 Okay, so we've done two modules, mostly sort of introductory ones are focused as I said today and this lab is on quantitative metabolomics. Here's the standard set of slides about creative commons licensing. And as I said this is the lab that we're going to be doing. And we have a total of two hours allocated for the lab so we're starting at three and ending at five in Montreal and we're starting at one and ending at three here in Edmonton. Now for the lab I'm going to give you guys roughly I don't know about a 35 40 minute introduction to the lab. This is where I think if you can please download the slides so you have them internally on your computer because you're going to have to look at things might have to write some notes down because you have to go to certain websites. The lab and I'm showing you the slides partly take you through what you're supposed to do so. If you're stuck, you should be able to find a slide that says click here. So that portion work, talk to your neighbor, they might be doing exactly the same thing, or they might not it depends on the groupings here. But this is intended to be interactive. So it's going to be open, open mic for the first 35 40 minutes and then we're going to close the mics in Edmonton and in Montreal so that each group can work with their TAs and go through the different exercises. So roughly an hour and a half to finish up, you need to be you're going to do three different exercises for three different platforms like explain this in a bit more detail so we're trying to do quantitative metabolomics you're going to be processing experimental real experimental data from real samples that were collected and harvested if you want for this particular project. So you're going to do NMR. You're going to do LCMS and you're going to be LCMS for NMR you're going to use the software that I just introduced called magnet for GCMS they're going to introduce the software called GC out of it. And for the LCMS we're going to use LC out of it. So three platforms, three software types. Now, because these things are pretty compute intensive. It's not as if we can get everyone there's 43 of you I think in in Montreal and Edmonton to take all these queries all at once. So that's just not possible with today's computer technology so we're dividing you into several groups. And so it's kind of around Robin. Three groups, three platforms. That means that roughly gets 13 to 14 users at a time, which is what these servers can tolerate. So there's going to be two files were two experiments that you're going to get for the GCMS and the LCMS. And you guys are going to work with one file for NMR. So it means that NMR partly takes a little longer GCMS should be fast. I don't know how long the LCMS will take that hopefully it's reasonably fast. Now the idea is you work on your platform so if you're assigned to the NMR platform, you're going to work on that file for 30 minutes. So it might take 15 minutes for it to get a result. And then for the last 15 minutes you can play around with the spectrum but you're also supposed to upload your data to a certain server so that we can pool everyone's results. So there are slides sheets or the CBW github page where you'll see your name and which technology platform and which software tool, you're supposed to be using. So if you're with NMR you do magnet if you're with GCMS you do GC out of it if you're assigned to LCMS use LC out of it. Don't mix LCMS data with NMR magnets. Okay. So you're going to have your files the files will have your name associated with them. You can download those onto your computer, and then you're going to start your designated web server. And that means when you start it you're going to have to upload your files to the web server can process it. Now you're processing, sometimes it's interactive, most of it's going to be automatic so you just click and kind of wait. And then the answer appears, and it's once the answer appears that's when you can kind of explore and check and do things. So you can download the files onto your computer, those are all going to be CSV files or Excel files, comma separated value files. So you can take a look at them, and you can save them. Some of you know a little bit about HMDB or other resources you can compare the values you got with what's what's known. And then you're going to be uploading your CSV files to the CBW Dropbox page. Those are all going to be consolidated, because what we're doing is that sort of crowdsourcing science. You're running these things. You're doing it for many different samples. And for, you know, three different experiments with three different subject groups. And then tomorrow you're going to be able to use Metaboanalyst to analyze those data sets. And you're going to see what Metabolomics reveals what's different about these what's unique. Maybe find some biomarkers, maybe discover something that we haven't discovered. So for the GCMS files, there's 84 urine samples from patients. And we were looking at a study with sleep apnea. So some people had sleep apnea, some didn't. Evidently, there is a difference in terms of their metabolomes. And this was analyzed on an instrument up on the eighth floor here, which is an Agilent 7890 GCMS. And I think measures about 7075 metabolites. Do you guys remember exactly? The NMR files are going to be 42 serum samples. And this is nonhumans. We're looking at sheep. And we're looking at pregnant sheep and non-pregnant sheep. And the idea here was to see if we can find a marker that can distinguish between pregnant and non-pregnant sheep in the early stages of pregnancy. So if you look at sheep, you can't tell which ones are pregnant, especially at sort of 50 days. So it's a little blood test than you could. And that makes a difference for sheep farmers because they have to change the feeding rations and other things for pregnant sheep. So this is part of a study to develop