targets so we're not going to talk about your research today unfortunately that would take another whole hour to focus on validate thank you thank you very much for having me Thank You Mona for the introduction so one big part of this whole RNA see experiments is obviously validating and the first question I usually get whenever I present any of my RNA seek data and I have a lot is yeah but how do you know that's true how do you know that's real are you sure that such as an artifact you know these kinds of things so definitely a major question here is how reliable is our RNA seek data and what does it really mean I guess most of you will probably work on differential gene expression and obviously you want to know if you find a differentially expressed gene is it actually differentially expressed or is it some kind of technical artifact you get here I don't know if anyone is interested in different splicing isoforms but you can also detect novel isoforms by RNA seek and how do you know that those are real so it's very it's a very tricky question actually and there's a lot of different methods to validate them I'll give you an overview of what these methods are however I will focus specifically on qPCR because again when I'm giving my talk because usually people are like but but have you done qPCR on these genes have you verified those and this is the like probably the commonly used method and that's why we will focus on there today but I'm happy to answer any other questions as well of course there's a pretty good paper which I would like to mention although it's already four years old but they have done a very extensive study of comparing rnaseq data to microarray data to qPCR to Sanger sequencing to everything you could think of and have verified how well these methods correlate so it's a pretty good starting point if you're interested in anything else but qpcr now what are the different validation methods for RNA seek in principle you have two options one is qualitative validation and the other one is quantitative validation now let's start with the qualitative ones most obviously the first one would be reverse transcription PCR such as to stand at cDNA synthesis and then you amplify your cDNA and then you run it on a gel this is of course not quantitative it's just a qualitative approach to see is it actually expressed and maybe you can get some glimpse of how strongly it's expressed but it's not it's really not quantitative so I wouldn't use it for that if you are interested in finding new exon exon junctions or maybe some snips at some stage you could also use classical Sanger sequencing to verify these kinds of things and finally if you would like to see it if you would actually like to see your transcript in the cells you can do in situ hybridization like RNA fish and I just quickly want to show you a picture just actually from a publication from our left the specter lab and it you can quite nicely see those red dots are a transcript called mallet one and there's that's our transcript called knit one and this displays already the second advantage of our nefesh is not only you can see your transcript to know it's there you can also tell where it is so RNA fish has an addition to the qualitative validation also a localization advantage now there's a number of quantitative methods you can use to validate your RNA seek data and the gold standard is actually still northern blotting although it's super old has been used for many many years but it gives you not only actually quantitation it also gives you different splicing isoforms for example and it also is necessary for some specific applications let's say micro RNA sequencing or small RNA sequencing which you might not be able to amplify very well with PCR so you actually have to do a northern blot however not a blotting is as you I'm sure you all know technically not super-easy the it's very elaborate you need radioactivity in most cases which you don't want to use and most importantly I'm sure many of you are working with tissues or like just low yields of RNA and then it's simply not feasible to do a northern plot if you have just a picogram of RNA so you can also use microarrays it depends a bit on what you actually want to answer which question you want to answer so maybe if you're interested in a range of genes you could design a custom microarray and just test all your genes on the array micro arrays are a lot more quantity there's an RNA seek in fact if you're just interested in one or two or ten yes yeah it depends a lot on what we are comparing so if we compare this classical RNA microarray where we just hybridize just it's done at hybridization the double hybridization you label your control in green and your sample in red yes they do run into saturation but actually you have a control in in on your chip yes so if you're exactly exactly exactly of course there's different types of arrays and some do not allow hybridization with a control and those are of course not the ones I'm talking about I'm talking about the double hybridization ones because then how many have a control on the same cell and they all hybridize at least statistically they should hybridize with the same probability to the probes and then you can quantify a transcript like really like a qpcr just a wider range in the end it is if you just hybridize a control on one chip and then your treatment on another chip then obviously it's not gonna be as quantitative anymore but I'm talking about the very classical red-green ratio microarrays and oh no it's not more quantitative than CT values no no it's more quantitative than RNA seek so like okay sorry maybe maybe I said it wrong no of course it's not more quantitative than qpcr qpcr is the most quantitative method we have I'm just wanted to make the point between RNA seek and my pro race so quantitative real-time PCR is as I said the most quantitative option you have and it is also very commonly used just