 So, again, my name is Dave Wishart, you know, a few who just arrived. So I'll be giving a number of the presentations, especially this morning, and then later this afternoon we'll be handing over to Jeff Shaw and Nama Karoo to do some of the lab exercises. So as I said, I'm the director of the Metabolomics Innovation Center. I've been doing metabolomics actually before it had a name, and way back when we weren't sure what to call this sort of thing, we knew we were looking at chemicals and a lot of them were metabolites and we tried a whole bunch of different names. One was chemical genomics or kinomics, then we tried metabolomics and then it settled on metabolomics. So this is a bit of an introduction, partly historical, but also to give I think a number of you a bit of a background or on metabolomics. Many of you have had some experience, but there are some who I think are quite new to metabolomics, and so we have quite a range of those backgrounds and skills. And I think, as Anne pointed out, I think just to try and be courteous or patient, because we can only go as fast as sort of the slowest learners. And in many cases, if people are brand new to this, we'll have to try and get them used to some of the concepts and some of the terminology. There'll be opportunities for those of you who are at advanced levels to sort of work on your own, but I think this is just to try and see if we can get everyone up to speed. So if I could have everyone just bringing their laptops down. So typically each of our lectures will have an opening slide where we talk about sort of the learning objectives. And these are sort of outlined here, we're going to learn a little bit about what is the metabolomics, what defines a metabolome. We're going to look at some of the applications and utility of metabolomics. A lot of this particular lecture is going to focus on the operational principles of some of the major technologies used in metabolomics. So that includes liquid chromatography and gas chromatography, mass spectrometry, and NMR spectroscopy. Some of you are probably more expert in some of these than I am. But I doubt if anyone's an expert in all of these. And this is the intent is to try and give you a bit of a backgrounder in all of the technologies, because in fact, to do a good metabolomics study, you should probably use at least three of these technologies. And then we're also going to look at something or distinguish between targeted and untargeted metabolomics. So you've seen a picture of the schedule, we won't go over that again. Now we'll just sort of dive into essentially a description of where metabolomics sits in the pyramid of life. So I've used this picture many times, but I think it's a really useful one. If we can think of a pyramid at the base of the pyramid is the genome. That includes all of the genes in a given organism or cell or system. The study of the genome is genomics. Genes code for proteins. And the study of those proteins is proteomics. Proteins are intended to really either transport or catalyze the transformation of chemicals. And so the action of the proteome on those chemicals produces what we call the metabolome. I've put the metabolome at the top of the pyramid, because in fact, the metabolome is profoundly influenced by both the genome and the proteome. A single base change at the genome level can lead to a 10,000-fold change in metabolite levels. So that makes it exquisitely sensitive to what's going on, both at the genome and the proteome level. I've also illustrated, I guess, a gradient of influence as you go up the pyramid. There is increasingly greater influence of the environment. So what you're eating or have just eaten, what you're drinking, what you're breathing is actually affecting your metabolome right now. It's changing it. It changes it within seconds. Hopefully what you're eating right now is not changing your genome. Otherwise, you'll be in trouble. And somewhere in the middle, probably as you start getting fairly sleepy around 10 o'clock this morning after having eaten a breakfast, the effects of insulin and guerrilla and other proteins are starting to affect your body, which are also affecting your metabolism. But the proteome only changes slightly with respect to what your environment is. The other thing to remember is that humans and other complex organisms are not single cells. We have a physiology. We have organs. We have tissues. And in fact, many of our organs are designed to be metabolic engines with very distinct kinds of metabolism. Our liver has a distinct metabolism from our brain. Our brain has a distinct metabolism from our intestines. Our muscles have a distinct metabolism from the adipose tissue. So again, as you climb up this pyramid, in fact, you have a greater physiological influence. So in fact, the the metabolism is able to probe what's going on in physiology. Your genome, in terms of whether it's your liver or your brain, is identical. It will lead to different kinds of proteomes or different transcripts coming from. But in the end, your genome at every cell is essentially identical. So the fact that the metabolism sort of sits at this interface between the genome, the environment, and physiology makes it a great reporter of the phenotype. And this is a value that people have seen in metabolomics, is it's a great way of measuring the phenome or the phenotype of systems, whether it's cells or organisms. A definition of genomics versus metabolomics, they're both fields of life science research. They both use high throughput technologies. But in genomics, she was to try and characterize all the genes. In metabolomics, the idea is to characterize or identify all the small molecules, which can include metabolites. But increasingly, I think people are expanding it to include many, many chemicals. And you can do metabolomics or genomics on single cells. You can do it on tissues. You can do it on organisms. And in the case of metabolomics, you often do it on biofluids. Now, it's also important to have an agreed upon definition of what a metabolite is. And we've proposed this many years ago, and I guess it's sort of picked up to be a standard. This is essentially any molecule, primarily organic, although it's also possible to include metal ions. That's within an organism, but with a molecular weight of less than 1,500 Dalton's. That's an important cutoff, because many people start talking to me about metabolomics, and they're talking about proteins all the time. And I have to inform that really what you're wanting to do is proteomics, not metabolomics. Now, if you use the 1,500 Dalton cutoff, then, yes, it'll include a few short peptides and a few small oligonucleotides. But it also includes things that we think of more conventionally as metabolites, like sugars and nucleosides and nucleotides. And organic acids and aldehydes and amino acids and steroids. But because people and other mammals and amphibians and reptiles eat other things that includes components in foods, it includes for humans food additives, which give a lot of the color and aroma and taste to what we have. It also includes things we didn't intend to take in like toxins or pollutants or other things that we may deliberately or not so deliberately take in like drugs and drug metabolites. So these are xenobiotics. These are typically foreign molecules, but they also make metabolomics both interesting and challenging. In the case of humans, we also have a microbiome, a large collection of bacteria, actually 10 to 50 times more cells, bacterial cells in our body than our own cells. And microbes produce a lot of products, a lot of metabolites as well. Now, a metabolome is something that's always changing because the sensitivity of our instruments is always changing. So right now, the definition of a metabolite is something that's detectable. And what's currently detectable is between 1 and 10 picomolar. But as the instruments get more sensitive, it may be possible to detect things even at lower levels. The genome, of course, doesn't really change. It doesn't change with instrumentation. It's been defined. It's set for all ages. But the metabolome is sort of a moving target. So that collection of metabolites is what we call a metabolome. And it can be specific to a cell. It can be specific to an organ. And it can be specific to an organism. And that's really important to remember, because a lot of people mistakenly view that the metabolome is essentially all chemicals in PubChem or something like that. And that's really not the case. It's really something that is highly specific to an organism, to a species. And having that taxonomic connection is really important to doing good metabolomics research. I've mentioned this as well, that there's both xenobiotics and then conventional endogenous compounds. So we'll call them endogenous and exogenous. So the fact that humans eat other metabolomes means that we have, in some cases, exogenous compounds. The fact that we consume synthetic chemicals also means there are exogenous compounds. Some of these compounds persist for decades in our bodies. Others are there for mere microseconds, as they are transformed almost instantly. And more and more, we're realizing that there's lots of compounds that we know should be there because of basic chemistry and enzymology, but we haven't detected them. And in that