 Well, we might want to do just before diving into things We had sent some instructions for all of you about downloading software In a couple of hours. We're going to be starting a project. We were doing Some spectral deconvolution with what's called the kinomics software Anyone who was not able to Download that software just raise your hand Okay, so there's three of you who don't have four of you five of you five of you who don't have the activation license Probably hours I hope you already by today you send you the link to the license Okay, so you said last night how about you guys you just sent last night Okay, this is a weekend. I think most companies work Monday to Friday Well, they they're Edmonton Bay, so they only would have been on off on July 1st, but So hopefully they'll be sending you a key, which would be about 8 30 or 9 o'clock Mountain time which would be 10 30 or 11 Our time which would correspond to about when we start the the exercise If you're still waiting you might want to pair up with someone In fact, there's a couple of people here Who actually are experts in this and they infect me? May end up Helping other people and maybe donating their computers to people who don't have their access keys Okay, so that's sorted out I think the other one was downloading XCMS Anyone who was not successful in downloading XCMS Raise your hands Over the coffee break Okay, that's good Anyways, we're this is a new new room new facilities and obviously every year as well there's changes or adjustments to both the software that we commonly use and Operating systems, so hopefully we'll we'll squash those little troubling bugs Okay, so these are the two standard slides that Michelle has already talked to you a little bit about And we're going to be starting into metabolomics here All of the lectures I'll be giving will have some a slide like this, which is sort of the learning objectives Instead of an outline of what we hope you'll be able to learn over the next hour and a half And Again, this being an introduction that we're really just trying to get people familiar with metabolomics, so we're all on the same Plane here also to Look at some of the applications and obviously many of you are applying metabolomics, which is good And and then we'll look into some of the technologies and this is what the focus is and Metabolomics is really Analytical chemistry on steroids. We're trying to do lots of analytical analyses often with many many different tools So we're going to go through some of them and we've got people with very varying backgrounds So we'll try and get you up to the same speed and Then also close off with this difference between targeted and untargeted metabolomics and we'll talk about that quite a bit through the next two days You've seen the schedule. I won't go through that again, but This does give you day two So This is what we'll be doing and typically we'll wrap up by five o'clock And then on the last day, there'll be a bit of a survey Where are we trying to get your feedback? So if there are things that we'd like to see or what hope you could have seen in the next next year's Metabolomics Presentation or workshop Put that down take notes. It'll help students who come later on And it may also allow us to develop Different kinds of workshops as well So here's the first slide And this is a way I often begin with describing metabolomics What I call the pyramid of life The base of the pyramid is the genome and the genome codes for proteins And the collection of all those proteins is called the proteome Proteins are there really to function as catalysts in many cases or as transporters to move Small molecules in and out or to produce small molecules or manipulate and transform them So the proteome actually that largely helps code for the metabolome The Reason why I'm showing this as a pyramid is is in part because one depends on the other which then depends on the other But there's also feedback metabolites Effect gene function a lot of feedback control some of the very first Operons in genetics were identified as metabolite based operons There's also progression Small changes in the genome can have very profound effects at the top of the pyramid At the top of the metabolome and we'll learn about that. There's also progression in terms of environmental effects and physiological effects And this is something that's particularly important. So the genome is largely unaffected by the environment What you just ate this morning for breakfast is not affecting your genome Maybe if you ate the same stuff every day for 30 or 40 years it might subtly affect your genome What you ate this morning or what you're breathing is mildly affecting your proteome But it's only a few proteins insulin, glucagon, garelin But what you just ate or what you're breathing or drinking actually is profoundly influencing a metabolome even as we speak It changed almost the instant that you took in any Anything and it will change your metabolome for up to several hours afterwards So that's one thing that's particularly important is that the metabolome is significantly influenced by your environment And in that regard the metabolome is an excellent indicator of the phenotype. It's a quantitative measure of the phenotype phenotype is the sum of the genome and Environmental interactions Another aspect that we tend to forget is that Physiology is reflected in metabolomes arguably genes coding for our organs are important for the development of organs different organs have different proteins and proteomes But we have specific organs like the