if you're looking okay sorry can everyone hear me if you just interested in one gene or a couple of genes you just want to verify the most interesting ones from your RNA seek screen then qPCR is perfectly fine now how well do RNA seek and qrt-pcr actually correlated with each other there's a lot of studies out there who compared exactly that and they correlate reasonably well that's what we can say you have here on the x-axis a differential expression as measured by qpcr and on the y-axis the differential expression is measured by FPGAs by RNA seek and you get reasonably well correlation so the validation rate in this particular case was 88% and this is also what you more or less get so between 80 and 90% is what you usually should get between RNA seek and qPCR the blue dots here are the validated targets purple ones are the so called housekeeping controls and then there are some red dots who are not validated which means they did not show up differentially expressed in the qPCR but they did show up differentially expressed in the RNA seek and usually that's also what you get so that's a classical example it happens sometimes that a gene shows up differentially expressed in your RNA seek screen but might not be you might not be able to validate it afterwards so yeah yes yes very good very good question and you're absolutely right there is and usually all these correlations are best when you are in a medium range of expression very highly expressed in very lowly expressed genes have some technical issues in both methods RNA seek and qrt-pcr and that's why the correlation is diverging both up and down so the medium range like the standard protein gene expressions that's what where the correlation is the best okay approximately approximately I would say if you're looking for protein coding genes for example I would probably not go for anything which is expressed lower than 10 F pkms for example because that's usually it's oftentimes considered as noise its background you don't really know is this a lowly expressed gene it is a lowly abundant gene or is it actually noise so it's it's it's hard to give values but I personally wouldn't go for anything below 10 fpkm I don't know helps thousands yeah so the cutoff line here would be by one and this is a lock to fold so it means these are two fold or more overexpressed as determined by RNA seek however they could not be found in the qpcr I don't actually know if there is a statistical there's any statistical correlation and you say oh if my Q value is basically zero then for sure this is actually a differential expression or not I guess the best way to find that out is by doing biological replicates and see if you see the same effect more often then I would rather believe in it then if I just see it once and I wouldn't be too hung up on their statistics in that case okay now let's get technical in general you have two different options for performing a qrt-pcr it's either one step or a two step qPCR and depending on what you're trying to do one or the other method might be more beneficial basically what it means in a one-step rtq pcr you do both the cDNA synthesis and the qPCR in the same tube so you have your RNA your reverse transcriptase a DNA's a DNA polymerase attack and of course buffering 20-piece and you only have sequence specific primers so you add your PCR primers right to the seed unison to this which has the benefit that only those get amplified and you have a very sensitive measurement in that case however you're not very flexible because you don't have any cDNA which you can use afterwards if you want to run a different gene a different target afterwards you basically have to do the same thing all over again so it depends if you want to verify loads of targets with the same cDNA maybe you want to go for two step if you want to just really take one but you have a panel of 96 different ones and on a plate let's say if you're doing a screen it might be a bit faster but it's really up to you which which one you prefer to step qPCR would be just as I laid out you have an r/t se DNA synthesis step beforehand with the oligo DT s random primers or you can also here use sequence specific primers and then you use the cDNA subsequently for a classical qPCR you can multiplex a two step you can not multiplex to one step so depending on your application either the one or the other might be beneficial it depends on your cue PCR machine so depending on how many filters it has so in my all that we add one which could do six different colors about the r/t reaction oh okay okay I see yeah I guess I would actually then not multiplex it I would prepare one individual reaction that's what I would do just to make sure not yeah so again it depends on your application I personally use random primers because I'm breaking on non-coding RNAs so I want everything to be amplified so random hexamers of course amplify your are RNAs your T RNAs everything which doesn't have a poly a tail whereas oligo DT only amplifies things which have a poly a tail so if you're only interested in those transcripts maybe all the go DT is the better option especially because then you can be sure you're amplifying a full-length transcript whereas the frenum hexam is you also amplify everything which has been degraded and it's only present in fragments anymore okay now how does this whole process of qrt-pcr work this is a typical qpcr curve you have in the beginning and exponential face which means at this stage you're neither your primers or your empty piece or anything else in your reaction is limiting and that's why you actually have an exponential growth so out of let's say one nanogram of cDNA you get in one cycle two nano grams and then you get four and it's just growing exponentially however it's very soon you're entering the linear range where then primers and all other components in your reaction will become rate limiting until finally you reach a plateau which this