 So we have to give this Creative Commons thing, and then this is our standard picture here for the first module. So this is really just an introduction to metabolomics. Now all of you guys should have books and you can follow along with your binders, and hopefully you'll mostly keep your laptops down. So part of this is just to give people an idea of what metabolomics is, how big different metabolomes are, some of the applications. Again, people have already mentioned some really cool ones, but also to talk about some of the technologies. And metabolomics is different than most other fields, including proteomics, and in particular transcriptomics or genomics, because there's a huge diversity of instrumentation. Many people will only learn or use one. We're trying to expand your horizons to realize in fact that the best metabolomics experiments and projects generally integrate at least two and usually three different technologies. We're also going to talk a little bit about the differences between targeted and untargeted metabolomics, something that you guys may have heard a term like that before, and we'll try and distinguish them a little bit more. Michelle's already shown you this schedule, so it's officially 9 o'clock right now, so we'll have about an hour and a half to cover this introduction to metabolomics. So one way that I usually begin presentations about metabolomics is this picture of a pyramid. I call it the pyramid of life. And it sort of shows this relationship between genomics and the genome, which is the base. So your DNA or plant DNA or microbial DNA is essentially what codes for everything that comes up above that. Typically the genome is very large, humans 22, 23,000 genes, 3.1 billion bases. Genes code for proteins. Now people who do a bit of proteomics will argue in fact the proteome is not 22,000 proteins, in humans it's 200,000, which is correct. But what we're particularly interested in are enzymes or isozymes, and these are the things that manipulate the enzymes. These are the workhorses in the cell, and there's about 5,000 or 6,000 of those in say human or mammalian cells. Those enzymes manipulate the metabolome. So you have this progression as the gene codes for protein, protein codes for metabolites if you want a study of each of them, genomics, proteomics, and metabolomics. Part of the reason why there's that pyramid is that when you count the number of critical or endogenous metabolites that are typically shown on most biochemical pathways is about 1,000 key ones. Humans have many more than that, most plants and animals have many more than that, but there's 1,000 that are critical or essential. So 20,000, 5,000, 1,000, so we'll go up in numbers. The other thing that we also have to remember is that a small change in the genome often has a pretty profound effect at the metabolome. So it ripples up just like DDT, which was sprayed in the 1960s. Insects took that in, but then it moved up the food chain and was starting to kill off owls and falcons and hawks. So that's partly this impact of how the gene affects the metabolome. But there's also two other things I've shown here, which is the physiological influence and the environmental influence. So above or at the top of the pyramid is the environment. And the environment profoundly influences the metabolome. It mildly influences the proteome, but it barely touches the genome. If it did, if the environment affected our genome, we'd all be mutants and zombies and things like that, so it just doesn't happen. Genes are highly stable things. Yes, there are epigenetic events. Yes, there are mutations, but they're rare. And the metabolome, on the other hand, is changing profoundly. You guys have just eaten breakfast. It is changing your body chemistry quite profoundly. What you're eating, breathing, drinking. What you're hearing is also changing your metabolome. There's also an element, what I call a physiological influence. And this has to do with the fact that humans, plants, most organisms are not single cells. We're a collection of trillions of cells, and these assemble into organs. And in the case of humans, we have metabolically designed organs. We have liver, which performs very specific metabolic functions. We have a stomach that performs very specific metabolic functions. We have an intestine with its gut microflora that has very specific metabolic functions. We have parts of our body that consume only glucose, mainly our brain. Other parts of our body that depend critically on fats. And the metabolome, and the metabolomes associated with the organs, or the fluids that bathe at those organs, so urine, saliva, cerebral spinal fluid, blood, all are quite different. However, if we were to look at our genes in each of these organs, they're all the same. Yes, there's maybe differences in gene expression, but they're all the same. So the metabolome is profoundly influenced by physiology, by organ structures, and organ composition. And that's important to remember, especially when we're looking at organisms that are multicellular and as complex as plants and humans and other animals. So metabolomics, I don't think I have to give this definition, but it's sort of the de facto thing that we have to talk about. So the relationship in essentially metabolomics grew from the term genomics. It was a wannabe thing. So you can use exactly the same definition that we use for genomics. It's a high throughput field of science. Technologies are used to characterize genes in cells, tissues, and organisms. Metabolomics, same words, and we've just replaced the word genes with small molecules or metabolites. It still has a requirement for high throughput technologies, and it's still typically looking at multiple systems, tissues, individual cells if we want, or entire organisms. What we study in metabolomics is the metabolome. What we study in genomics is the genome. What we study in proteomics is the proteome. One of the biggest issues, I think, still today, and I have respectable PhD scientists, professors coming up to me, asking if I can do or work on a metabolomics project with them, and then they list a whole bunch of proteins that they want me to measure. And that's because they don't know what a metabolite is. So technically, a metabolite is a small molecule. Usually it's organic. It doesn't have to be, so I might be changing that, but most of the time it's organic. And it's typically the cutoff we use is about 1,500 Dalton's. Some people use 1,000 Dalton's. We use 1,500 because there are a lot of lipids, actually, that are above 1,000 Dalton's, like 1,100, 1,200 Dalton's. So metabolites cover some very short peptides, 5,6,7,8 residues, oligonucleotides, 2,3,4 sugars, carbohydrates, but also a lot of the other things that we think of as metabolites, so amino acids and ketones and aldehydes, steroids. It also includes a lot of things that we eat or that other animals eat, so that includes compositions of food, food additives, includes toxins that we didn't intend to eat, pollutants that we didn't intend to inhale or drink. Also includes drugs. So there's both an exogenous and an endogenous metabolome. The endogenous is the one that would be the critical key metabolites, roughly about 1,000 that are critical or essential. And then a lot of other ones, which are just around for kicks. And then all these other things, which largely represent other metabolomes. So humans and other omnivores eat other metabolomes, therefore our metabolome is rather complicated. It's also important to remember that we have gut microflora, most animals do, even plants have microflora inhabiting them that are critical to their function and existence. So there are endogenous human products and there are also endogenous microbial products. For a metabolite to be metabolite, it has to be detectable. It can be fictitious, and some of them sort of are. But ultimately we want to be able to show that it exists. It has to be, in these days, the typical limit of concentration is maybe about 1 to 10 picomolar. So I've described this already, but the metabolome is both this collection of endogenous and exogenous molecules, covers all of the things that you'll find in tissues and cells. It is defined by the technology. So if we can't see it to our end, it does not exist. So because that's this complexity where it's defined by detection technology, unlike, say, the genome where we can read everything and arguably even the proteome where we can essentially now detect everything, there are many thousands of metabolites we believe are at this stage almost undetectable because of our technology. Some of them undescribable again because of technology. And so that essentially means that metabolome, unlike the genome and unlike the proteome, is always going to be ill-defined. So lots of people ask me and may ask you how big are these metabolomes? And this slide has changed every year I've given it. It'll change next year as well just because of the fact we're dealing with an ill-defined set. So if we look at all mammals, we can say right now that a detectable number of metabolites is about 60,000 compounds. That's give or take 10,000 or so. The human metabolome database has about 40,000 chemicals, but there's at least 20,000 that might include some exotic lipids, some undescribed food components and things that we're still finding in the literature. Microbes are complicated although any given species of microbes typically has about 2,000 different metabolites that it works with. But there's lots of different species and there are lots of different ecological niches. So the estimate now is perhaps 100,000 chemicals, although I tend to think it's lower than that. Plants are perhaps the most profoundly complex organisms in terms of their metabolomes. Estimates around 200,000 to 300,000 chemicals. So again we see this gradation from the most complicated plants to the least complicated being mammals. And the reason why plants have so many more metabolites than mammals