a biomarker for sheep pregnancy. Second, a third one is 84 serum samples. These are patients with early stage lung cancer and then controls. And lung cancer, which is normally quite fatal and is usually not discovered until stage three or four, can be cured quite routinely if you identify it at stage one. Unfortunately, we don't have any markers to do that. So this is a study to identify markers for early stage lung cancer. And as you guys will see, I think, I hope, well, when you do your study tomorrow, you will find markers for early stage lung cancer detection. So this was done with LCMS. It was done with a smaller study just to make it simpler for you guys. It measures about 145 metabolites using LCMS, using a Q-trap 4000. Again, all these instruments are either up on the eighth floor or the 700 megahertz is just down the hole in the third floor. So we've talked about magnets. I've given you a little bit of an introduction of it. This is the same text, but the access that you can have is either teaching one dot magnet or teaching two dot magnet. That's a sign to you. So if you've written it down or if it's in your slides, make sure you go to that. Same thing with GC AutoFit. Explained it before already. Again, two servers. Make sure you know which one. It's either the GC AutoFit one or GC AutoFit two. And those are the URLs that you go to, depending on what you've been assigned to based on the slides I've just shown you. LC AutoFit. Again, we've done a quick review about what it's like, what you do. You go to the LC AutoFitWishartLab.com. So everyone can do this one. It's the same server saying it's not two, it's just the one. We've explained the general overview as well for NMR. Your focus for the lab is from the processing the NMR spectrum to get the list. All the other stuff has been done for you guys already. Obviously we can't have a lab with everyone working in the lab. So this has been done for you. So if you're doing NMR, you can go to this website. Again, the link should be on I think the Slack channel or GitHub or both. And you're going to see your name. So, yeah, you're at the top. And so that's your file. And that's the one that you're going to download. Delissa and Amber, I think it's all alphabetical. So you can, you know, find your name, download your file if you're doing NMR first. If you're not doing NMR first, don't download the NMR file. So let's say you're doing NMR. Then you're going to go to teaching one about magnet or teaching two dot magnet. And these are just pictures of that. So this is the instruction. Now, I'm not going to go in a lot of detail. You guys can download this, but this is what you'll see, you'll see, you know, at about menu. It'll explain things. And, you know, you can look at it or you can race through it depending on whether you're into that sort of thing. So once you've started, then you're going to be able to submit your spectra. So it says welcome to magnet. Thank you for using it. We're so pleased. And then it goes on say upload how to submit your files. So if you scroll down, it'll explain this and you'll see this instructions to analyze. It does have some examples. There's no need for you guys to run example files. You have real files to work with you've hopefully downloaded them. So, you're going to select your zip file, you're going to enter a submission name. You're going to select none for the pre processed option. Because we're analyzing serum. These are sheep syrup that you have. So that's your bio food and we're using a 700 megahertz NMR instrument so the data were collected on a 700 megahertz. In NMR you can collect it 500 600 700 800 and so on. But you were collecting on 700 or the data for you as a 700 so don't choose 600 don't choose 800. And make sure you choose serum because that's what you're analyzing. So once you've uploaded your data, then the computer starts churning and it's doing that spectral deconvolution talked about. So it's going through each peak and there's hundreds of peaks in the spectrum, and it's picking out the peaks, matching to things in the library, trying and guessing and checking like a human would, but it's doing it automatically. So after probably about eight or 10 minutes for you guys, it'll produce an answer. So during those eight or 10 minutes you can have a cookie some coffee walk around chat with some friends. But eventually you're going to get a result would look like this. This will be a spectrum where it's been fitted. And you can view those spectrum or detail. You can click on the spectrum you can download the spectrum. It's giving you a list of you know what's been going on what's been processed whether it's completed or whether it's being processed. If it was completed, this one is still being processed. You're only hopefully submitting one file so you should just see one I think listed unless if everyone's submitting at the same time I guess it will at least all of them. I can't remember. So if 12 of you submitted you're going to see 12 files make sure you know which one is yours, and you don't steal someone else's. Yes. And my frequency select by the preprocessed is optional. Did you have to submit your chemical shift reference I think it should know the chemical shift reference that's that should all be there. It does ask for. So just the default options. So it should have chosen DSS. So you should just take those defaults. So maybe we'll put a six and seven is choose keep your defaults for that. So once you're completed, you can click on that it'll pop up your spectrum, you can explore your spectrum, and there's a control panel so there's a big down arrow. And that'll expand it to the control panel which allows you to view things. This is a tool called J spectra view. It was a Java spec JavaScript tool that was developed in the lab that allows you to view spectra it allows you to do NMR spectra GC mass spectra LC mass spectra interactively