basically means that your detector at your Q PCR machine is saturated and can't detect any additional fluorescence anymore and if you have a good Q PCR reaction that's how you want it to look so you wanted to actually plateau out if you're just if it just ends here around it maybe you'll want to use more RNA for your cDNA synthesis or have a higher concentrated cDNA in your reaction there's a lot of troubleshooting you can do but in the optimal case it doesn't work in all cases okay granted but in the optimal case you reach a plateau here and there's a relatively arbitrary set threshold which is just defining what is actually a product here and what's big noise what's background and as soon as your product is reaching this threshold this is called other cycle where it reaches a threshold is CT value and that's what you actually use for calculations afterwards this would be a dilution row just to show how this would work if you have ten picogram of starting material you your CT value might be relatively late maybe at 35 if you have 100 picograms it's coming up already cycles earlier and then one man of runs 10 nanograms a hundred and a thousand so you just can see that it is very dependent on the concentration and in the optimal case it should be linear correlation if you plot the total RNA concentrations versus your CT values that's kind of a quality control so you want if you get this plot you want that the R square the correlation coefficient is very close to 1 to 0.99 0.98 something like that and you want you just want to have a lot a nice linear correlation here this is just a quality control to see if your reaction works at all or not - 3.32 or something it just it is yeah so this is this has to do with primer efficiency and I will actually cover that later on so can I postpone this to later or okay fantastic and there's a lot of different ways of performing qrt-pcr and that depends on your system on whether or not you have a rush light cycler or applied Biosystems or whatever you have in your labs they have different technologies I guess the most commonly used apart from cyber clean would be taqman chemistry I will only cover cyber Korean in the interest of time and I only know about tech mint otherwise so please don't ask me about molecular beacons or something like that I don't know it so cyber green is the commonly used fluorescent dye for a qrt-pcr and a good thing about is that it has almost no fluorescence when it's not found to any DNA and as soon as the primer and kneels also cyber green can intercalate it integrates only double-stranded DNA and then in the course of amplification you get more and more cyber green intercalating and thereby more more fluorescence and that's directly correlated to the amount of cDNA as long as you are not exceeding the exponential phase of the linear range okay now I was advised to give an introduction into primer design so that's why I'm putting a lot of time into that now I hope that's alright please if it's all totally obvious let me know so there's a couple of general guidelines on how to design the best optimal qrt-pcr primers if there is something like that and if you look it up online you'll get a thousands of advices like how many GS and C's should be the 5 from another 3 prime end and whatnot I personally go with these criteria and they have always worked fine for me so I guess in 99% of all cases you will be fine and not run into any problems now the most obvious criteria always is it should be unique for your target sequence so unless you're working with a pseudogene or anything which is very repetitive you shouldn't have any problems but you still should verify that your primers are actually specific only for your target and I'll come to specificity in a bit secondly it's beneficial if the two primers you design are not prone to primer dimer formation otherwise it could easily destroy your reaction and also they should not form any secondary structures if possible because again annealing will be inhibited usually the recommendation is there should be between 18 and 25 paise per sets enough to get them specific for the target in the most cases and they should have both formatted reverse primers should have a comparable GC content because that ultimately results in a comparable melting temperature and of course you don't want one primer have a melting temperature of 50 and the other one of 80 that's not going to work an optimal range of melting temperature is usually said to be about 60 degrees but in my experience anything between 55 and 65 works fine the length of the amplicon again you will find a lot of different advice here for me 60 to 200 has always worked and I think it's a very common range to use and finally if you can then your primer should spend exon exon junctions and here's why even though even if you treat your CDN erection with DNA's or your RNA extraction with DNA's and I guess many actually don't do that to start with you might have some contamination of genomic DNA in your sample and now if you have primers who spend exon exon junctions like this one does so half of the primaries and X and one and the other half binds to exon 2 it can not amplify genomic DNA but it can finally perfectly fine amplify cDNA now the same is true if you have a very long intron in between in case you can't design anything which borders exon exon junctions you can choose some which are separated by very long intron in that case it's also very unlikely that any product will appear from genomic DNA I have a list of thousands of databases which you could use and of course you all get those in the end so you don't have to copy my slide here there is a number of non-commercial ones which means they are free and that's what I always recommend to use I have pretty good experience with primer Bank from Harvard University they have a very extensive set of qrt-pcr primers and the good thing is you usually also have a gel picture of those primers