is plants don't move. So essentially for them to defend themselves they can't run away from a threat or in bacteria swim away from a threat. So plants use chemical warfare and that process has led them to essentially create the most complex metabolomes. That complexity also manifests in us in the sense that we eat plants. So we do take in a portion of their metabolome and many of those 60,000 chemicals that I'm listing for mammalian metabolites are essentially plant derived. So they're exogenous. So I mentioned that humans and other omnivores, your dog or cat or whatever, will eat other metabolomes. So this is a graph I've used many times to sort of illustrate this collection of different metabolomes. So there's maybe 30,000 endogenous metabolites. That includes microbial metabolites, things that our body produces. Many of them are lipids and lipid variants. So as I say, bottom line, we need about 1,000 to live but there's another 29,000 that are sort of there for the ride. The concentrations range from picomolar to almost molar. So the most concentrated metabolite in your body is urea, which you find in urine, and that gets up to about 200 or 300 millimolar in some cases. Next layer up is drugs. We do take drugs, not all of us, but there are about 1,400, 1,500 drugs known. So these can in a given population show up and some drugs are relatively higher concentrations, sub-millimolar generally, but they can also show up down to picomolar levels. I've indicated in parentheses or brackets some of the databases which we'll talk about later on that house this information. So in terms of endogenous metabolites, there's the human metabolome. In terms of drugs, there's a database called Drug Bank which catalogs a lot of these. For the number of years we've been working also on another one called the food database, and this represents the plant chemicals, food additives, and other things that we get exogenously. And right now there's about 32,000 chemicals that we know of in the food database. And those are also about as equally concentrated as you might get in some drug doses. So they range from tens or hundreds of micromolar down to picomolar. Drugs and foods are metabolized. We do know particularly in the case of drugs there's lots of drug metabolites and we've been cataloging those for a while. They're typically at least one-tenths to one-hundredths the concentration you'll find with drugs. And then at this top is essentially toxins, environmental chemicals. Hopefully those are at much, much lower concentrations than drugs or food or endogenous metabolites. If they aren't, then you're very sick. So we try and keep them low through environmental protection laws and through limiting exposures, but they are there. And there's at least 3,000 that can be and have been detected. And a lot of that information is a database called the T3 database or the toxin target database. So these are the ones that we can detect, ones that have been described. People have measured them or described them in some way. We know their structures. But there's a lot of other metabolomes that are or components to that metabolome that we are pretty certain are there, but we just don't have the technology to either detect them or fully characterize them. So one example is lipids and lipidome. There are a lot of exotic fatty acids that come from other plants and animals. And so even though we might say there's about 30,000 lipids that are routinely detectable, describable, or we get structures, it's probably larger than that, perhaps 100,000, 150,000. I'd mention there are maybe 2,600 drug metabolites. Those are the ones that we know about. But we also know enough about drug metabolism to know that each drug will produce anywhere between 5 and 10 metabolites. So if there's 1,400 drugs, you can do the math. It comes up to something between 10 and 15,000 drug metabolites that probably exist. The food metabolome, this represents secondary food metabolites. Food that we eat gets transformed. It'll become glucuronidated. It'll become sulfated. It'll be cleaved in different ways with hydrolysis and esterases and other things. And so these are what we'll call secondary food metabolites. And so if there's, say, 30,000 food chemicals, we can expect anywhere from 3 to 5 different metabolites of each of those food chemicals. The other thing that we tend to forget is the fact that the endogenous metabolome that our body produces, those 1,000 or 5,000 key chemicals, also are treated by our liver and by other organs as exotic. And so they'll also be transformed. They'll become glucuronidated. They'll become hydrolyzed and sulfated. So we'll call this the secondome. But these are representing secondary endogenous metabolites that have been modified, methylated, hydroxylated, mostly by accident. Sometimes they're useful as signaling molecules. But just like we talk about junk DNA, there are junk metabolites, things that just happen to be in the wrong place at the wrong time, and they get processed. And, of course, they show up in our spectrometers and instruments that detect them. So add these up, and we're talking about, you know, several hundred thousand other metabolites. So 60,000 describable, the theoretical metabolome is probably closer to 200,000, 300,000 compounds, which gets into the realm of the size of the plant metabolome. Any questions about some of this? Am I coming off clearly, or are people losing my voice in the back? Okay. So the next thing is, why? Why are we here? Why are we wanting to learn about metabolomics? Why do we think it's important? I think it's useful to have some understanding of some general statistics, at least as it might pertain to human health, which we're all concerned with. So if you go to the doctor's office, and you've never had any tests, most of the tests they're done commonly, there are rare tests or expensive ones, but the common clinically diagnostic tests are all targeted for small molecules. So there's the cholesterol test, there's the blood glucose test, there's the blood gases test, there's the creatinine test to look for kidney function, and the list goes on and on and on. Almost any of you who are younger than 25 probably had a metabolomic test done on you, and these are essentially the newborn screening tests that have become mandatory in every province in Canada, every state in the US, and most countries in Europe. So most diagnostic assays look for small molecules. The blockbuster drugs that everyone talks about, those antibodies that are being discovered, yes they are useful, but it's the same number for the last 40 years, the same proportion, 90% of all drugs are small molecules. And most of those drugs actually are derived from metabolites. They're versions of compounds found either in plants, or animals, or microbes that we've modified, and we've made them into cool drugs. So we are inspired by nature, we are inspired by metabolites. So 90% of drugs are small molecules, half of which were derived from pre-existing metabolites, some of which already are metabolites. If you do account on the number of genetic disorders that people track, particularly inborn errors of metabolism, which is probably the most frequently collected set of genetic diseases, they account for something like... 1 in 6 people has technically a genetic disorder of metabolism. And that could be anything from lactose intolerance, which is fairly common in some populations, to hemochromatosis, which is iron buildup. And when you add them all up, some of them are 1 in 10, some are 1 in 100, some are 1 in 1,000, we come up to something like a 1 in 6 frequency of genetic metabolic disorders. So that makes them the most common genetic disorder by far. But then when we tally all the different genetic disorders that we've classified, some of course might be cancer, some might be things that are related to physical disabilities, but about a third of them relate to small molecule metabolism. So again, it's a huge chunk of genetic diseases and the efforts that genetic and genomic projects are currently undertaking. And the other thing that we tend to forget, especially if you're growing up around the world of genomics and proteomics, or everything's about proteins and protein networks, everything's about genes and gene networks, most of these genes and most of these proteins would not function without small molecules. Almost every enzyme needs some metal cofactor or some heme-like group or some vitamin cofactor to function. And so these are absolutely critical for most of biology. I mentioned this before, but it's still something I think to emphasize again, that metabolites function as the canaries of the genome. So you've heard of the term canaries in a coal mine. In the 1800s, miners would go down with little canaries attached on the top of their helmets and the canaries would be whistling all the time and when the canaries stopped whistling, that was your signal to run because essentially the canary had died because of essentially a gas leak, usually methane or carbon monoxide. They were the more sensitive detector than your nose. And so in the case of metabolites, there's a more sensitive detector of what's going on. So a single-base change can lead to a 10,000-fold change in metabolite expressions. So that's one reason why the vast majority of clinical tests are small molecules tests. Things are conveniently amplified for us. So we don't have to do a whole genome sequencing to figure out what the problem is. We can just simply do a blood test and say, well, creatinine's too high or glucose is too low or whatever. But it's again because of that amplifying effect. I've also mentioned this time sensitivity with the metabolome. Again, what you eat changes your metabolome quickly and profoundly. Many, many levels. Many metabolites, literally hundreds or thousands, rise and fall over the course of minutes or perhaps most a few hours. If you eat something, yes, proteins do change. Levels of insulin, garellin, a couple