stomach, the liver, the kidney Which are developed specifically for metabolism meaning that metabolism is compartmentalized in many different parts of our bodies Same as true with plants. Same as true actually even in some complex single cellular organisms and We tend to view unfortunately in the world of metabolism and genomes and proteomes as everything is just a single cell and it isn't and in fact What we measure Metabolically in plants or animals by looking at biofluids represents often the sum of many different physiological influences And we also also have to understand that measuring metabolomes in different fluids blood, urine, saliva, cerebral spinal fluid By all of those reflect what's going on in these different organs Therefore physiology can be captured particularly well through metabolomics and is significantly influenced by metabolism The term metabolomics we all use it sometimes people call it metabolomics where the L is replaced with an N Although largely most people are using the metabolomics term Definition I think everyone knows But I just to reiterate it's largely in the case with the case of genomics Which has been around for a lot longer. It's high throughput technologies Looking at all the genes in a given cell tissue organism and so for metabolomics Same definition high throughput technologies, but it's not the genes we're looking at its small molecules metabolites But we are looking at cells tissues or organisms, and I think everyone here who's doing metabolomics is looking at probably those different Aspects metabolism a Definition of a metabolite. This is one that's challenging because there's some people who have very different views and I still have People collaborators coming up to me and saying I hear you do metabolomics. Can you measure these proteins for me? No Metabolomics is not measuring proteins. Typically. It's measuring small molecules Tabloid or molecular weight we usually use the cutoff of about 1,500 Dalton's So we can measure peptides, maybe 12 or 13 residues that counts Large lipids some actually get up to 2,000 Dalton so that 1,500 Dalton cutoff isn't perfect But I think most of us understand what we mean by small molecules Unfortunately the rest of the world does it so we'll look at this as I say we'll be talking about And if you look through databases that talk about metabolites, you'll see things like some short peptides illegal nucleotides and sugars Nucleosides and organic acids and the list goes on it includes endogenous molecules But also includes a lot of things that we find outside So that includes foods and food products food additives toxins pollutants drugs and drug metabolites challenge with metabolomics for at least omnivores like humans or mice and rats Or plants growing in the modern world is that they we consume other metabolomes late several fruit metabolomes this morning and If we eat meat who are eating a tabloom of beef or lamb or fish There's other components to the metabolome we produce our own metabolites we consume those metabolites But we also have gut micro flora that live in our large intestine and actually many different compartments of the body all through the gut And they produce products. So it also includes micro floral microbiome products typically with modern technology Metapolites include the things that we can detect and limited detection is something on the order of pico molar But most people and most instruments it's more on the level of nano molar So that's a metabolite metabolome Formal definition. That's a complete collection of small molecules. So less than $1500 and that can be in a cell an organ Tissue or an entire organism And it helps to define that so you might say I'm looking at the urine metabolome That's a component of say an organism's metabolome But to say that you're looking at the human metabolome largely means you have to look at all of the compartments all of the organs all of the tissues As I said when we talk about metabolomes We are often looking at both the endogenous molecules things that are produced by the organism Or are needed by the organism as well as the exogenous which includes things that the micro flora produce includes things that we eat There are many transient molecules in metabolomics. We know about some of them. We still have yet to measure others These are molecules that may exist for fractions of a second. There are others that persist for decades There are other molecules that we have yet to actually measure but based on the chemistry Then we have inferred their structure. You'll find those sometimes drawn in in pathways Metablo itself is also something that's defined by current Technology so the genome Has been defined the human genome is fully measurable. We know pretty much the exact number of bases It's not going to change or grow except maybe over a course of millions of years by small increments The metabolome is something that will vary as our technology improves the size of the metabolome will increase Because we can detect more So we still really don't know the true size of the human metabolome or the Arabidopsis metabolome or the mouse or rat metabolome We don't know the complete collection of secondary metabolites and plants So it's still ill-defined. Whereas we have a pretty good handle on the exact number of genes bases Even proteins in many organisms So in terms of plants based on detection and reports in the literature We think there are about 200,000 different chemicals Clues primary and secondary metabolites in the plant kingdom Could be much larger in fact. I'll argue