so they actually run it on a gel show you the picture and then you can decide whether or not this is according to your standards so sometimes you look at primer sequences you see the gel and there's no product I wouldn't order those well it happens it happened to me I promise yeah exactly yeah so that's that that's exactly what I wanted to mention most of those public databases are for mouse or human because they're just the most commonly used organisms some of them are also for other organisms but if you are working with something very exotic you might not be successful with the public databases there's a lot of commercial ones as well and if you're working for plat bio systems you might consider they actually take some primers which they are validated for you they could do cost money but the at least they will work for sure too so if I do I guess the most famous one is primer 3 and I have only one webpage here but it's open source so many other universities there's definitely at MIT and then probably at any University everywhere they are all using primer 3 and it is pretty good it is these are free completely free yeah so primer free is an open source which means it's web-based you just copy paste your sequence there yep so it it is true that if you design anything online they will give you an estimated TM and then you probably order the premise and you get a different TM that's true they shouldn't vary too much at least they can vary by one or two degrees five freely ok the time that doesn't happen to me and but ok I usually go still I try to resign my primers aren't 60 degrees even if they say it's 59 I still run above 60 it has always worked so far but of course something to keep in mind that whatever they predict as melting temperature might not be the actual melting temperature and I'll come to that too later there is also a web-based a surface called primer-blast and primer-blast is ncbi based it just combines primer 3 software with a blast search so that's also quite handy I use it a lot if I because I'm regularly pseudogene so I need very specific primers and you just copy paste your sequence there you have to adjust the length of the amplicon because by default it's I think a hundred to a thousand and that's too long for QP CRS at least but otherwise it's really good and it will give you not only primers that work but also promise that are specific to your transcript right away there is an open source software called benchlink which can be used also for cloning and I think maybe nowadays even for high-throughput sequencing and analysis a little bit at least so you could have a look at that it's for free and then there's something which I have used a lot in the past because it's one of the few software's which run on a Mac and that's called teenies however that's not free it is also based on prime of 3 though and it also has all these vector nti opportunities like we can do cloning or also have who put analysis actually with that with this one I know for sure you can do that now about primer specificity even if you do not design your primers with primer blast it is a very beneficial if you afterwards use primer blast or just a basic blast search to verify my primers are actually specific to my target and they don't accidentally pine to 100 other transcripts you will see that sooner or later but maybe you want to know that before you order the primers into the whole reaction another point and that's coming back to the melting temperatures you this is pretty much out of fashion nowadays no one is optimising melting temperatures anymore I have done it a lot actually and it definitely my undergrad and also my grad work so you actually could run a gradient if your PCR machine is suitable for a gradient run of course it is and tried different temperatures from 55 to 65 for example is a common range and see if your primer is specific or if it amplifies your target at a certain temperature range better than in another one yes yeah yeah definitely definitely exactly exactly so usually if I have four mismatches I don't bother because they shouldn't align to they can maybe in theory alight but honestly they will not in practice at least I haven't seen that so far I am running into similar problems with my pseudogenes sometimes because they can be up to 99 percent sequence homology to the protein coding gene and what I do then is I actually blast my entire sequence of the gene are looking for not prime a blast but use the actual blast and then look whatever is closest and then go through the FASTA alignment and check for snips and actually check ok maybe there is a gap so this is a good region to decide my primers across or if there's really just a couple of snips I designed both forward and reverse primer on a snip and try to have them as the last a second last base of my primer and then on the 3 prime end because it it still could amplify and definitely in these cases you should probably use standard Sanger sequencing to verify what your hear sequencing here or use some other kind of verification of what your where your amplifying but it is helping in getting the right target sequence at least in my experience you can try you can definitely try and see if it in one in at 55 degrees amplifies in my case the protein coding gene and at 65 maybe only the pseudo gene because there is a 82 CG a base change or something like that so that's definitely you can also play with magnesium concentrations that can also help if you do a gradient like that but I have the feeling that most people just nowadays order primers to have qPCR don't bother too much so there's still a lot of things you should actually bear in mind I believe another good way and this is actually also be used for snips is a melting curve so I would always recommend to run a melting curve after you're actually cue rt-pcr reaction because it will tell you how many products you get and you can do a high-resolution mouth curve if