regard, they're often called theoretical molecules. And in fact, there's vast numbers of the compounds that you see in your textbooks for biochemistry, which are no longer available or purchasable. So in some respects, they are also theoretical molecules because we don't have authentic standards for them anymore. Metabolome is defined by their detection technology. So very sensitive instruments will create a larger metabolome, less sensitive instruments will work with a smaller metabolome. And because of that, because of those differences with instruments and temporal persistence, the metabolome is always gonna be ill-defined. But people always ask me how big is the metabolome? And so you can come up with some rough numbers. If we look at sort of the endogenous metabolites in mammals, there's between maybe 50 or 60,000 compounds that we know of. We suspect there's probably closer to half a million that'll eventually be detected. Microbes have a slightly more complex metabolome. Any single microbe typically has a metabolome that's maybe one-tenth the size of a human. But because microbes fit in so many ecological niches and produce so many different kinds of chemicals or grow off so many different substrates, the microbial metabolome is probably larger than mammals. And then the largest and most complex of all is the plant metabolome. And the estimates in terms of the number of compounds that have been documented in the literature is about 300,000 now. And the reason why plants are so much more chemically complex than humans or microbes is that unlike microbes which can swim or humans which can run, plants can't move. And so in order to defend themselves, they wage chemical warfare to fend off pests or to cope with environmental stresses. And that's essentially why plants are so chemically rich. And this is a slide that I forgot to update, but this is a list of what we'll call a human metabolomes. This is from a year or two ago, but the numbers haven't changed too much. So right now, in terms of endogenous metabolites, which is tracked in a database called the Human Metabolome Database, is about 20,000. They range from concentrations in pico-even femtomolar ranges up to almost molar ranges, such as urea in urine, which is up to sometimes 100 millimolar or more. In a narrower concentration range, you'll also find drugs and food additives about roughly the same levels. And these represent things that you eat, things that you consume, whether it's over-the-counter medications or whether it's candies with their various color, flavor additives. And the numbers are quite large in case of food additives and known phytochemicals that are in the body. There's about 25 to 28,000 compounds. In terms of the number of drugs that can be found, they're known. It's now actually a little over 2,100. The list of drugs is maintained in a database called Drug Bank. The list of food components and chemicals in a database called Food DB. There are also drug metabolites. There are also food metabolites. Drug metabolites are also tracked and there's about 1,200 that are known now. At much lower concentrations, and hopefully, because otherwise we'd all be in trouble, are typically the toxic compounds, the pesticides and herbicides and pollutants that ideally are not supposed to be in our food or in our water, but actually are. And so those are generally at sub-nanomolar levels in healthy people. And we track those in a database called the Toxic Exposome Database or T3DB. And right now there's about 3,700 compounds in that particular database. So these are different things that are in the body. Some from food, some from drugs, some from inadvertent pollutants and others produced through endogenous metabolism that we learn about in biochemistry. They have considerable ranges in terms of their concentrations. And there are certain classes that sort of fit based on that. And the people interested in these metabolomes are also different. People interested in toxic compounds, toxicologists, environmental chemists. People interested in drugs and drug metabolites or pharmaceutical chemists and medicinal chemists. So people interested in food additives and phytochemicals or typically nutritional chemists or phytochemists. And the people interested in endogenous metabolites could be just about anyone. Now those are the numbers of compounds that we know. Ones where we have clear structures, where we have records in the literature, references that say they're there. But then as I said, there's also a large number of compounds that people believe have to be there based on basic chemistry and enzymology, based on other pieces of evidence that have been gathered. So many of them are at lower concentrations, nano and picomolar. But there are many, many lipids. And so the assumption now is there could be up to 100,000 different lipids in mammals and humans. We know that given that there's 2000 drugs, many drugs are converted into anywhere from three to 10 different metabolites. So we could say five times 10, about 10,000 different drug metabolites. Food, food products also are metabolized. So after about 25,000 known foods, if you multiply that by four or five, we'd end up with another 100,000 food metabolites. If you want metabolites of metabolites. And then there's just other ones where our endogenous metabolome is modified by promiscuous enzyme reactions. And that could also lead to anywhere from 10,000 to another 100,000 or so different chemicals. So these theoretical metabolomes may add up anywhere to 250 to 500,000 compounds. So these are yet to be found. And this is where metabolomics is more challenging than proteomics or genomics because there's a huge space of unknowns, the dark matter of the metabolome as we call it. So does anyone have any questions about what I've talked about so far? Are you able to hear me all right? And okay. Yes. Can you elaborate more on the toxic chemicals? Have they all considered kinds of metabolites? Because I assume they would be complex chemicals. That's right. So within the toxins, and let's just say, a pesticide or herbicide, which is intended to kill off things, they're not generally supposed to be highly toxic to humans, but some are. So some are specific to plants. Glycophosphates or Glycophosphates are used to kill off plants, but they're not that particularly harmful for humans. Although at high levels, they appear to have some carcinogenic capacity. Many of them are then transformed in the body to produce secondary metabolites, some of which are less toxic, some of which are more. Example would be a drug like acetaminophen, which is relatively non-toxic, but when it's metabolized, produces very toxic liver toxins. And these are again where transformations happen, partly through the liver, partly through the gut that can change things from toxic to non-toxic or from non-toxic to toxic. All right, so we're gonna dive into as to why metabolomics is important. And obviously many of you are here because you believe it's important. Here are some statistics you can use to maybe help convince your supervisor or your friends that you're actually doing important work. So in the case of a medical field, more than 95% of the common clinical diagnostic assays are actually tests for small molecules. And that's something that people generally aren't aware of, but even if you think about it, when you go to the doctor's office, as they ever do a urine test or blood test, almost all of the things that they're going to measure are actually small molecules. Almost 90% of all known drugs are small molecules. And even though the focus is on new antibody or protein therapeutics, the proportion over the last 30 years has still been the same, about 90% of all drugs are still small molecules. When people think about or try and discover new drugs, they usually work from existing metabolites. And many of the chemical features about metabolites are very, very similar to the most successful drugs, whether it's their solubility or their rule of five structures, metabolites in many respects are the inspiration for most of today's drugs. About 30% of the identified genetic disorders that are listed in databases like OMIM involve diseases of small molecule metabolism. So roughly a third of genetic disorders are related to metabolites. And then, I mean, we might say that proteins are the engines of life, but small molecules are the fuel of life. They also serve as the cofactors and the signaling molecules for thousands of protein interactions. And this role of signaling is one of the, I guess the other great neglected components of metabolomics. Really that the role of most metabolites is to mediate biology. And while we tend to focus on proteins as being the mediators, in fact, in most cases, it's the small molecules. Another reason my metabolomics is important is because metabolites are the canaries of the genome, just like the canary in the coal mine, which was used to tell coal miners that there was a gas leak when the canary stopped singing. In this case, metabolites, because they're so sensitive to single base changes or mutations, they are exquisitely useful in its as readouts of what's going on in the genome or the proteome. The time sensitivity of metabolomics is both useful and sometimes challenging. I brought this up before as in terms of what you're eating, what you drink, it rapidly