on the web. That's been a lot of work, because most resources require computers full time CPUs to do this so this is a lot of work to get this interactive. You can expand and blow up your spectrum you can zoom in you can drag across the spectrum. You can see how things are fitting or some cases not fitting. The interactive zoom is available, essentially for all the spectral views. So we've zoomed in here you can see where things are closer. You can show your compound file. You can adjust your fit you can lock your view there's a bunch of options that are there. So again, once it's done you should spend a few minutes just exploring things. You can mouse over with your mouse or trackpad and click on peaks. Those will also identify things for you. You can search so this thing will have identified about 55 compounds and will determine the concentration for all of these compounds in these sheep serum samples. You can type in a name here we've typed in veiling. There it is listed. You can click and highlight that. And when you clicked on the name of the compound in the screen below, it highlights the veiling chemical shifts in the spectrum above. Those are the dark blue peaks. So you can see how well it fits or isn't fitting most of the time it should fit very, very well. And so this is done this spectral deconvolution. You can see some cases where peaks are, you know, there's peaks directly overlapping or not overlapping. Once you've played around you can then download things so you can download the magnet file you can download the spectral database in a CSV file. And so the CSV file will be a list with the hMDB identifiers and name the compound the concentration. You can get the magnet data as a JSON file. You can drag and drop that JSON file back into a standalone viewer to see things again. So this allows you to revisit your data. So that's the NMR one. So this is what the NMR groups should be doing. At the same time, there's going to be a GCMS group. So we talked about how GCMS is done and how things are processed. So you're focusing on this last part process GCMS chromatogram analyze where you see out of it, get your list. The other stuff was done for you to collect in this case urine samples. So if you're in the GCMS group for part one, your name is here. You can download your GCMS file. And this is the link. You don't have to type in all 18,000 characters. That link is on GitHub or Slack. Is that correct? I'm good. So each of these long ridiculously long links. They should work on your slides if you've downloaded them. Okay. Okay, again, I'm just trying to give you guys an overview. If you're feeling a little lost after I finish the overview, then you can start asking questions that Mark and Alan and others. Okay, so in this case, if you're doing GCMS, you go to the GC auto fit server and there's two different ones, and you've been assigned to one of them. You don't know which one it is. So you don't overload the server. So in this case, you're going to, going to download your GCMS spectra first, and then you're going to upload them to the server. So those spectra have to be taken from the cloud. And then these have been zipped for you guys. So it just keeps it simple. Didn't want to complicate it. So that means we put in the alkane standards, the blanks and the sample file all into one. So three files have been put together. And it's now in a single zip file, you're going to upload that zip file for you for us. And once you upload and say go, it's going to start processing. It should take one or two minutes. We'll see if the smoke and fire starts coming up out of the servers or not. And once it's done, then you can start, you know, profiling. And you're going to click at this stage. It's going to show you some of your alkane standards. I say you have to wait till the profiling is done. Once that profiling is completed, and you can see there's, they'll say complete. If it's not completed, they'll say profile. You can download some of the results in a CSV file. So here's an example of what the CSV file is. And so it's names metabolites. But you can also view it through the spectrum viewer. This also uses J spectra view, but for GCMS. And so this is what the annotated results will look like. So it looks a little bit like the NMR one. So there's your spectrum, but this is not an NMR spectrum. This is a GC chromatogram. And it's not where we're looking at, you know, all the peaks are integrated. That's not how an GCMS works. But the peaks are labeled, and then you'll see the list of the compound ID, compound name retention index and everything else. So you can scroll up, click again to view the spectrum. I think you can click on peaks, I hope, and it'll show you where things are. Click on the names and it should show you the position of those peaks. So that's the GCMS one. It's probably the simplest of them. And in terms of the server, it's still probably evolving a little bit. But it's in principle very similar to the NMR. They both use a lot of similar code. Now for the last one, that's LCMS targeted metabolomics. You've seen the slide already. This is outlining the protocol. Gain, this was done for you in the, in the lab, the kits for run, things are driven ties, things were processed to the LC chromatograms through the MS. And now you have the data files. So those data files are gain in this ridiculously long URL. But that's on the Google Drive that you can then download your file with your name on it. So these are zip files. So choose the one with your name, not someone else's name. So then you can go to the LC auto fit server. So unlike GCMS and the NMR one, where we had to choose one of the two. This one, everyone can use the same server. I hope. So to use that, you're going to load your file. So first download your file, and it's on your computer, run the web server. Web server wants your file. So choose that file, click here. And