you are bothered about sequence specificity so if you have only if you even if you have only one snip in the entire amplicon a high resolution marker should tell you which one it is or if you if you get one or two products so that there could be an option to and also essential is always run your primers on a gel it sounds old-fashioned again but at least if you have new primers even if you have a perfect melting curve you might want to see it on the gel and see this is only one product and this is certainly not a primer dimer what I'm seeing here yeah the appreciat product yeah okay sorry I do both oh yeah you can run afterwards after the qpcr is finished you have the plate or your tubes and you just run it on the tongue oh no no no no you can still run it on the jail yeah only in case as we talked about the sequence-specific effect if you have if you're not sure if your primers amplifying a pseudo gene or protein coding gene something like that then yes otherwise if you know that your primers are specific and they can only amplify this one transcript and you have the right band on your gel and your Molotov looks fine I would not send it for sequencing I don't think that's necessary in that case and finally a good way to verify that you're actually getting the right product would be multiple primer pairs per target and if you get the same result of all these primer pairs then you for sure you know that what you're measuring here I'll show you an example about from our publication this is by the way Tom Watson not James in case anyone is wondering and it's about the expression of a visual homeobox gene in Mouse and they compare reginal to cornea here and I'm just showing this because it's a nice example of how qpcr can look like they use three different primers which are these three blue lines they are not reaching a plateau so that's the only downside of this paper I find but otherwise they have a contra great gene this is Gabe th in that case which they normalize it to and they have also a row control which is specific for the retina and not for the cornea so that's how they verify their results and that's that would be a good example to do it and it's quite obvious that this V is X protein is only expressed in the retina and there's absolutely nothing there's baseline in the cornea so if you get something like that you could be reasonably sure of what we're doing here they also run it on the gel afterwards the free primer pairs I hope you can see there is one product two and three so it's a single band there's no primer dimers there's and it looks really clean and same we take up the age and roll primers as well yeah actually you can't see it very well in the screen it looks a lot better on my computer so I think it's just a matter of display this is a little weaker but it's pretty similar same range as the other two are on my computer I can show you later if you like so that would be prime up here too and as you can see it has a little lower CT that's the other two so it does affect them but usually if you run into saturation in the end you shouldn't see a difference of your primaries on the channel okay now just a comment about melting curves this is a specific product so you get nothing nothing nothing and then at a certain temperature you get a peak and then that's it that's how it should look like if you get two peaks like one here and one there you have probably some primer dimer contamination here this is in this case this is the negative control and then the negative control it is fairly common that you do get that because there's nothing to amplify really so that's okay here but in your specific product you really just want one single peak okay I have to hurry up it I'm sorry the last thing I want to talk about is how to analyze qrt-pcr data and most people I guess still use the delta delta CT method so that's why i'm talking about that one and i will give a short note about primary efficiencies and other methods in the end so basically you need four different things you need a control gene for basically every targeting you need always a control gene but you also need actual biological control of your treatment so this could be an untreated case in a treatment case or two different genotypes or whatever you're working on basically so red line here would be the treated target gene and the red dotted line is the treated control gene and the blue dotted line is the control control gene and the solid blue line would be the target gene in the control and you can see that basically the controls the control genes both in treated and untreated case are pretty similar and that's what they should be so if you have a large variation of your control genes already you can probably not really quantify anything here and you can see also there's a bigger difference between the target gene so actually your treatment seems to have some kind of effect in that case now how do you calculate this exactly as the name applies delta delta CT mine means you have to make two deltas and the first delta you calculate is the difference between the CT value of your target gene and your control gene and routine again could be like a p' th for example and the target gene could be this visual x protein we just had in the example and then you calculate the difference between your target gene and your control gene and you do this for both your biological control like the untreated case and also for your treatment or your condition whatever you want to compare to each other so you have two CT values and they call it calibrator here so that would be your control or your untreated case and then your ever unknown condition or your treated once and you calculate Delta Cities for both and then you can carry you can calculate a value called relative quantitation rq2 calibrator would be to the power of minus delta delta CT and delta delta CT is the difference between the calibrator your untreated control and your