changes your metabolism, but it changes many, many components and all of these are affected by each other. On the other hand, the proteome is only subtly changed and it takes often hours for it to change in a substantial way. And then in most cases, you hope that the genome never changes over time. In terms of catabolism and anabolism, metabolism is understood. There are wall charts that some of you may have seen and textbooks that describe all of the reactions that take place in living cells. You can't draw this kind of network diagram for genes or proteins, but this has been around for more than 50 years in terms of metabolism. I think those of you who are interested in system biology also understand that metabolome is highly connected. It connects to the proteome. It connects to the genome. And in terms of that connection, you can think of it whether it's the small molecules from the nucleotides and nucleosides, which are the constituents of the genome and the transcriptome. The amino acids, which are the primary constituents of the proteome. The lipids, which give cells their shape and integrity and structure and represent the lipidome. And then as the fuel for most cellular activities, whether it's ATP or sugars or lipids, they're the source for that fuel. The cofactors and signaling molecules. And in fact, I think there's an evolving views that really the genome and the proteome as they were generated a billion or more years ago, essentially evolved to catalyze chemistry. And so they're just facilitating the chemistry that could have or might have happened very slowly without genes or proteins. So the driver of evolution has really been the chemicals and then the passengers have largely been proteins and genes. So we tend to think of genomics, proteomics and metabolomics as separate fields. But they aren't and they shouldn't be. And I think one of the tools that helps bring these together is bioinformatics. Or because metabolism involves a lot of chemistry, we'll also use tools called cheminformatics tools. But using these computational tools, which hopefully will give you a chance to test and try, it essentially allows you to do systems biology. There are lots of applications for metabolomics and when you guys were describing some of the research you're doing, I think we covered many of these things. So metabolomics originally evolved to help and still is involved in genetic disease testing. Tabalomics is used widely. Anyone under the age of 25 has had a metabolomics test. Most of you wouldn't know it because you were only newborns. But that is the process. That's newborn screening. So given how pervasive metabolomics is, it's surprising that hardly anyone knows about it. On the other hand, I guess you could say, how many people have had a genomics test? Has anyone ever taken a genome test? One, one. So even though it's popular, or you can do this through 23andMe and other things like Ancestry.C, it's still something that metabolomics has essentially a leg up on. Nutritional analysis, blood and urine analysis, cholesterol testing, these are all examples of sort of metabolomic tests. Chemical shift imaging, magnetic resonance spectroscopy. Tox testing uses a lot of metabolomics. Many clinical trials use components of metabolomics. Production of beer and wine, or even just a production of many chemicals through fermentation, again use metabolomics. Water quality testing is also involving environmental metabolomics. And some of the very, very first applications of metabolomics are actually doing, we're done on petrochemical and oil analysis, which is incredibly complex. So those are some examples of applications and those are some reasons why metabolomics is important. So now we're gonna dive into metabolomics methods. And we're gonna talk about some of the technology. And again, some of you are very expert in some of these technologies, but some of you probably have never heard of each other. The intent really is to give you a flavor of the diversity of tools that need to be used in metabolomics and to give you a little bit of background on how some of these tools work. So the general workflow for metabolomics is to start with some kind of biological sample. Could be living tissues, of course it could be soil or water as well. And if it's a solid, typically we extract it. We'll homogenize it with the intent of sort of converting it into some kind of liquid. Another way of getting a liquid is to actually just sort of, to tap a tree to get it sap or to get urine or to get blood. That's another way of essentially getting through a couple of steps a little faster. You can also do cellular growth, then you can take the cells or you can take the growth media. That's also possible. So once you have an extract or a fluid, then you can start doing chemistry on it. And the chemistry of the chemical analysis is actually relatively conventional. It's been around for decades, whether it's GCMS or LCMS or NMR. Traditionally they are intended to analyze single pure compounds. And so where metabolomics sort of turn the field upside down is you're looking at mixtures primarily. Some of those mixtures are simplified through chromatography, other times they aren't. The major breakthrough in metabolomics has really happened in the last 10 or 15 years, which is this last step, which is how do you take those mixtures spectra and actually analyze it? How do you have the tools, the software, the databases to make sense of these mixtures? And that's where metabolomics has sort of hit its stride. Now metabolomics still isn't a solved problem and when you compare it to other omics techniques, it's kind of feeble. So in the case of genomics, we routinely sequence human genomes and we can measure and identify all the 22,000 genes. In proteomics, it's possible, yes. In the previous slide, you showed the LCMS or GCMS, but there is some plate-based technology that is called biograde. How they separate things? Like they are using same sampling, but they're getting different compounds. There is no separation on the steps. So tools like biocardies, they sometimes will do direct injection mass spectrometry and so they're just separating based on mass. The newer versions of the biocardies kits or plates actually have an LC column as part of their step. They also do some chemical derivatization to also help with the separation of the compounds by mass. NMR, we don't use chromatography either, it's just putting the sample in and just collecting the spectra of the mixture. So the different technologies will sometimes use separation, others don't. So just going back to this point about, I guess the level of coverage or the completeness, the fact that genomics technologies evolve so quickly, it is possible to measure thousands of genes relatively routinely. Proteomics is evolving so between five and 10,000 proteins can be measured. And in the case of metabolomics, well it is possible to measure 10,000 or more features. Most people are considered fortunate if they can measure or identify more than about 200 chemicals. So there's this diminishing set of returns and this completeness when we talk about the metabolome being 50, 60,000 compounds measuring 200 is not great. And this is I think a constant knock about metabolomics and it's a legitimate one. And it's one of the areas where there's active research going on to try and change that number from 200 to 2,000 or even 20,000. So why is it that metabolomics does so poorly at least relative to other omics? And this has to do with essentially the chemical complexity or diversity. The genome is defined by four bases. It's just four chemicals which are essentially very, very similar. And so if you know how to do the chemistry on one, you can do it on all four and because it's highly repetitive or polymeric, genomics is chemically trivial. Same with the proteome, there's just 20 amino acids. And again, they form different polymers, there's incredible diversity in the sequences but it's just the chemistry of 20 different similar compounds, what you're called amino acids. Once the chemistry for cleaving amino acids or peptides was known, then it became pretty simple to sort of do protein analysis. In the case of metabolome, there's anywhere from 200 to 300,000 different chemicals. There are amino acids, there are bases, there are sugars, there are aldehydes. If you think of it in broad terms, there's about 4,300 major chemical classes of which may contain anywhere from 10 to 100 different compounds. So the chemical diversity in metabolomics is profound and this is why it's fundamentally more difficult than proteomics and genomics. It also means that you have to use a much wider range of technologies. So for DNA sequencing, a single machine is all you need. Some kind of Illumina sequencer. And in the case of proteomics, yeah, a single mass spectrometer is adequate for a proteomics lab to function. Yes, you may need an LC system but in the case of metabolomics, people use everything in the kitchen sink. They'll use Fourier transform infrared spectroscopy, they'll use GCMS, they'll use X-ray crystallography, NMR spectroscopy is common. A whole range of different types of mass spectrometers, ICPMS, LCMS, FTMS, triple-quad-S, capillary electrophoresis is commonly used in metabolomics. Many microfluidic systems are being developed from metabolomics. And U-PLC and