of other ones will go up and down, but gradually. And then hopefully what you ate this morning isn't changing your genome. So it is essentially intended to be something that is static and does not change. So I could show this picture with people eating, but also show the same picture with a bacterial infection or a viral infection. Typically, if you are getting infected with something, your first response is not your white blood cells and not your other immune organs. It's essentially a chemical response. And this has something to do with the fact that the ancient cells didn't have immune systems, but they did have ways of controlling the flux of metabolites in and out of the cell. So this is the fast and quickest response that a cell can generally have. And this is one of the reasons why metabolomics is sort of hitting its stride in terms of biomarkers because when a disease happens, often it's the metabolite that changes or metabolite levels that change most rapidly. Another thing is that metabolism and metabolites in addition to being very useful for drug leads, very useful for monitoring events is that we actually do understand metabolism. Most of us probably wouldn't want to try and memorize this, but this is a wall chart that you can find in some of the older biochemistry labs. And it describes metabolism, basically human metabolism in exquisite detail. This chart is 40 years old. And so we've known about metabolism and all the enzymes and all the processes and pathways for a long time. We're still adding to the chart, but the fact that 40 years ago we knew this much about metabolism is pretty remarkable. It's what you learn in biochemistry textbooks. It certainly tells us that it is very well understood. It's very different than the network diagrams that you might get from cytoscape or the hairball diagrams that are typically drawn in network biology or systems biology. And I have a real pet peeve about hairball diagrams and network diagrams because they almost tell me nothing. They just tell me lots of things are connected. These have arrows. These have pathways. They have directionality. They have beginnings and ends, products and reactants. That's what we're actually aiming for in systems biology. That's what we should be trying to describe in terms of genes and protein interactions and pathways. So, whether it's that pathway diagram that I've just shown or just some of the connections I've tried to point out is that metabolism is highly connected. It impacts the genome. It impacts the proteome. It impacts the transcriptome. And that connection is sort of seen in the fact that, first of all, all of these small molecules, AMP, GMP, CMP, TMP, they're the primary components of DNA. So, without these small molecules, DNA doesn't exist. Without the small molecule 20 amino acids, proteins don't exist. Without lipids, cells don't exist. Without sugars and other, if you want nutrients, cells don't stay alive. We don't stay alive. They are the energy source. And as I've already said, enzymes, at least two-thirds of enzymes need cofactors. Those cofactors are small molecules, metals, vitamins. And these are the things that essentially help the genome and the proteome and the transcriptome work. So, you can kind of turn it around and say that, really, the genome and the proteome evolved to help chemistry. And that, really, metabolites just aren't there by accident. They are the reason why we are alive. And we are simply vessels holding these collection of chemicals. And the reason why we're not just bags of water is we essentially have a little bit more to do to make sure that those metabolites are put in the right place at the right time so that we can think and eat and breathe and move. And it's kind of the way that I think people are also turning around the idea of the microbiome, which is saying that we're also vessels just to hold microflora, which is probably not too far from the truth, either. So, there is this connection, and we are sometimes brainwashed to think that, in fact, the molecules, small molecules are just there for the ride, and everything revolves around genes and proteins. And I would really like people to think that it's actually the reverse that the genes and proteins are largely there for the ride, and that it's metabolites that are doing all the things that allow you to think, breathe, and listen. Yes. So, whether it's genomics, whether it's proteomics or metabolomics, making those connections, bringing those links together, and maybe raising the bar to what metabolism has already been able to do for biology and biochemistry, the idea of pathway diagrams, of metabolic networks, where things have a begin and an end, arrows, products, and reactants is a way of helping, I think, move the concept of systems biology forward. However, to make these things right now are largely separate and distinct disciplines. We're having a metabolomics session now. If you've been part of the CBW before, you've probably taken proteomics ones, and you've taken genomics. We've never had things grouped together, but ultimately it's going to be through bioinformatics, and it's going to be also through cheminformatics that these separate disciplines are brought together in the right way, and that we think of biology more in a systems biology perspective. So, in terms of applications, metabolomics is used pretty widely. We've heard a number of examples where people are using metabolomics in microflora work, in human biomarker studies, in plant work, in avian physiology. It's pretty diverse, but it's even more diverse than that. People are using metabolomics and toxicology quite a bit. Food and beverage testing is actually critical. It's also quality assessment, adulteration. People can identify a person's drug phenotype by giving them sick and or cocktails of drugs and determining whether they're fast and slow metabolizers. There's a field of environmental metabolomics, which just simply because ponds and other things are places where living organisms live, we can do a lot of things like water quality testing. In Alberta, where I am from, one of the very first applications of if you want metabolomics I was ever aware of that started in the 1990s, was to look at oil. Oil is an organic product. It is the product of plants and animals growing and sort of getting cooked, but there's a lot of really unique organic material there. And it's been used for quite a number of years, using NMR, primarily, more recently, FTMS, to distinguish different petrochemicals in different oils. We use metabolomics in things like genetic disease testing, obviously in a lot of clinical analyses. Cholesterol testing and even lipoprotein testing is a branch of metabolomics. It's used in transplant monitoring, and now there's imaging of metabolites that's appearing, both at NMR and also with mass spec. So lots of applications, some of which we're aware of, some of which you may not have been aware of. So I'm going to switch gears. Hopefully this has inspired you to say, let's learn and stick around for a little bit longer. But I'm going to switch gears and talk about the metabolomic methods. And this is partly an introduction. Again, I don't think everyone knows all of the methods. Some of you know the methods that I'll talk about very well, so you can kind of shut down. Other than this, for you, this will be the first time you've heard about some of these techniques. So the standard work for it from metabolomics is to start with typically a solid sample. It could be tissue, it could be plant, it could be a cell pellet. And typically the next step is to liquefy that solid sample. We'll call it rather than liquefying, we'll usually say it. We've extracted it early, homogenized. But the idea is to convert the solid to the liquid because chemistry works better in liquids. Most of the tools that we work with, like mass spectrometers in NMR, work with fluids. Sometimes we can save a step or a few steps if we can collect the fluid right away. So in the case of humans, we can collect blood or urine or saliva or supraspinal fluid. That's great. It's usually the preferred way to do metabolomics. But either way, it's trying to get to a starting point where we've got a fluid. There's lots of issues with that fluid. Most fluids are metabolically active. And this is where a lot of people forget, particularly clinicians or field people that are collecting the samples. So it's important to freeze those samples so that all metabolism stops. If you have to work with the liquid sample briefly, make sure you're working with it cold. Some fluids are essentially metabolically sterile. So urine is generally sterile, so metabolically stable. Blood is absolutely not. And this is where we see some of the worst cases of metabolic malpractice or metabolomic malpractice. So things have to be extracted or worked with fairly quickly. The next step is the chemical analysis. And in fact, you could probably go to just about any standard chemistry lab or analytical lab and do metabolomics, at least collect the initial raw data. So you use HPLCs, you use GCMS, you can use NMR. Where the real trick to metabolomics has been and is is this last error, going from the spectrum to essentially a collection of assigned or identified peaks and metabolites. And that identification process is what has revolutionized metabolomics. Everything else is standard. So without the tools, largely the software and the databases that are coming out or we'll talk about, metabolomics could not happen. So that's why we're having a course on informatics for metabolomics. So when we look at metabolomics, and this is a few years old, but it really hasn't changed much, there is a big difference in how complete our coverage is. So with high throughput next generation sequencing, it is routine to be able to sequence all the genes and genome to do RNA sequence and get all of the transcriptome. And if you aren't doing complete genome sequencing, we'll kind of look at you