treated control does that make any sense I'm actually coming to that can I postpone it for a few slides almost there no worries I just want to make a point about primary efficiency because there was a question about the slope so obviously for the whole delta delta CT method this just works if you're comparing apples to apples and not apples to peers which means the primers of your two genes you're taking your control gene on your target gene should amplify with the same efficiency and I agree it does not in all cases very easy to get you can I can calculate this e value where you plot on the y axis the delta CT and then the x axis the log of the relative quantity and then you should get a linear correlation and you get a slope and now the slope in the optimal case is something - three point three to something around that basically it just means that in the end if you solve this equation you want that your efficiencies one or a hundred percent and that would then mean you have an exponential decrease from one nanograms to two to four and so on and now most primer pairs don't have a hundred percent efficiency but you should aim for something between 95 and 105 if you're in that range you're pretty good and then you can put you can compare these different ones if you are not in that range you have two options one option is designer primers and that's what I would always go for the second option is if for some reason this really doesn't work you can use efficiency normalized method and I can give you a hand out where they actually have a paper of registers described in a very detail and very long story short it just basically means that instead of two you have a different value here because you can't assume that you get the same thing exponentially so maybe if you have a primer which only amplifies with 90% efficiency it would be 1.9 to the power of minus minus del minus the other diversity something like that it's not the optimal case but many people do use it too oh there's a whole paper about it and I'm happy to contribute it like if you if you I said oh yeah exactly it's from Michael pop buff I think he sounds German ffs I think it's PF AFF or PFF PFF I'll give you the reference anyway so you can read it up up on it if you like I would aim for 95 to 105 percent if it's only 80 percent or something like that I shouldn't certainly would not use that primer pair mathematically you can you just have a higher fluorescence value then you should have for a double as much product as you so you would expect that your fluorescent is linear to your increase of product so for one to two to four nanograms you also would expect one to four to two for fluorescent units so if that's not the case for example because your primers form a little bit of primer dimers and you have additional fluorescence or for any other reason which happens in your tube you have too much fluorescence then you get an efficiency 105 or even more than that that's a very good point that's a very good point I actually stopped using any primers from any publications because they never worked from here and I always get like triple bands and and it can of course depend on different buffer systems different qPCR machines different people pipetting I don't know but I I don't have very good experiences with those either it depends on how difficult it was to design your primers if it is a standard protein coding gene and you can design easily a new primer pair I would go for that if it's a bit more tricky I would run a gradient it's also not every machine can run a gradient so it depends on what equipment you have in your lab a little bit actually and so in my entire PhD that's what I did I optimized I'd optimized every single one of my primers to the optimal TM and just use this TM in the future because then you have maybe 95% in one case but 100 if you just increase the temperature by a degree so that's what I did I'm not saying everyone has to do it of course but usually 10 a dilution of 10 samples that's what I do and in triplicate as everything always in qrt-pcr always triplicates if there's unless you have a very good reason why you can't use triplicate sure the last thing I want to talk about is housekeeping chains or control chains and in this example as you can see both control chains both in the untreated in the treated case they are pretty much the same and that's what you would like to have now how do I choose which is a good control chain and which is not a good control team I don't actually like this word housekeeping genes because they are not really housekeeping they vary between tissues and they vary from one cell to the other sometimes and classically used ourself is something like up th but it's not necessarily a stable control it depends very much on your sa whether you're looking at general any effect which would for example mess up your general gene expression pattern then it can easily also affect any of your housekeeping genes so general guidelines again would be expression levels must be similar in all tested samples they don't necessarily have to be similar in all tissues in your organism or across organisms but at least in the things you're testing they should have a similar expression level they should be resistant to the experimental condition so if you see there's a big difference after you treat your samples for example then don't use it as a housekeeping chief and finally as we talked about before you want primers which are giving the same kinetics or the same efficiency for both your control and your target gene so if by the efficiencies they should all be in the 95 to 105 range if if you have two primers which both have 95 that's perfect then you're comparing the same range although they don't amplify with 100% efficiency the amplified with the same efficiency involved still if you have two primer pairs which amplify with 80% I still wouldn't go for it it's just not good qpcr result now we talked