HPLC and UH-PLC systems are also used. So to have a comprehensive metabolomics lab, this is what you need. So we're gonna talk about some of these things. So we're first gonna dive into chromatography. And this is sort of the silent partner in much of metabolomics. And it's absolutely critical. This is what takes your mixture and simplifies it. So you're essentially taking a mixture and passing that or put that mixture in what we call a mobile phase and pass it through a stationary phase. Stationary phase is essentially the column matrix. Could be a powder, could be a plastic, could be anything. But that stationary separates the components, the analytes from each other. The process or the technique is differential partitioning. So chromatography's been around since the days of paper chromatography for more than a hundred years. You can run it through columns, you can run it through plates with thin layer chromatography. You can have gas and liquid chromatography. You can have affinity and ion exchange, size exclusion, verse phase, normal phase, the whole range. It's an amazing technology. It's always evolving. But most people today, whether it's in the case of small molecules or peptides, even oligonucleotides and proteins, use liquid chromatography. And they use high pressure or high performance liquid chromatography. So that's been around for more than 40 years. And the concept is to use high pressures, 6,000 pounds per square inch, and really, really small particles, about five microns or three to five microns. When HPLC was developed, it was a revolution for chromatography, it allowed compounds to be separated exquisitely, also to detect things at very low levels. And to, with the different types of columns, to separations of both polar and nonpolar. So with HPLC, you've got reverse phase and normal phase and helic, or hydrophobic interaction liquid chromatography. So for reverse phase, this is essentially to separate nonpolar molecules. This is the one that's primarily used in metabolomics, in part because you get great separations. And quite consistent ones. There's normal phase, which is quite rare, not commonly used. It was intended initially for nonpolar molecules, but then has largely been replaced by helic, or HILIC. And in helic, basically, you have a polar stationary phase. That means that the column matrix is polar. And then the mobile phase is a mixture of solvents that have both polar and nonpolar. Columns in HPLC, you have to be able to tolerate high pressures. Glass historically was used, but no longer. Stainless steel is primarily used. It's also polymer peak columns that can be used. They're analytical columns and they're preparative columns. Again, in metabolomics, almost all of it's done through analytical columns. These are generally quite small, few millimeters or centimeter at most in diameter. And the length for most HPLC columns is maybe five to 20 centimeters. Now, if you're to look inside a column or break it accidentally, you'll see all kinds of what looks like sand falling out. If you could look under a microscope, you'll actually see that these are silica beads. They're very porous. They're also very round. And they're also chemically activated. And in the case of reverse phase, you'll often call or hear the columns as a C8 to C18, C4. This represents the length of the carbon chains that are attached to the silica beads. So on the big picture, you can see a silica bead with a C18 or 18-carbon alkane chain sticking out of it. So this is a greasy silica bead. But there are different kinds of modifications, chemical modifications you can do on those beads and some of them are shown below, which are barely visible to me actually because the picture is so small. But some of them might have these biphenyl-like structures. Some will be slightly hydrophilic, some are aromatic. And the intent of changing the chemistry on the beads is to sort of change the affinity or the separation. Aromatic molecules will be attracted to other aromatic molecules. Linear polymers will attract or dissolve other linear polymers. Now, when these beads are all packed into your column, you run your mobile phase through them, but the separation efficiency will vary. It will vary with the length of the column. It will also vary with the size of the beads in the column. So a short five-centimeter column will generally give you a modest separation. A longer column, which will take longer to run, will produce a much nicer separation. If you shrink your beads from five microns to 1.7 microns, you'll also get really nice separations over a short column. This is the basis to UPLC. Now, when you shrink your beads, you have to apply much more pressure, and that's why it's ultra-high pressure liquid chromatography or UPLC. So if we could get really, really tiny beads, then we can have very fast separations over very short columns, and that's the advantage of UPLC. With HPLC, typically they'll have a solvent, that's your mobile phase, you'll have a pump, which moves the liquid through, but also applies the pressure. You'll introduce your sample sort of simultaneously into the liquid, and you'll get your separation, and you'll have a detector. It could be a fluorescence detector, a UV detector. The vapor light scattering detector or a mass spec, either or any, could be used, and that essentially allows you to generate a chromatogram. Now, many cases in metabolomics and in proteomics, people will use gradients where they'll have two or three different solvents, a binary for two, ternary for three, even quaternary, where they'll have up to four different solvents that are introduced in a programmed way and mixed to lead to even better separations. And that's where there's a lot of trial and error and a lot of experimentation. But with chromatography, what you'll typically get is you can start with a really complex mixture, and you'll end up with something like this. This is a chromatogram. And typically, an HPLC run will take between 30 and 50 minutes, and this is often the slowest phase of any metabolomics experiment. And the intent is, as these things run off, either your mass spec is then collecting information of each of those peaks. Now, typically, under those peaks, there could be anywhere from one to 20 different compounds. So that's liquid chromatography. There's a technique called gas chromatography as well. And this, unlike in, say, proteomics, is commonly used. It's a very powerful technique. And instead of using liquid, you're using gas. But in many respects, you'll have a similar detector. In many cases, it's a mass spectrometer at the end that allows you to detect things. For gas chromatography to work, you have to have your compound vaporized or volatilized. It has to be a gas. It's injected into a column, and in terms of the mobile phase, it's not a liquid, like a seed of nitrile and water, it's a gas as well. It's an inert gas. It could be helium, it could be argon, even nitrogen. The column itself is very, very long. So instead of five centimeters, it's up to 10 meters in gas chromatography. And instead of being maybe one centimeter across, it's only a couple millimeters across. The columns are metal, but on the inside, there's various polymers that are attached. Now, to get gas chromatography to work, because most compounds are not volatile and not easily vaporized, you can volatilize them by derivatizing them with a chemical, tetramethylsilane or TMS. So unlike in liquid chromatography mass spec, where you're basically working with pristine samples, in gas chromatography, unless you're working with very volatile aromatic esters or things like that, you have to do this step of derivatization. There's often sometimes another step where there's methanolysis. So there's a chemical modification step for GCMS, but it converts, in this case, a sugar into something that is decorated with silil groups, TMS groups. So this particular sugar has gone through two steps and there's one, two, three different TMS derivatives. Now, in some cases, there could only be one, sometimes two, and this is also one of the challenges of doing chemical derivatization is that you can't often control how many derivatives are being produced. So after you've derivatized things, made them volatile because the TMS actually makes, gives things a very low boiling point. They can now be converted to gas and they're put through the column and they separate just like they would with the liquid chromatography. So there's this partitioning by interacting with the sides of the column. So with a GC system, you don't have a really long 30 meter single tube. You coil it up, that's just more efficient, but it is 10 to 15 meters long. The inside of this very narrow tube, which is only a couple of millimeters across, has outside as a plastic or sometimes metal coating, fused silica and then sort of a polysiloxane polymer on the inside and that is where the interactions happen. Now when column, when material moves through a column, whether it's liquid chromatography, gas chromatography, compounds are retained, peaks come off at different times. Those peaks in their position is called the retention time. So that's the amount of time it takes for something to pass through a column. So different compounds will have different retention times. Different columns will lead to different retention times. Different