cock-eyed because it's so easy to do now. In the world of proteomics, a good proteomic experiment typically measures a few thousand proteins now. A coverage is at least, in humans at least five and now reaching up to 10,000. Complete proteome coverage in bacteria is fairly routine now. When we do metabolomics, although some people will claim 5,000 peaks or 10,000 features, in reality, if you read all the papers, the typical maximum number of compounds that they identify is about 200. So there's this gradient in terms of what our coverage is. So metabolomics is only typically measuring about 1% of the metabolome, where genomics is measuring 99 or 100%, and proteomics is typically measuring about 20 to 30% of the entire proteome. So we're pretty bad still in the world of metabolomics in terms of having complete coverage. And it's probably one of the reasons why metabolomics is sort of the lap dog in the omics world. But it's also something that I don't think people in the world of genomics or proteomics appreciate is that metabolomics is extremely difficult because of the diversity of chemicals. We're talking about hundreds of thousands of chemically distinct species that have to be identified. In genomics, the reason why we're so successful for sequencing entire genomes is we just had to work out the chemistry for four types of molecules. And we've also been able to, because enzymology has also figured this out, so polymerases and synthetases, are able to manipulate those things because it's just four bases pretty easily. And we exploit those enzymes to help with our sequencing. Proteomics also is easy. It's just 20 amino acids. The chemistry for that was worked out by Fred Sanger in the 1950s, largely, at least for sequencing. And then we can adopt it for proteomics via mass spec. Again, it's easy. The math, the informatics, and the chemistry are pretty easy. But when you're dealing with 200,000 different chemicals, it's tough. And so, in order to deal with that diversity of chemical structure, we have to use a lot of different techniques. We use chromatographic methods, HPLC, UPLC, capillary electrophoresis. We combine these with mass spectrometry and we use different mass spectrometers for different situations. We use NMR spectroscopy. We'll use gas chromatography for volatile things. We'll use infrared spectroscopy to help with stuff. And when we're really, really desperate, we'll even use crystallography to figure out the structures, the molecules that we've isolated. So, we could talk about every one of these. We're not going to have enough time, so I'm going to talk about some of these infrared sweeping strokes. Some of this again, if you've had a bit of a background in analytical chemistry, you can shut down, read your Facebook page or whatever. For some of you who are relatively new to this, hopefully you'll kind of tune in. So, the first thing I'm going to talk about is chromatography, because it is integral to so much of your tabloid mix. And typically, in your very first experience in chromatography in high school, usually you'll do a little bit of it in an organic chemistry lab with thin-layered chromatography. But basically, it's a matter of separating components. And you have what it's called a mobile phase, and you have a stationary phase. The stationary phase represents like silica, might represent plastic beads or drivitized beads, something that looks like sand to your eyes. And the mobile phase is usually some fluid. And it might be something that you've dissolved your mixture in, or it might be something like blood or urine, or a plant extract. So there's many types of chromatography. Probably everyone at some point has had to do thin-layered chromatography if you did a university chemistry course. If you've ever had to purify proteins, you've probably done column chromatography. If you've ever had to work with smaller molecules and spend some time in an analytical chemistry lab in second or third year university, you may have done gas chromatography. You can do variety of forms. Some of these are reverse phase, normal phase, some can be gravitational pressure, others are simply high pressure. All of these techniques are used and can be used to help with the separation process. And what's chosen depends on the chemical that you're looking at, the extract that you're also working with, how much money you have, and how much time you can afford. The most common method for separating compounds, small molecules or metabolites is high pressure or high performance liquid chromatography. So we'll just call it HPLC. That's been around for 40 years. And the essence is to use not atmospheric pressure, which is gravity feed chromatography, but to pump things up to almost a thousand times more than atmospheric pressure. About 6,000 pounds per square inch. And to work with tiny pressure stabilized beads, which to our eyes look a lot like sand. And with HPLC we're able to not only separate things, but when it's coupled with suitable detection systems, we can detect things down to the parts per trillion level, which is pretty impressive. Depending on the particles or beads that you're using in HPLC, you can separate both polar and non-polar compounds. And usually you'll try and use different columns for polar and different columns for non-polar molecules. So there are three major types of phases that are the solid or immobilized phase, or stationary phase. So the reverse phase is probably the most common. This is when we say a C4 or C18 or C12 column. That's a reverse phase column. And usually in that case the column is very hydrophobic. The mobile phase, the fluid that's pumped through the column is usually quite polar. There's water or acetic acid, acetyl nitrile, some kind of mixture. Normal phase is actually the very first form of HPLC done. It's rarely done anymore, although you can use it for lipid separations particularly. The normal phase, the stationary matrix is relatively polar. And the mobile phase is relatively non-polar, which again is useful for lipids. So it could be chloroform or methanol or something like that. One that's picking up a lot in popularity is helic or hyalic, depending on how you want to pronounce it. And that stands for hydrophobic interaction liquid chromatography. And the reason why it's so popular is it separates the polar molecules. And it turns out the vast majority of compounds that you find in urine, cerebral spinal fluid, are polar. And those are the things that are most detectable. And in fact, because plants and animals live in an aqueous environment, that's the way most metabolites are. So in this case, the stationary phase is usually a polar matrix. And the mobile phase, the solvent that you pump through, is kind of a mixed polar, non-polar phase. So again, water methanol, water ethanol, something like that. HPLC columns have to take a lot of pressure. So usually they're made out of something that's a little stronger. You can have HPLC columns out of glass, but I haven't seen one. Almost all the ones I have seen in my life are stainless steel. They're usually relatively narrow. And they're two types in HPLC. We can have preparative HPLC and we can have analytical HPLC. Almost everything that you would do in metabolomics labs would be analytical HPLC. If you're someone who is a natural products chemist, you might have some preparative HPLC systems. Preparative ones have broader columns. So the columns are measured in anywhere from 2 to 50 centimeters and they'll have diameters ranging from 1 to 50 millimeters, most being more like 5 to 10 millimeters. So the columns themselves are usually purchased, although some people actually try and pack their own. Has anyone ever tried to pack their own HPLC column? Just Michelle. It dates you then, I think. But these days everyone buys them and what you'll see here are the beads. So these are the things that are 5 microns across. Very tiny, pressure sensitive, but they are somewhat porous. So they have to be porous, but they can't compress and you're applying 6,000 PSI. So the best way to make something porous is to make it sort of out of silica. So it's almost like clay if you want or baked ceramic. But what we do is we derivatize these ceramic beads or silica beads with something hydrophobic. In this case we're attaching an 18-chain carbon moiety or a lot on the surface. So it makes a greasy glass bead. It's basically what you've done. And then you pack these greasy glass or clay beads into the column and you now have your stationary phase. Now you can play around with this. You can have something that's very hydrophobic like C18 or less hydrophobic like a 4-carbon butane attached to it. You can also change it so it's not allophatic, but you can change it so that you're sticking aromatic molecules like bifenyl molecules. Or you could even stick on these polar moieties. So nitrile groups can be stuck on. And so that becomes a polar stationary phase. So people can play around with the chemistry and the idea is the surface you're surrounding the bead with is to make it into a somewhat chemo selective if you want. Like dissolves like. So molecules that have C18 groups on them like lipids would probably like to stick onto this bead. Whereas I don't know caffeine which is a very polar molecule would not like to stick onto this C18 bead and it would come out in the void volume very quickly. So we can play around not only with the beads and what we stick on them but we can also play around with the column how long it is and also with the size of the beads. So long columns give you much better separation. So we're seeing a 50 millimeter column and a 100 millimeter column. So if you want good separation long columns are great. Problem is it's usually this issue of time it also takes longer to separate on a long column. So your boss may not give you that much time so then you're stuck with a 50 millimeter column. So if you're stuck with a 50 millimeter column you can improve things by getting smaller