about gap da and other commonly used housekeeping genes or control genes would be beta-actin tubulin then oftentimes people use either our RNA directly or any a transcript of ribosomal proteins and they can be good controls in some cases but then are not necessarily good controls so if they work for your system if you don't see any change ever and they're totally stable fantastic you can go for it in case you don't and that happens many times you can do the so-called basket normalization which is a lot more resistant to treatment and it's a more it's a lot more accurate and therefore you use three different control chains or however many you want to use but three is general what people do and you use the geometric mean of those three control genes and use that for normalization of your target genes and you can use any three which you think are for your sample stable there's usually nowadays there's a lot of publications out there which you can read up which genes are stable in which tissue at least definitely that's the case for human mouse hamster and so on that I don't know that I don't know but there is a lot of publications out there where people checked which genes are expressed stably and accurately across tissues and across organisms and you can definitely start from there and try of some of these and if you then see they're still varying just choose a different one so maybe you want to start with five to beat your test and choose the best three of them okay if you have something that works well don't bother but yes no no no no that's just no you don't have to do that no you don't have to do that it's just it's May it's making it more robust and more accurate in case you're worried about your control chains if you're perfectly fine and if it always comes at the same Ct whatever you do you don't have to worry about it yes yes so it's always best to choose your control chain with a similar CT to your target gene it should be at least in the same range of let's say two to three cycles in either direction that's fine too and obviously if you do a knock out you have like no targeting in one case or maybe 10% left then you can't do that but then you would go for the for the normal normal expression in normal tissue yeah yeah that's that's absolutely correct that that's a very very good point Soph all these things although the qpcr is actually quantitative the seediness synthesis is not at all so what you absolutely have to do is a consistent amount of irony a study material so you always have to check what's my RNA concentration and use the same amount of RNA for all of your reactions you also have to be careful about the linear range of your C DNA synthesis reaction they usually tell you that in the manual this cDNA Center this reaction is stable between 10 nanograms and a thousand nano grams let's say and then you shouldn't use 5000 nano grams because you're not gonna have a linear amplification anymore so that's certainly one thing if you're really worried and that's also what I did to my PhD I did a dilution row so I actually made cDNA of ten hundred thousand nano grams or you name it whatever you want to use and check if there is any difference if I use different amounts of RNA starting material and what would be the optimal range for me for my essay so if you really want to make it 100% correct go for it and to RNA dilution row first for a cDNA synthesis and see what's your stable linear range so usually you're good with nanodrop unless so the the range of nanodrop is between 10 nanograms per micrometer in 2000 so if it's a lot less or a lot more is a diluted or you have to use a different method which would be by analyzer or cubed fluorescence based for the RNA for the our name for the cDNA you cannot necessarily measure the concentration because you have your primers in there and your dntps which don't have which have not been used so it's difficult to afterwards estimate how efficient it was you can only really estimate by diluting your RNA and then running a qPCR and I've only one more slide and then I'm done I just want to point out this paper because it will I think aren't there any residual questions which you might have if the miq guidelines are minimum information for publication of quantitative real-time PCR experiments and it is a very extensive list and you can take the boxes of what you should do with your how you should how you should prepare your RNA make your cDNA which controls should you use for aq PCR reaction which control genes would be appropriate anything and it's very extensive I would not recommend to adhere to all of these points because otherwise you can make an entire paper just out of your Q PCR reaction but it is certainly helpful if you have any questions it has a lot of good background material and that's the last thing I wanted to show you today if you have any additional questions please feel free to ask me trust me if you read it you will understand what I mean what you actually should do I think the minimum would be melting careful analysis rania primers on the gel make sure they are specific make sure they're i ficient make sure that you're using the same right control genes for your essays or something which is stable something which is accurate if you can't find anything use the basket normalization that's the main points I think and there's a lot of more things which you can optimize but if you adhere to these three points you are in 99% of cases covered maika Arnett analysis actually there is a for micro rna's I don't think there is anything out there for micro rna's as far as I know like there's no such guidelines as there is for standard long jeans as far as I know micro RNA people usually use northern blotting to verify that data in some cases you can also use qrt-pcr but I would if I was working with small RNAs I would definitely go for northern port although it's elaborate and not fun but I think it has it has its justification there