flow rates, different pressures. Carrier temperature, the carrier gas, all of those will affect it both for gas chromatography and for liquid chromatography. Now in liquid chromatography, we still use retention times, but in gas chromatography, people have decided to move towards a retention index. And this is essentially taking the retention time and normalizing it or scaling it to the retention times of a set of alkanes, a standard set of alkanes. So in many respects, gas chromatography is light years ahead of liquid chromatography because they have standardized things in a really nice way, but the separations in gas chromatography are also much better than liquid chromatography. They measure separations in terms of plate counts and this is a form of measuring the resolution and a gas chromatograph is many, many times better than a liquid chromatograph. So gas chromatograms look like HPLCs. You can measure them in terms of the retention time or retention index. You can measure the area under the peak just like with liquid chromatography to determine how much of a given substance is there. And as I said, the plate counts for GCMS are astonishing. Peaks are many times narrower. The separations are many times better. And the typical GC run will be about 30 to 50 minutes just like it is with an LC or HPLC system. So those are some of the separation tools. You can combine two GCs to make a GC-GC. You can do LC-LC as well. There's variations that can be done. Now after you've separated things, typically you want a character, yes, question. Sorry, I'm curious about when you merge things, like when you have GC-GC or a GCMS or an MS, MS, I don't know if that's the thing. Just put lots of letters together. Is it that the output of one column goes directly into the next one? Is that how that works? That's right. That's essentially right. And it's a little challenging and often you'll change different phases. So you might have a hydrophobic, so in liquid chromatography, it'll be a C18 hydrophobic reverse phase separation followed by a helic separation, which is for hydrophilic separations. So your stationary phase, sorry. Your stationary phase change. It could change. Yeah, you could change your mobile phase, but also your stationary phase has changed as well. So anyways, separations you can get are quite impressive and particularly with GC-GC. So in the end, you still have to detect this stuff and you can detect these things by UV or fluorescence, but the more useful method of detection is mass spectrometry. And essentially mass spectrometry is just a magic weighing scale. It allows you to measure molecular atomic weight of samples. All kinds of different mass spectrometers. This is a time of flight mass spec, an older one. And they have different designs, but the concept behind them all is that every compound has a different molecular weight and that can be sufficient to identify what those compounds are. So here's three different chemicals, three different molecular weights, each unique to that chemical. It's the same sort of thing as if we had everyone weighed everyone here, probably there's no one who is identical weight. So if we just have your list of your names and your weights, someone just gave me your weights. I could tell you who you are. So this is the concept of what mass spectrometry is. Precision of modern mass spectrometers is quite astonishing. Typically things can be determined, the molecular weight can be determined within one part per million. And with that precision, you can determine the molecular formula of that compound from the mass. If you're looking at proteins with that kind of one PPM accuracy, that means you're able to measure the molecular weight of proteins to about one Dalton. So one Dalton is the weight of one hydrogen atom or one atomic mass unit. So just as you pointed out, we put all these hyphens together. We hyphenate GC to GC to MS, LC to MS to MS. They have different utilities, obviously gas chromatography, liquid chromatography is essentially we're separating compounds and then measuring their mass, in tandem mass spectrometry, we're separating compounds first by their mass and then we fragment them and then separate them again by their mass or the fragment ions. Now molecules are made up of atoms and atoms have different isotopes. And there's just like there's hydrogen and deuterium or carbon 12 and carbon 13 or nitrogen 14 and nitrogen 15 or chloride 35 and chlorine 37. The result is that when you have these different isotopes, you're gonna have slightly different molecular weights. So a given compound, a given molecule will not just have one weight, it will have four or five visible weights or atomic masses. It'll have the most abundant one, in this case it's 1155.6 and that is the biggest peak that you see and that's called the monoisotopic mass. If you have a lower resolution mass spectrometer like a triple quad or Q-trap or linear ion trap, you don't see the monoisotopic mass. You don't have the resolution. So what you see is the average mass and that's kind of marked there. In this case it's 1156.3 and that is the weight of both their intensities and their masses to produce this general peak. So that's the weighted mass of all the isotopic components. So high resolution mass spectrometers will produce a spectrum here where we have four or more visible isotopic peaks. Lower resolution mass spectrometers will just produce a big mound which tells you your average mass. Now in high resolution mass spectrometers, as I say multiple peaks, you can get some pretty interesting patterns, particularly when there's things like chlorine involved because most cases isotopes like carbon-13 and deuterium are rare, but in chlorine it's about a third of all of the chlorine atoms are chlorine-37. So you can calculate based on the number of... or the isotopic abundance for these different isotopes, the number of molecular formula and the intensity of those peaks. So in this case this chlorobenzene has up to six peaks that can be detected on a conventional mass spectrometer and you can calculate those intensities and you can see that yes, the monoisotopic one is the most intense, but then there's another peak that's located two mass units away that's also pretty intense. And so this is what you see for chlorobenzene. If you had a pure collection, you'll see yes, the big peak, that's the monoisotopic one, but then the second peak which is further up and the intensities are given in scaled terms. So yes, a single molecule can have multiple weights whereas humans just sort of have one weight. Now in the case of mass spec, there's some general principles. Samples are fed in through the liquid chromatography or gas chromatography system. They are ionized and that's critical. In mass spectrometry we're not actually measuring the weight as we would measure our weight on a waste scale, from measuring the mass to charge. You have to charge or ionize molecules and if you can't charge or ionize the molecule you can't see it in the mass spec. That's critical, that means there's a number of compounds that are essentially invisible to mass spectrometry because they are not ionizable. Once things are ionized, then they are put through a mass analyzer. Usually it's a set of electrically charged quadruples sometimes it's magnets but they separate the particles according to the mass to charge ratio and then you'll typically have some kind of detector which is also electric and signals are picked up from the detector. So a mass spectrum looks a lot like a chromatogram. Lots of peaks but they're very, very narrow instead of measuring time at the x-axis you're measuring mass to charge. So they have sharp, narrow peaks. The mass to charge or M over Z or M over Z ratio is given. The height of the peak gives you a relative abundance. It's not useful for quantitation. It just simply is a measure of the ionizability of that particular ion. So some ions fly better than others. Now in mass spec we talk about resolution or resolving power and that's essentially how narrow are those peaks. Just like within chromatography we call a plate count is how narrow are the peaks. Resolution is a measure of how narrow peaks are. So narrow peaks are good. The better the resolution, the better the instrument. The more expensive it is usually. And essentially that resolving power is defined by the mass and the width of that mass peak. So how narrow it is. So if we're looking at this, this is essentially looking at that delta M at 50% or the delta M at 5% but this is the width of the peak. And if the peaks are very, very close then there reaches a point where you can't really resolve those two peaks. And so at the bottom there you can see that those two peaks are probably overlapping when they sum together they would look like almost a single peak. Whereas at the top those are well resolved. Obviously if these had no width at all you could resolve almost anything. So here's our real picture of a low resolution iron trap instrument. And you can see it looks like one giant mound or a cross section for a mountain. And then you can look at a higher resolution time of flight instrument. And you can see seven or eight distinct isotopic peaks. So that's the difference between a high resolution and a low resolution. High resolution we can measure the mono isotopic mass. Low resolution all we