beads. And small beads are actually the basis to UPLC or ultra high performance liquid chromatography. And now they're getting beads that are roughly 1 micron in size. So you can still have the same column length and still have very short times 10 minute separations but get essentially the equivalent separation as you would on a 100 millimeter column using 5 microns. So that's the basis to UPLC which some of you have heard. Yeah, both, both. So if you want really really good separation, really tiny beads really really long columns it does. So typically most HPLC systems are set up there's many different vendors that produce them usually you have a solvent so this is your mobile phase there's usually different recipes and you'll have a pump that delivers the solvent at that high pressure that will inject the sample while the column is being pumped through. So the sample joins in and then it gets pushed through the HPLC column under this high pressure and then you'll detect it. And the detection systems can be anything they can be ultraviolet absorbance fluorescence which is incredibly sensitive they can be evaporative light scattering, ELSD Some people even connect them straight to a mass spec without any kind of visible detectors. So the detector can be any one of four or five different things and then as you're detecting you're tracking and producing what's called a chromatogram and collecting the material in could be waste or you could be saving the material in a fraction collector for future measurement. So this is the essence of simple HPLC most HPLC is not done with a single solvent but it's done using a gradient. So you're actually making you have two some cases even three solvent mixtures that are programmatically stirred produced over half hour or however long you're running your experiment. And this allows you to play not only with well essentially how to modify both the solution and the interaction of the solute with the column so you get preferential retention. So gradient HPLC allows you to greatly improve the separation of many many mixtures and it's a recipe that is done where there's no absolute. People will play around with different solvents different mixing regimes it's more of an art than a science but if you've got someone who's good at this they're like gold in a lab they'll work out some protocol and you can get spectacular separations before you just saw one big lump. So gradient HPLC is the other trick beyond playing around with columns and column packing and column material. And so it gives you lots of options. So the end result of HPLC, liquid chromatography with solvents like water and methanol and cedanitrol and cedic acid, gives you this. This is a chromatic round. This is one that was run about an hour, 50 minutes and you can see as measured by absorbance units or millieabsorbance units, this means we are monitoring by UV, a bunch of compounds came out. Typically each peak may correspond to one to up to a dozen different compounds. The intensity is an indication of how much material is there. Because we're using absorbance there are some things that do not have or do not absorb and so in fact they may also show up but they're not detectable using your detector. So these are things to consider and they're important when you're doing isolations and separations. So we can separate in liquid but we can also separate by gas. And one of the oldest analytical techniques around is called gas chromatography. In this case we are working not with solvents but with vapors or gases and we're separating still in columns. But the process with gas chromatography and maybe I should, how many people have done gas chromatography? 1, 2, 3, 4, 5, 6, 7... So most of you or half of you are at least familiar with this but the point about it is it's ideally suited or it was originally developed for volatile compounds. Things that are already easily evaporated, things that you can smell. But it can also be adopted to making things that are not terribly volatile like vitamins and lipids and other things, amino acids by chemically modifying them so that they can be volatile. So if you can get either something that is volatile, smelly or something that can be volatilized then you vaporize it into a gas and then you put it into a column column is a little different than an HPLC column and rather than pushing a solvent you have a carrier gas, usually something like helium that pushes the vaporized material through the column. Material interacts with the column interface some of it will stick, some of it won't stick and so things will start separating. So the mobile phase is gas, the stationary phase some kind of polymer that's absorbed to the surface of the column. Columns aren't very wide, they're usually a couple millimeters at most. Internal diameters are very thin they're very very long usually 10 even 20 meters so not like HPLC columns which are 50 or 100 millimeters. These things are 100 times longer obviously if they were stretched out as straight columns they wouldn't fit in most labs so they're coiled very tightly and they're put in an oven so as to keep things in the gas phase. If the stuff isn't smelly and easily volatilized, you derivatize it with a trimethylsilane or TMS and this is how TMS derivatization is done. You can take an organic molecule in this case an amino sugar which is not very volatile, it would be basically a solid. But if you hit it first with methanol hydrochloric acid for a while you can modify it and essentially once it's been modified in this way then it's more susceptible to being silated or trimethylsilated. So this trimethylsilo group actually makes many compounds gaseous or easily volatilized. Heating them up to maybe 75 to 100 degrees it'll just easily go into into a gas whereas if you didn't you'd have to essentially well probably try to cook sugar too long it just turns to charcoal so not the best for being volatilized. So like liquid chromatography the sample is introduced but instead of using liquid use helium gas and you push the volatile sample down through the column and so things that stick to the column move slowly things that don't stick to the column move quickly and come out and so just like the liquid chromatogram that we saw you will get a gas chromatogram where there are peaks and you can measure things using absorbance or fluorescence but usually flame ionization detection or mass spectrometry so the intensity of the peak is an indication of how much is there. To say the columns themselves are long and narrow they look like wires electrician's wires if you want coiled up inside their hollow and they usually have essentially a silica or a few silica interior that then is derivatized with a stationary phase and many of them are polysiloxane. So these are things that are a mix of silicon molecules polymerized to have either benzene or methyl or combination of those groups so they're fairly hydrophobic and this mix of aromatic and aliphatic essentially allows different compounds to stick with different affinity therefore to separate. You can play around with what you derivatize the column with and that too just like with HPLC with normal or helic or reverse phase gives you a different separation efficiencies for different types of compounds. So you'll notice that in chromatography things take time. We were talking about measurements of minutes 50 minutes for HPLC in GC same sort of thing separations typically on 30, 40, 50 minutes so we talk about retention times and so some things are retained, some things aren't. That retention time is affected by how fast you're pushing the helium gas it can depend on the pressure depend on the temperature of the oven how much you've put in all of those things What's particularly useful in gas chromatography and the reason why it's still done today is the retention times can be converted to retention indices which are surprisingly consistent from column to column lab to lab city to city country to country. It's not the same with liquid chromatography so by normalizing to a standard set of alkanes that you can standardly buy and inject you can make your retention times or retention indices consistent with everyone else around the world and so there are tables of thousands of retention indices or retention times for thousands of compounds and simply matching a retention index to some of these in these tables you can have a pretty good idea what your compound is you don't have to do any detection you don't have to do any mass spec characters and just simply saying this is a compound, this is its retention index I'm pretty certain it would have to be this compound now that's not ideal but you can't do it with liquid chromatography columns vary too much packing differs from layer column widths vary too much solvent pressures vary too much it's just too hard to be consistent from lab to lab so there is no universal table with retention times for HPLC so this is an example sort of zooming in but this is a GC chromatogram you can see one sample here we can identify based on its retention time and maybe be characterized or put an authentic sample so we know that this compound is acrylamide then we can do another run again from a different biological sample and then we see there's more acrylamide but we see that it essentially comes off almost identical or identical retention time or retention index that's this reproducibility of GC which is so useful and if we've got a sensitive detector measuring the area under the curve allows us to actually quantify so if we produced a standard that we ran first where we knew exactly how much there was in this sample B then we could use the information from sample A to actually quantify things so calibrating with standards means that GC is quite quantitative so this is again a picture of a GC chromatogram you can see that peaks are probably narrower than what you'll see with HPLC and what's the numbers essentially identified the compounds in many cases the compounds were