then load it up. And then you're going to do something that's good. Things you click which is pre process and select your quality control level. That's the first thing that does is starting of the pre processing and it's looking at all those qcs that were in the plate. And then it's going to ask you to process these calibration control level selections. So again, it's going to screen pops up and you say, it's very small there. And then you're going to say, once you've done that, then you can start profiling. So those qcs have been uploaded. You can start profiling. And just like with all the other ones, it's going to process process process. And you may have to wait a minute, you may have to wait two or three minutes, I don't know for sure, depending on the load. There's a bunch of these files being loaded up here but for you, well, I guess you'll look for your file and wait till it's completed. And as they wait till things are completed, don't press it while it's still saying processing. Once it's completed, then you can click on your calibration profiling, I guess, and you'll optimize things. So this is the case where you can spend a bit of time interactively looking at how are my calibration curves looking. And most cases they should be very linear. And because you're going to be looking at a few dozen different molecules, you can just quickly browse through some of those, see how things are looking. Here's an example of one where it didn't work out too well so for whatever reason. So this is where peak drift or some erroneous retention time was entered. So then you would have to do some manual adjustments so you can see that it's trying to integrate just on the left hand of the peak. But in fact the peak is actually all of this. So this is where the program made a mistake. And so you can adjust some of those parameters to make sure that you're integrating the full set. So here's the retention time on the left and update the max and the minimum retention time. So here the max minimum was too low. Now you've adjusted it. And after that you can submit and get the actual correct integration which will give you actual or correct concentration. So these are looking at these MRM pairs, the product ion, qualifier ion, paradigm, and the retention times. So once you're done with the concentration you have to do the integration. Once you've finished some of those fixes then you just can click on your results and download some of those results files in a zip format. So that's if you're exporting and you'll find an LC file with your metabolite names and the concentrations. Again you can open that up. It doesn't use all of J spectra viewer and the way that we saw with GC out of it and NMR. But the integration graphs and other things I think use elements of J spectra view. So again, you'll have a spectral viewer and the CSV file it allows you to look at things. So once you're done, you guys will have spent roughly an hour and a half I've given you about half an hour of this field so you've got an hour and a half now, roughly 30 minutes each. Each of you will have processed a few files for GC mess LC mess and NMR. And then you're going to go to drop the CBW drops box folder, and these are the sequence and this is the link. So you can see that it's crowdsourced science, you're analyzing all this data, sort of simultaneously, you'll have looked at by the end of this. What is it 160 200 different samples. And these things are going to be uploaded into three different batches GC LC NMR. And that data set or those three data sets are the data sets that you can use tomorrow for metabolism analyst to do things like biomarker identification. Multivariate statistics PCA PLS DA heat maps whatever else you want to do. But this will give you quantitative metabolomics data for three different disease conditions and three different systems. So the whole point here was to show that you can use targeted metabolomics methods, NMR GC mess LC mess. We're not giving you a barge scale study, we're not making you guys do all of the wet work. But it's, it's to give you a flavor of what it's like, what the fields have evolved to. If you were doing this as a, you know, masters or PhD thesis or to do a paper. There's a lot of examples and what we're doing here. And I think, particularly for the LC MS one there's usually a bit more manual work that is required. It's not quite perfect. Usually, at least with the GC mess and LC mess or the NMR, there may be a little bit of tweaking that you would do. So we've tried to give you example files that are pretty high quality. So you're not going to get spend three hours trying to figure out why this peak looks so strange. But we're trying to give you a flavor. So this is more like a sampling menu, as opposed to a full smorgasbord. I think what I wanted to also highlight with this because you'll have done some quantitative metabolomics now and this may be for some of you the very first time you've done it. I want to reemphasize some of the myth busting concepts here so that targeting metabolomics is quite a bit cheaper. It is quite a bit faster. It's quite a bit more sensitive. It tends to be more comprehensive. And it can lead to significant discoveries. And all of these projects did lead to significant discoveries. And you'll have a chance to have those eureka moments yourselves. It also means that because this stuff is quantitative, these things in some cases are already being translated to clinical practice or veterinary practice. As opposed to things that were untargeted where, yeah, it's a nice paper and then people just kind of put in the dustbin and move on to another thing. The point about targeting metabolomics is to try and get it to a point where it can be used in practical applications that other people can replicate it, other people can do it, and so that it can be applied in regular types of testing or meaningful applications.