can see is the average mass. So this is an example where we're looking at different resolutions. So this is a blue one which has a mass resolution of a thousand and it looks like one giant mound. There's a red peak which has a resolution of 3,000. But then if we look at the black or the green everything is distinct. We can see all of the mono isotopic masses and those have resolutions of anywhere from 10,000 to 30,000. So good resolution makes a big difference. And then there are certain very expensive mass spectrometers that have excellent resolution. So we've talked a little bit about mass spectrometers. We've talked also about how you get the sample into a mass spectrometer whether it's HPLC or gas chromatography. That's the inlet. The inlet then feeds into an iron source which is a way of ionizing or converting those compounds into ions. Now almost all mass spectrometers are under high vacuum. So whether it's the ionization, the mass analyzer or detector, they're all under a strong vacuum. And that's sort of shown above there. And that's usually the thing that breaks down. It's the thing that causes all kinds of pains in mass spectrometry. And in mass spectrometry, you're putting stuff in and sort of burning it up. And so that gunk starts accumulating also on your ion source and your mass analyzer. And so there's always a need to clean a mass spectrometer. So ionization is done through many different methods and people have been exploring ionization methods for more than 70, 80 years. There's the classic method which is called the electron ionization or electron impact. It's called the hard ionization method. It's ideal for very small molecules and it fractures and shatters molecules quite distinctly. And it's primarily used in GCMS. There's chemical ionization. That's a semi-hard method. It doesn't use electrons, but it uses chemicals to ionize or collide with other molecules. It's not as common as EI or as ESI. So electrospray ionization is a very soft ionization. It often doesn't even break up the molecule. It leaves it intact. It's been used for proteomics for decades and instead of having a mass limit of 1,000, you can go up to 200,000 Dalton's. Another ionization method is called MALDI or Matrix-Assisted Laser Disruption. And this is where you use lasers to vaporize molecules and to ionize them. And these are generally used to look at much larger molecules like proteins and DNA and up to even a mega Dalton can be measured by MALDI systems. All of these, as I say, are trying to convert a molecule into an ion so that it can be sent into a mass spectrometer. So electron impact or electron ionization, you send electrons into the system and they collide with the molecules that are coming off from your gas chromatogram. So there's an inlet. There's this vapor that's being diffused into this box. And then there's this filament, some kind of tungsten or rhenium filament. And it will generate electrons, just like your old cathode ray tube for the TV would generate electrons and hit the surface to generate an image. These hit the gas-phase molecules, fragment them, and then they're sent through a series of electrodes because now they're charged so they can be sent off into the analyzer. So the sample goes into the instrument. It's bombarded with those electrons at a precise energy, 70 electron volts. That energy fragments all of the bonds because the bonds in most molecules are anywhere from five to 10 electron volts. And those are sent to the mass analyzer. So if you send in some kind of volatile molecule like methanol and you fragment it, you're gonna get all kinds of tiny fragments and the ones that are charged like the CH3 charge or the CH2OH charge or the CHO triple bond H, those are ions that you're gonna see. So this is an example of an electron impact or electron ionization of methanol. And you can see some peaks at 32, 31, 29, and 15. So this fragmentation is actually predictable. And for very skilled mass spectrometrists, they can actually look at these spectra and identify what those fragments are. There's not too many of those people still around, but it was something that analytical chemists often had to learn to do. So that's a hard ionization method. Then there are the soft ionization methods. Maldi, which uses a UV laser and uses a matrix made of cyanohydroxycinematic acid or other variants. You can also use polyphenols as matrices. And they vaporize the molecule. There's essentially heating it up and causing little explosions and molecules fly off or alternately send things through the equivalent as a charged aerosol can, which is ESI. And you use essentially a high voltage, send the liquid through and it sprays out. So this is an example of sort of the details of the spraying that's typically done with electrospray. Again, a series of charged plates that are in front of it. As things spray from the nozzle, the droplets start evaporating because it's in a very, very high vacuum. So the liquid evaporates and condenses. And then the charges that are in that liquid actually lead to even further shattering because you can't bring charges any closer. So they split up to even smaller droplets. So eventually the droplets just have a single ion in them. And those are the ones that are ultimately detected in electrospray ionization. So to do electrospray ionization, you usually put a volatile buffer. Could be things like acetonitrile water. And you'll pump them through a very thin tube, a capillary at a very low rate of microliters per minute. Strong three to 4,000 kilovolt voltage across the nozzle. It helps aerosolize things. And then that high vacuum rapid evaporation allows these things to fragment or shatter even further into tiny single ion droplets. And so you can get the ionization to happen depending on both the volatility of the solvent and its viscosity. So whether you're using acetonitrile in water, if there's more water, you have to use higher voltages. If there's more acetonitrile, you can use lower voltages. You can go to nanospray systems, which are very popular, where you're working with even less than 10 microliters even down to one or less than one microlitre per minute. In electrospray, you will have either positive ion measurements or negative ion measurements. And this depends on the solvent or solute you add. So if you add, say, formic acid, a small amount to your acetonitrile water, you can measure things in the positive ion. If you put in ammonia to your solvent, then you'll be able to measure things in the negative ion mode. And there's different additives that people will use. Electrospray ionization is best done without salts and without detergents. So you wanna have as clean a system as possible. So ESI is not as well standardized as electron impact ionization. So people use different approaches, different ionization energies, which will lead to sometimes very different spectrum. So after you've ionized things, you have to run your ions through a mass analyzer. And this is essentially the thing that separates its form of chromatography if you want, but it separates ions based on their charge. So the original mass spectrometers use magnets. These are called magnetic sector analyzers or MSI. They're still used today and they're very, very high resolution, but they're expensive and they weigh a lot. And there's a lower cost alternative, the quadrupole. So this is using electric fields for metal rods. That's the quadrupole. And they're usually called Q instruments. The term Q means quadrupole. So a triple quad means a triple Q. That's three quadrupole sets in a row. There's a Q-toth meaning a quadrupole coupled to a time of flight analyzer. And then they can also couple all kinds that there can be toff toff. There's two time of flight instruments in tandem. So higher resolution mass spectrometers are things like time of flight, which measures how long it takes for an ion to move through a flight tube, which might be two to three meters long. There are orbit traps, also very high resolution mass spectrometers. It's a different principle, but basically, ions start spinning around. Similar to ion cyclotron resonance, or FTICR instruments. Again, they'll use a magnetic field to spin ions around and measuring how long it takes for them to go through this cyclotron, which gives you very, very precise mass measurements. The basis of using how I can say the particle error is like, who is the top system or Q-toth, Q-system is necessary, how I can understand it. Which one would you want to use, you mean? Or, so it's basically defined on your budget. And so if you've got a lot of money, you might as well get the best, highest resolution instrument. They are giving the, for everything, but it depends on the regulations. Yeah, so the most robust mass spectrometer, is a triple quad. It's, never breaks down. It's widely used in clinics, environmental testing labs. It's primarily used for targeted analysis, and it has relatively low resolution, about 0.1 Dalton, 0.2 Dalton's. So it's the most popular instrument in the world. If you're wanting a high resolution instrument, and your budget isn't infinite, most people will either get a Q-toth, because it's a higher throughput instrument, which gives high resolution. Whereas they can afford it, they'll get an Orbitrap, which is a slightly lower throughput instrument, which gives you better resolution