identified by mass spec but in other cases just by looking at the retention time and retention indices they also could be identified and the nice thing about GC is you could run this that's chromatogram you get this morning you could run it tonight you could run it two months from now almost the same one highly reproducible as long as the column is still cared for any questions about chromatography ok so we're going to talk about the analytical methods that are typically attached to chromatography and so you separate a compound now you need to characterize it the most common method is mass spectromach and I think probably everyone here that's doing metabolomics probably uses mass spec so it's a method for weighing samples there's all kinds of different mass spectrometers this is an older one a time of flight instrument but they're usually the size of a refrigerator either laid on its side or standing up and they usually have computer equipment attached to them really what you're doing in mass spectrometry is you're identifying compounds by their molecular weight every compound has a fairly unique molecular weight just like everyone in this room probably weighs slightly differently so if I can't remember your names if I had a table of your weights and your names I could suggest that ask what's your weight now I know your name we're doing the same thing in mass spectrometry we have and we can calculate the weight of any molecule if we know its formula obviously there are isobarque systems or compounds that weigh identically so leucine, leucine or isomers and they weigh identically mass spec can't really tell us what they are unless we do something special but that's the that's the essential point about mass spectrometry measure something, weigh it you can probably figure out what it is with the best mass spectrometers we can measure molecular weights or atomic weights down to one part per million and that's sufficiently accurate now to determine not only its weight but to calculate its molecular formula and that's pretty cool just being able to say this is its molecular weight this is how many carbons it must have this is how many hydrogens it must have this is how many sulfurs, nitrogens and so on for peptides and proteins with that mass accuracy it means we can generally determine the molecular weight of a protein in one Dalton which is pretty good so what we usually do is we attach mass spec to other instruments so we'll attach gas chromatographic equipment to a mass spec we'll attach HPLCs and UPLCs to a mass spec so that becomes LCMS or we can attach a mass spec to a mass spec and that becomes tandem mass spec and mass spectrometry also separates molecules so we have one that sort of separates and then one that sort of then analyzes or weighs the molecules so these are different types of forms of mass spectrometry that we'll use and are commonly used in metabolomics a good metabolomics study probably would use all three of these so with the instruments of today the modern ones, we measure what's typically called the monoisotopic mass so that's the mass of the compound the molecular weight using the most abundant isotopes if you have a low resolution mass spectrometer like a triple quad or a Q trap you'll tend to measure the average mass and that's essentially the mass of all of the isotopic components so what you're seeing here is some heavy molecule with a maybe it's a lipid or something but 1155.6 Dalton that's the monoisotopic mass, that's the tallest peak roughly in steps of one Dalton to the right are the different isotopomers and these represent things that may have deuterium or carbon-13 or something like that or combinations of them they're less abundant because deuterium and carbon-13 are rarer what's marked in red is essentially the weighted sum of those isotopomers and that gives you an average mass of 1156.3 so if you had a low mass or a low resolution mass spectrometer you'd get an average weight of 1156 this just illustrates chlorobenzene where we have a list of the isotopic distributions hydrogen, deuterium carbon-12, carbon-13 in this case chlorine chlorine-35 and chlorine-37 carbon-13 and deuterium are fairly rare but chlorine-37 actually is very common about a third of all chlorine atoms are there so based on the abundance of these hydrogen, carbon and chlorine we can actually figure out what our isotopic abundance would be so if we had all of the lowest molecular weight just proton just carbon-12 just chlorine-35 the mass of this molecule is 112.007 we could also get two other versions one Dalton Moore where either there's a single deuterium or a single carbon-13 in the molecule then you could imagine that we could also have two carbon-13's and two deuteriums and so in that case the molecular weight is two Dalton's more or we could have a chlorine-37 which also boosts the molecular weight by two Dalton's but because chlorine-37 is so much more abundant we'll actually see a much more intense peak so this is actually the isotopamer pattern that you actually see for chlorobenzene so instead of this rapid tailing off that we saw here which is more common you see this mass peculiar shift where you have what's marked as the intensities but if you have a high resolution mass spectrometer you normally see multiple peaks for a single compound and that's because you're seeing these isotopamers being detected in these very high resolution instruments all mass specs are constructed in this three step structure they all have some kind of ionizer they all have some kind of mass analyzer or mass separating device and they all have a detector they all produce a very sharp or finely described peaks where what's shown at the bottom is the mass to charge or M over Z ratio and then the intensity which is measured by the frequency of these hits so here's this is an EIMS spectrum so you're seeing the fragments of a compound no so the primary ion would be typically maybe it's the one around 180 in this case and then yes you're seeing fragments from electron ionization so the fragmentation of aspirin but again just the point it's a little different than a chromatogram in that we're measuring not time but mass to charge we do see intensities we see multiple peaks so it looks like a chromatogram we can measure mass spectra I mean most of them have these narrow peaks so said there's this height some ions fly some don't so this is one of the challenges of mass spectrometry not everything is detectable not everything flies from the matrix as a result the intensity of the peaks don't really indicate the quantity whereas the HPLC or GCM that's usually the intensity of the peak is a good measure of quantity we can measure the quality of the mass spec or the quality of the spectra by resolution or resolving power just like when we look in optics can you resolve something in focus or out of focus so we use the same optics term called resolving power or resolution so delta M is the width of the mass the two mass differences we can separate or distinguish and then mass is the observed mass of the molecule that we're looking at so most people use this 50% cutoff 50% height and it's the same thing that's typically used in optics but if you have two peaks that are close together two dots on a page how close can you move the dots before they seem to merge to one same sort of thing with mass spec how close can you move the two mass peaks before they seem to be just one big lump so that's a way of measuring resolution and so you can see here in a real example this is a lower resolution mass spectrometer, an ion trap where the resolution here is defined as 700 and then you have a higher resolution time of flight mass spectrometer and what you're able to see are all of the isotopamer peaks so the resolution in the lower one is almost 10 times greater yes we're assuming that we have our separation of the complex mixtures you can't generally in fact this is one of the challenges the you will often have many situations where there's multiple masses sort of overlapping and so that's the challenge about separation and then the spectral deconvolution so we'll be talking a little bit about that tomorrow that's the question now in terms of quantification particularly in terms of absolute quantification it really doesn't matter too much usually you're going to be using an isotopic standard to help with the quantification so you have an authentic standard and you're looking at relative peak areas and by the time you're quantifying you know exactly which peak you're looking at, what time it's coming off on the chromatogram so most of that has been sort of worked out so this is another example just illustrating a range resolution going from in this case a thousand to 30,000 which you might get in and say an FTMS instrument so we're going to blue to red to green to black and you can see that the blue level looks like one big lump and by the time you're looking at the black peaks every single peak is clear and sharp easily differentiated so a real difference when you're working with high resolution mass specs and there's a lot of information there that we'll learn about a little later so different kinds of ways of ionizing so remember there's the ionizer the mass analyzer and the detector so there's different ionization techniques that are used one is called the electron ionization this is characteristic of gas chromatography it's generally done for small molecules it's actually a great way of determining the structures of compounds so if you're working with unknowns it is possible to actually figure out the structure or at least partly the structure there's also chemical ionization methods CI that's also used in gas chromatography frequently it doesn't fragment things quite as badly as electron impact then there's ESI and MALDI these are two soft ionization techniques originally developed for proteomics and proteins but they can be and are frequently used in metabolomics so they don't fragment the molecules quite as badly often the parent ion is still detectable and this is a technique usually where we're just seeing some cases just a single peak just the parent ion thus identify particular metabolites this is just a picture of electron ionization and we can see that this is very similar to GCMS instruments so you have a standard electrode 70 volts that's being applied you put in your normally neutral molecules but electrons then fly off from these electrodes and impact the neutral molecules and ionize them once they're ionized then they can be characterized by mass spec because we're always measuring charged molecules in mass spectrometry so EI ionizes by hitting things with an electron beam typically what's submitted at this time everything is volatilized mass spectroscopy or spectrometry is everything has to be volatilized, nebulized vaporized in some way so that things can be further ionized usually it's a pretty low vacuum high time not really low for GCMS for some of the LCMS which we don't usually use for EI it's somewhat much lower microtor 10-6 10-7 so EI is used for gas chromatography MS we bombard things it could be just the equivalent of a tungsten filament not unlike what you see with a light bulb an incandescent light bulb but it's at a set energy it's always 70 volts or 70 electron volts electrons moving 70 electron volts will shatter bonds which are held together at 5 electron volts so you break up the molecule and so the fragmentation is sometimes pretty severe in the case of something like methanol you can fragment it into 2 or 3 or 4 different molecules or fragments you're knocking off the hydroxyl group you're knocking off the hydrogen you're knocking off changing things to double bonds very bizarre looking molecules they're very metastable they're stable for very short periods of time but they are all ionized they all have in this case a positive charge and when they are sent through the mass spec they fly through the electrodes and they will be detected by the detector and you will see masses so the CH3OH parent ion 32 Dalton's the CH3 methylene group 15 Dalton's you'll see that so you're seeing these different peaks the intensity partly reflects the level of ionization relative abundance partly how well they fly it's not as I say the best way of measuring quantity but at least in EI molecules break up in predictable ways and because it's so standardized there are tables and databases of thousands of compounds with standard spectra so you can compare an EI spectrum and identify a compound by just looking at that sort of fingerprint soft ionization methods are different Maldi you put a substance on a sort of a photo absorbing matrix you'll hit it with a laser that absorbs heat things sort of blow off and you ionize what was deposited on the matrix or in the case of electrospray ionization you'll put things through sort of like an aerosol spray can or equivalent there's a high voltage at the end of the aerosol tip and this causes things to further vaporize in a low vacuum environment low pressure vacuum environment and the vapor turns very much into a gas this is sort of shown in detail here where you'll have sort of a fluid that's pushing past the through capillary and then you'll have a gas that surrounds the fluid and a very strong electrode thousands of electron volts that essentially just causes this to work like an aerosol spray can so you get fine particles of fine mist spraying out and then as the mist goes into the vacuum the particles which are the stage tiny liquid drops start evaporating they contain charges because you've applied a fairly high voltage at the tip of the electrospray device and as they evaporate they essentially shrink down to just carrying a couple or one ion and of course then they'll fly through the electrodes and the mass spec and then it ultimately be detected different compounds will have different charges and some of them will be multiply charged so it depends on luck and happenstance of which ones are going to be singly charged or doubly charged or triply charged ESI is something that's typically used with substances that have no salts so you can't put urine into an ESI straight away because there's lots of salts you don't want to have detergents in them but if things fairly clean processed usually through or cleaned up through an HPLC then you can pump them through these capillary tubes at a few microliters a minute you apply this really strong voltage and then you create this aerosol and the aerosol if you've ever put your hand against a spray can with off or something like that some kind of insecticide yes it's liquid but if you put that aerosol spray can in a vacuum then the droplets will rapidly evaporate and then you essentially get these essentially ions gas ions that are essentially atomic size of course they still carry charges you can play around not only with the voltage the size of the capillary the sheath gas, the flow but you can also play around with the solvent so depending on how much water or acetonitra or other volatile solvents you add you can change ionization begins so if you have more of a hydrophobic solvent ionization will happen at a lower voltage you can make it now nano spray devices where they're looking at less than a microlitre per minute it's extremely sensitive as I've mentioned this issue of salts and detergents so you have to do quite a bit of cleanup for ESI to work and then you have to switch between two different modes often positive mode and negative mode some molecules ionize better under the positive mode some molecules ionize better in the negative mode depends somewhat on the structure of the compound so these are the ionization methods then you can do analysis where you'll have the terminology and mass spec so almost none of you probably would have heard of the magnetic sector analyzers but these are the original mass specs from the 1900s still occasionally you'll find a few of them but they use big magnetic fields to separate things now everything is done through electric fields typically although in the case of ion FTMS they also use magnets, huge magnets so the quadrupole analyzers these have lower resolution one atomic mass unit then there's the time of flight we have higher resolution of 5000, 10,000 and then the FTMS instruments which are the most expensive and generally have the highest resolution available so we talk about resolution we can also talk about mass accuracy and they kind of go hand in hand and this is just this table of the accuracy so you can see that FTMS and orbitraps give you your best mass accuracy so we can talk about mass accuracy or precision around 1 ppm no one has magnetic sector ones but they're pretty accurate then there's the time of flight mass specs so somewhere between 3 and 5 ppm is typical some triple quads can get down to 5 ppm but you have to operate them in a different mode and then mostly ion traps around 100 ppm so lower resolution so when we collect a mass spectrum we actually can collect the different ways and play them in different ways we're measuring we have a detector we're seeing things coming off in this case mostly electrospray so there's a time chromatogram element to this if we have an LC and it's pushing off and we don't put a UV detector but we put a mass detector then we're essentially detecting mass is not UV absorbance and so we are measuring over time and so we can see what are called total ion chromatograms that's the red one which is just some of all intensities across the entire range one that is generally preferred is sort of the base peak chromatogram so it's mostly displaying the most intense peaks and then there's the extracted ion chromatogram where you might just have one or two ions that are extracted from the full set so this is what we would see in terms of a chromatogram so we've seen an LC chromatogram we've seen a GC chromatogram here is an MS chromatogram where we're seeing time and in this case we're seeing masses for specific compounds so the masses correspond to a compound so we're just using mass spec instead of UV or fluorescence as our detector and we're running over time because we're using in this case LC so I've only got about five minutes here and people are okay because we kind of were a bit delayed with the break if I could run on for another five minutes beyond so about 1035 people aren't going to die so I'm going to talk about sort of the last part here which is just NMR so this is another technique and I don't know how many people have ever used NMR in their metabolomics one, two, three which is kind of the normal breakdown so it's not the most popular method for metabolomics but we're going to try and convince you that's a really useful method so NMR also produces chromatograms uses big magnets and we detect spectroscopic features that use strong magnets we'll work with liquid samples under an intense magnetic field of tens of Tesla or hundreds of thousands of Gauss and what we do is we send radio waves into the sample under this strong magnetic field and then we look to see where those radio waves or radio frequencies are absorbed so it's a little bit like UV absorption where we measure what's absorbed and just like the different colors are absorbed so in radio frequency there's different colors of radio frequency so some compounds absorb at high frequency some absorb at low the unique feature is that it's not a single peak for a single compound it's usually many peaks for a single compound and that's the special strength of NMR it allows you to uniquely identify many compounds that way so NMR measures nuclear magnetism it's not electromagnetism it measures the magnetism in the nucleus it's not working with ultraviolet light it's not working with infrared light it's working with radio waves and it only works when something is under a strong magnet so has anyone ever had an MRI scan done on them? a few of you, okay so you're putting a big magnet otherwise you're invisible so they magnetize you and