than a TOF. And that's sort of illustrated here. So this is this mass accuracy. If you have an infinite budget, you would get an FTICR, Ion Cyclotron Resonance Mass Spectrometer. And those are the very highest resolution mass spectrometers. They're also very slow. So there's different compromises that people will work with. And depending on some cases, the type of work they do, many people doing targeted mass spectrometry will use a triple quad or Q-trap or linear ion-trapped instrument. People doing untargeted mass metabolomics will use a Q-toth or an Orbitrap. So those are just general rules. But people can do targeted mass spectrometry with a high resolution mass spec as well. It's up to sort of the different circumstances that you might have. But you can even use biocrities on an Orbitrap if you want, which is a targeted QIT system. Okay, so as I said before, that the result from a mass spectrum analysis is similar to what a chromatogram looks like from HPLC or UPLC. You can think of it as sort of the x-axis corresponding to retention time because things are coming off from your liquid or gas chromatography, and then the y is a signal intensity. So there's three different types of mass chromatograms that can be produced. So there's a total ion current chromatogram, a TIC or TIC, a base peak chromatogram, a VPC, or an extracted ion chromatogram. And people rarely, well, I don't know, depends on complex. People will show either a total ion chromat, current one or base peak chromatogram. And it looks just like an HPLC system. You're seeing lots of peaks, and instead of measuring the UV absorbance, you're basically measuring the mass and the masses and the intensity of those ions as they hit the detector. You can also extract out, computationally, just one or more analytes from the TIC or VPC, which is shown on the far right. So the total ion or TIC chromatogram is red, base peak one, which flattens things out and has a baseline brought down to near zero is in blue. And then the one that looks like it's just a single peak, that's the extracted ion chromatogram, which contains just typically a single molecule, maybe two, and again, that's sort of computationally extracted. So this is an example of what would look like just like a GC chromatogram or an HPLC chromatogram, but this is now where we're measuring the masses and this would be a base peak chromatogram. So I'm gonna switch over from mass spectrometry to another technique called NMR spectroscopy. So maybe before I do this, I just wanna, how many people have used GCMS in their metabolomics work? One, two, three, four, five, six, seven, eight, nine, 10, 11. Okay, about half of you. And how many people have used LCMS in any of the metabolomics? So about the same number. How many people have used NMR spectroscopy in their metabolomics? One, two, three, four, five. So this is typical, but it is in fact where metabolomics actually got its start. The very first metabolomic studies were actually done using NMR spectroscopy. So in NMR, a sample is put into a big magnet and you collect again what looks like a chromatogram, but in this case, you're measuring signals, magnetic signals from the compounds that are dissolved in a liquid. So to do NMR, you basically have to put a sample in a strong magnetic field, then you have to send a pulse of radio waves into the sample and the sample will absorb those radio waves. And what you measure are essentially the absorbences of the different nuclei. So it's like a UV absorbent spectrum, except the peaks are a lot narrower and usually there's many more absorbences. So I've sort of put a color gradient there like you would see with visible or UV light, but in NMR we're seeing things in much finer detail. So NMR is a non-radioactive method. It measures nuclear magnetism. That's why we have the word nuclear. It measures the absorptions of light, but not visible or UV light. It measures radio wave light or radio frequency light. You can only get NMR phenomena to work in a high or strong magnetic field. And different nuclei absorb different frequencies. Most of the magnetic properties of molecules have to do with the protons, which have a positive charge and they spin. And any spinning charge or rotating charge creates a magnetic field. Protons can either spin, have a spin up or spin down, counterclockwise or clockwise. So the effect is that protons in the nuclei of different atoms will have essentially little mini magnets. So when you send in a radio frequency or light into an oriented sample, the light will be absorbed by those protons and it'll cause them to flip. It'll change their spin orientation. They'll go from, in some cases, a downspin to a higher energy upspin. So you pulse the sample with a radio wave, you change the energy of the sample, change the number of protons that have upspins, and then you observe them flipping back to their downspin state. And the net effect is you get this sort of oscillation that's measured over time over a couple of seconds. So it's equivalent to hitting something, a bell, and letting it ring, or multiple bells and letting them all ring. And from there, we can decompose the frequencies and determine how many bells there were and what the size of those bells were. In NMR, we have to use very big magnets. The bigger, stronger the magnet, it's better. Typically, the magnets used today can pick up a city bus. And you're going to be measuring different frequencies, just like with light energy, low frequency, red light, high frequency, blue, or ultraviolet light. Bigger magnets lead to higher frequency measurements. And so a big magnet is like the equivalent of an orbitrap mass spectrometer, and a little magnet is equivalent to a single quad mass spectrometer. In NMR, you'll start with samples. You're often using robotic systems. You'll take a sample, upload it, inject it into the NMR, a giant magnet, which is the silver can. You'll have radio waves sending in through a radio wave transceiver, transmitter receiver. It's like a giant ham radio. And then you have a computer to collect your spectra. The magnets are large. They're as big as a typical refrigerator, or in some cases, as large as many refrigerators stacked together. They can weigh several tons. They're cooled by liquid helium. Liquid helium is surrounded by liquid nitrogen. And inside are especially niobium tin superconducting wires that are wrapped, several kilometers of them around a core. And inside that, you will put your sample. And the sample is also put into what's called a probe. Probe has a bunch of electronics in it, zisters, and inductance systems. But basically, what it is is about a, I don't know, half a meter long tube with what's called a saddle coil, which is just a couple of little wires, which allows you to generate a radio frequency of sufficient strength to perturb the sample. So in NMR, I typically use something that's about a little tube about the size of a pencil, length of a pencil. You fill it up with about half a mill of the fluid. And that tube is dropped into the probe here, which is in the magnet here. And you pulse away at it. You send radio frequency energy into that tube, and then you collect the sample, the signal. And when you collect the signal, this is what you get. This is an NMR spectrum. It's not mass to charge that's plotted out. It's radio frequency in megahertz, or they convert that to parts per million. Those peaks then have chemical shifts. They also have very characteristic splitting patterns, and they have different intensities. Intensities are related to the number of hydrogens or protons. Splitting patterns come from what's called spin coupling. And the chemical shifts are characteristic of each of the nuclei. Chemical shifts are the key, equivalent to mass to charge and mass spectrometry. But chemical shifts are unique to each molecule. Different hydrogens in different positions in a molecule absorb at different frequencies. And just about every compound has a unique chemical shift fingerprint. They're affected by the charge or electronectivity of neighboring atoms and bonds. And anyone doing an NMR spectroscopy sort of has to memorize these sorts of tables where they look for the very characteristic chemical shift patterns, depending on whether something is a carboxylic acid or an aldehyde or it's part of an aromatic group, whether it's attached to an amine or whether it's in a saturated or an unsaturated methyl or methylene group. So those chemical shifts are sufficient to determine structure. And so in fact, NMR is still the gold standard for determining the structures of all molecules. Mass spectrometry can get you some way there. But if you want to officially determine what your structure is, you actually have to use NMR spectroscopy. So NMR spectra, I say this is a very characteristic group. You can see there's a triplet corresponding to the CH3. And then there's a quartet of four peaks corresponding to the CH2. And this has to do with the couplings, the N plus 1 rule of the neighboring hydrogens. People assign NMR spectra. So this is a pure compound. They have a reference, which is called TMS, which is the same thing used to derivatize with GCMS. It's added to NMR. So there's a nice way of reusing reagents at chemistry. But this is just highlighting the positions of the aromatic hydrogens, the methyl and methylene groups. NMR