all of your nuclei so they can actually see what's there different nuclei will absorb for different frequencies so carbon absorbs differently than hydrogen which absorbs differently than fluorine and so on NMR takes advantage of the fact that nuclei protons which are in the center of every nucleus spin around like tops and because they have a charge when you have a spinning charge that creates a magnetic field so some of them will be spinning up clockwise others will be spinning counterclockwise so by convention we have a spinning up or a spin down so that spinning creates one where the north pole is up or other cases the north pole is down so again it's a random collection of nuclei or protons from random substance much with up spins and down spins when they're under a strong magnetic field you have a Boltzmann distribution of up and down spins but if you send in a radio wave it'll cause some of these nuclei to change their orientation in the spin they'll flip up so hit something with energy just like when you warm something up adds energy so these spins flip up and that incident radiation is the way that we measure the absorption and then when you turn it off they relax and some of the spins flip back as that flipping back happens we're able to detect radio frequencies that have been absorbed so you can increase the magnetic strength and in fact the stronger magnets actually give you better NMR spectra they go from lower frequency to higher frequencies and we really like high frequencies really really big magnets help a lot so in a modern NMR instrument you'll take in a sample you'll inject it into a magnet you'll bombard the sample with radio waves and then eventually detect these signals the magnets themselves are about the size of a refrigerator they are superconducting magnets they're giant thermoses wrapped around with a liquid helium bath to keep things super cool with another bath of liquid nitrogen surrounded by space blankets and metal containers so a typical NMR magnet is about half a million dollars and they're maintained regularly by filling up with liquid helium and liquid nitrogen and at the core of this giant container is a magnet maybe about the size of a toaster wrapped with a tin niobium wire and this is sort of the cross section where you can see the collections of vacuum liquid helium, liquid nitrogen just to keep things ultra cool but then this superconducting magnet at the center of it in the superconducting magnet it's about four kelvin so very cold but the probe which is where your sample sits is at room temperature but the sample inside just like if you had an MRI you're not chilled to four degrees kelvin but it's room temperature so this is where the electronics it does although all the ones that they're using now are just this tin they haven't used the new ceramic superconductors so they're just using tin and niobium so it's unfortunately they haven't been able to figure out the necessary flux yeah but inside the magnet you have a little probe and this is where there's all the radio receiver transmitter stuff that probe is at room temperature and in that probe is a little wire called a saddle coil which is where your sample actually sits so the probe goes up at the bottom of the magnet and then there's this bore which is about 10 millimeters across maybe sometimes up to 50 millimeters across depending on the sample so this is that's an NMR tube so very thin pencil size tube that's dropped into this coil the coil is an antenna it's a radio antenna and it allows you to transmit radio signals and to detect radio signals so it's like the optics in a UV spec it's not light, it's radio waves so that's kind of the meat or the core if you want so you've just got this massive thing just to try and give you a magnetic field but the essence of of the NMR is this coil here so what you'll get is a looks like chromatogram again a bunch of peaks in this case instead of measuring time we're measuring chemical shifts and these are frequencies super million, we'll see certain splitting patterns we'll see things with different intensities the chemical shifts are the things that tell you what you're looking at different hydrogen atoms have different chemical shifts depending on the chemistry, the bonding and every molecule or just about every molecule has a unique set of fingerprints or chemical shifts so NMR even today is the standard method for determining the structure of small molecules it always beats mass spec, always will beat mass spec because it provides you this detailed information about neighboring atoms, chemical bonds chemical structure most NMR spectroscopists memorize these kinds of tables so they know which chemical shifts correspond to which groups so they can often just look at a single NMR spectrum and figure out whether this is a methyl group methylene group whether it's located to a strongly unectronegative atom telling you how many protons there are again it's a practice but it is something that can be computerized where people can look at that spectra and assign them and their characteristic chemical shifts around 7 or 8 ppm where you're looking at aromatic groups methylene and methyl groups the numbers of peaks telling you how many protons are there the intensity and so on they don't give you clean results so when you first collect the NMR spectrum they're really messy the peaks are kind of all over the place they're warped this is called their out of phase and so humans actually have to do a lot of that fixing so they don't give you a nice clean chromatograms of HPLC or GC they kind of mess up with this so we'll learn about that later on for your lab and so this just sort of describes the fixing that we typically have to do to clean up a spectrum so it's presentable and usable and it's a manual process always has been hopefully in the next few years will become fully automatic so again if you look at each of the things I've seen I've shown you guys liquid chromatography gas chromatography MS, NMR they all have this characteristic spread out in some cases over time in this case over frequency and in all cases intensities vary and we're usually using that to identify what these compounds are the reason why I've talked about these different methods is that each of them is useful for different types of situations some are useful at measuring higher abundance molecules more polar molecules NMR is very good at that others are best looking at low abundance molecules someone might say based on the sensitivity of these systems why don't we just always do LCMS that's partly the reason why a lot of metabolomics is done with LCMS the problem is that 90% of what you're seeing with LCMS we don't know what it is whereas with NMR 90% of what you're seeing you do know what it is and so when you're working with knowns it's obviously easier to write a paper or to come up with a pathway or describe a phenomena and so that's one reason why NMR still continues to be popular even though it's not very sensitive the other thing is that each of the techniques specializes in different compounds so to get a full picture you often have to use all three or four different techniques this is just a comparison I'm not going to take the time to read through it so you guys can look at it during the coffee break but this just highlights the different volume requirements the types of metabolites the numbers of samples that could be done the process, the limits of detection the numbers of metabolites that people might see in a given sample and then the overlap and the important point here is that overlapping metabolites these are complementary methods they're not duplicating methods so it's worthwhile doing at least two and usually three different methods NMR can give you up to 200 compounds that you can identify and quantify GCMS can also match that in some cases you can use flow injection mass spec or liquid chromatography mass spec you can generally identify more with DIMS and LCMS although it's hard to quantify them lipid methods now because there's so many different lipids and they all are very similar in structure you can actually get several thousand identified the last part is that we're talking about identification and then something called chemometric methods there's targeted and untargeted metabolomics we're going to largely focus on targeted metabolomics here and that's where the whole field is moving it's trying to identify and quantify everything that they can see or as much as possible the older methods and some metabolomics were saying we're looking for pattern differences is this pattern different than that pattern and if you see pattern differences that's good and then you have to go back and do quantitative metabolomics which can take just as long as you started with quantitative metabolomics to begin with so that's why the chemometric or non-target methods are sort of being replaced this is a flow for untargeted metabolomics you collect lots and lots of samples look at things as other patterns or clouds see which clouds are more similar and then you go off and separate them using different methods which we'll talk about tomorrow and then eventually identify the quantitative approach is to identify first quantify roughly at the same time then use the data analysis so it actually saves you a step and then you can go right into the biological interpretation so what we're trying to do and what we'll do over today is to go from spectra to lists of compounds both the identity and hopefully their quantity and then what we'll do tomorrow is go from those lists to their pathways which is the biological interpretation and then at the very end we'll show how you can go from those pathways and lists to things like models and biomarkers things that are important these days so all of these are things that we're going to focus on for the next day or day and a half and we'll try and talk about each of these specific topics in each of the next sets of lectures okay it's coffee time