spectra, when they come off an NMR spectrometer, they kind of look warped and distorted. They have to be fixed. They have to have baseline correction. They have to have the water signal suppressed. They have to be phased. They have to have a reference. All of that is done often manually with NMR. So it's a little bit of an art. So with just like retention indices and retention time where you normalize with chemical shift, people will normalize. They calibrate to an internal standard, which is TMS or DSS. They'll optimize the magnet strength and magnet shape to get the shims so that the magnets and the spectra look clean. You'll phase signals so that things will all be pointing upwards instead of half up and half down. And because everything you do in NMR for metabolomics is done in water, you have to get rid of this giant signal of 110 molar water or hydrogen, especially when their signals are only about 20 or 30 micromolar. And then there's a baseline. Just like with mass spectroscopy, you'll see baselines that have warped. Same thing you'll have to do with NMR. So this is an NMR spectrum. Looks a lot like a chromatogram. LC, GC looks like a mass spec chromatogram. So they all look the same in many respects. You just have to figure out what the peaks are. So there's different technologies that we've looked at. We've looked at NMR. We've looked at GCMS, LCMS. We've looked at quadruples and time of flight. What you'll find is that each instrument type has different levels of sensitivity. LCMS is the most sensitive. You can measure things down to nanomolar or even picomolar concentrations. GCMS, GCTOF, next most sensitive. It's often useful for things in the nanomolar to micromolar range. The least sensitive is NMR, which is typically good from micromolar and more. So that might reflect whether there's fewer people doing NMR. But there's also another reason why people still use NMR in metabolomics. And that's because typically what you're measuring in NMR are the compounds that are known. Everything that you see in an NMR spectrum is identifiable. And with NMR, you typically identify between 50 and 100 compounds routinely. In GCMS, you might be able to identify 200 compounds or more, of which maybe half are identifiable. With LCMS, it's possible to measure 5,000 or 10,000 features of which only 200 compounds are identifiable. You get into the realm in LCMS of what we call the unknown. So in many cases, what you might see is a really cool signal. It's really important that differentiates the controls from the cases or whatever you're wanting to distinguish. But you can't figure out what the compound is. So you can't publish. In NMR, because everything is known, you can see the signal. And if it is differentiable, then you can publish. And so that's why still NMR is widely used. The other thing is that the tools are complementary. A lot of the compounds you detect by NMR, you can't detect by LC or GCMS and vice versa. Many of the compounds that people look at GCMS, the volatile ones, you're not going to see by NMR or LCMS. Again, it's a complementary technique. So using all three actually gives you a much broader window. And again, that's why all three are still used. This just sort of highlights these differences, whether it's LC or direct injection mass spec, GCMS or NMR. The types of metabolites they typically are used, whether they're hydrophobic, hydrophilic, water-soluble, types of samples that are used, and the types of sample volumes. You can see certainly with mass spec or LCMS, you can get away with very small sample volumes, whereas with NMR, sample volumes are up to 10 to 20 times more. Some methods are slow. Some methods are fast. With NMR, there's an advantage in that you don't have to actually do any separations. You can just drop things right in. You can collect samples relatively quickly with NMR, depending on the type, whereas depending on the type of mass spectrometry you're using, it can take a little time or a long time. Eliminative detection, as I said, with NMR is like five micromolar, with LCMS it's five nanomolar. The number of metabolites that can be identified is not substantially different between the methods. That's because in NMR, you're dealing with the knowns and in mass spec, you're dealing with lots of unknowns. The other thing, as I said, they're complementary, so they rarely completely overlap. There's about 10 to 20 to 30% that overlap between the different techniques, so they offer a way of not only seeing more of the metabolome, but also, in some cases, cross-checking what you're actually measuring. So this is just a summary of what's possible with the different types of metabolomic techniques with NMR versus GCMS versus LCMS versus elipidomics, versus sort of the specialty methods that people use these days. So some are more sensitive, but in the end, about the same number of metabolites are ultimately identified and or quantified. So just to wrap up for the last, maybe, few minutes here, I know some of you would like to have your coffee break and get recharged. Basically, there are two routes to metabolomics. There's a targeted or quantitative metabolomics, and then there's the untargeted method to metabolomics or what we'll call a more chemometric method. Both are widely used. Our focus here today is targeted, and as a heads up, most of the world is heading towards targeted metabolomics. And there's several reasons for that. One, the number of compounds that people can measure now is increasing, so in the old days, it wasn't very many, but now people are getting 500, 600, 700 compounds through targeted metabolomics. Second reason why targeted metabolomics is picking up steam is because it's quantitative, and it means it's comparable. Absolute quantitation has been the Achilles' heel to just about every field of omic study. It's been a killer for proteomics. It's been a killer for transcriptomics. So if you can get absolute quantitation, you can compare results from lab to lab, group to group, country to country. So those are among the reasons why targeted metabolomics is picking up steam and why it's becoming very popular. There are many reasons to pursue untargeted metabolomics. It still will be done, it'll always be done. It's a great tool for discovery, for novel compounds, for novel ideas. So in untargeted, typically you'll have lots of samples. You will have to collect many samples for untargeted metabolomics. The samples are generally not assigned or annotated. They're just the peaks are identified. The peaks are then compared through multivariate statistics and the most significantly different peaks or features are the ones that are then characterized. So you deal with maybe thousands of features. The data reduction reduces it to dozens of features that are interesting and then you kind of cross your fingers and hope that those dozens of interesting features are identifiable. So the identification phase is saved to the end in untargeted metabolomics. In targeted metabolomics, the samples are collected, metabolite identification and quantification is done in the first step. Once you have your lists of metabolites and their identities and their concentrations, then you do the data reduction to identify which ones were the most significant or important and then you go on to the biological interpretation. So in that regard, targeted compound identification is the first thing, untargeted compound identification and quantification is the last thing. So basically what you're doing, whether it's untargeted or targeted, you're still going from spectra, NMR, GC, LCMS spectra and you're trying to convert them to lists, lists of compounds, lists of relative or absolute concentrations or intensities. Once you have those lists, then you can start doing some of the cool bioinformatics, which is to going from lists to pathways or from lists and pathways to models or to biomarkers or to my biological insight. You can start integrating those things. So we're gonna look at some of the different challenges associated with going from spectra to lists. That's what we're gonna talk about mostly today. We're gonna talk about how to deal with data integrity and quality, we'll talk about some of the data reduction. Tomorrow, we'll look at how to go from lists to pathways and biomarkers. And that's when we'll be talking about Metaboanalyst. So what's the difference with absolute quantification and absolute, if you would like? Yeah, so absolute versus relative is really, I think went on a sort of emphasize, but so absolute means coming out with micro molar, millimolar concentrations, relative quantification is basically saying, the intensity is 30 for this and the intensity for this is 12. So you don't have any units. It has to be understood between labs, whether those units are some pseudo unit or something like that, but it's just simply saying one was 30 was bigger than the one that was 12. And relative quantification has been unfortunately used for too long. And it's, as I say, it's how most protein and proteomic studies are done. So you can't get absolute quantification with many proteomic studies. Jeff. So the relative quantification unit in one study, you did versus before. So that's why the unit probably do a ratio like that. Yeah, typically used as ratios for just, it's good for within one study, but not for any other study or between other.