it. It is and we'll show you a little later Microbes are not as complex as plants The number of 60,000 is sort of just pulled from the air. We're still not sure but microbial Complexity is quite large. We draw many antibiotics and exotic compounds from microbes Mammals have about 20,000 detectable endogenous metabolites. So whether it's humans or mice That's pretty much the same metabolome same Kinds of compounds the only thing that differs is probably their concentrations So mammals actually although they might be regarded as the most complex organisms probably have the simplest metabolism What is plants? Which we might regard as fairly simple Have perhaps the most complex metabolism and the reason for the complexity of Metabolome in plants is because They can't move they can't run away from predators Or from parasites So what plants have evolved is a capacity to wage chemical warfare And this is how they battle both pests and parasites. It's how they avoid being eaten By producing toxic compounds. It's how they protect themselves And so this lack of ability is this actually spawn this remarkable complexity in the plant metabolism in terms of the human metabolism and we could put the word mouse or rat or horse or cow Relatively similar or very similar We know This graph really illustrates both the range and concentrations We're remarking from picomolar to molar And the number of compounds that have been recorded or identified Either quantified in some level or confirmed through Experimental techniques, so there's about 19,700 in dodge this metabolites in the human tabloom That's in a database called HMDB Humans take drugs There's about 1400 that are known The range and concentrations that you'll find in various tissues or fluids typically is in the Sub-millimolar range down to picomolar Whereas the endogenous metabolites cover this full range of about 12 orders of magnitude People eat foods and in fact the concentration of food metabolites and food derivatives Is roughly comparable to what we find for drugs There's a lot more variety of food components is nearly 30,000 based on what we know and it's been measured Drugs are metabolized into drug metabolites. They're at much lower concentrations In order of magnitude at least Several orders of magnitude But drugs break down to drug metabolites in some cases the drug metabolites are actually the active component In other cases the drug metabolites are the toxic components So the very lowest and hopefully well below micromolar usually in nano molar You'll find various toxins pollutants environmental chemicals that are produced by modern society Those are also tracked. There's about 3200 that are common these include pesticides and herbicides and Poly aromatic compounds that are used in transformers which are in all of us whether you like it or not And as I say, hopefully these are very very low concentrations So the collection of Metabolite metabolites that might be in the human body at this stage would come to perhaps 40,000 And that wouldn't be in a single person. It might be in a population of several thousand individuals Because not everyone takes all 1400 drugs and not everyone eats every variety of food and hopefully not everyone's been exposed to every toxin So that's a fairly large variety But when we think that there's something on the order of 35 million different compounds that we know This is actually a small fraction and it's an important thing to remember Those of you who've heard of pub chem Often are drawn to it because of the sheer size But 99.99% of the compounds in pub chem have never left the lab They're not part of the metabolome of any known organism nor will they ever be And so this is an important thing when you're trying to search If you've ever done proteomics or genomics, are you gonna hunt for a gene by looking for that gene in every organism? No, you typically look for the gene for the organism of interest same thing with a protein experiment So in the case of them tabulum, it's important to know both what organism we're working with but also know that you're working with Natural compounds not synthetic ones Now though what I gave you in that graph is the essentially the known Metabolome these are things that have been detected or measured There's also what we'll call the theoretical metabolome and If we had infinite capacity to measure the lowest concentrations We know that there's probably on the order of a hundred thousand different lipids in the human body But as I said many of them are not detectable or have not been detected We also know based on the number of drugs that are used Roughly seven to eight metabolites for each drug that there should be roughly 10,000 different drug metabolites only about a thousand have actually been characterized We also know based on the foods we have Many foods are broken down through phase one and phase two metabolism So if there's about 25,000 different foods assuming there might be four or 25,000 different food components There might be four or five metabolites produced for those So there's perhaps a hundred thousand different food Breakdown products if you want or secondary metabolites And then our own metabolites our own endogenous Metabolites many of them are also modified our liver Things that they might be for and we'll transform them and that could be anywhere from 10 to 50,000 compounds We really don't know so the theoretical metabolome is not 20 or 40,000 but probably on the order of 300 or 400,000 molecules So that's a scale and those are estimates and every year we're modifying those estimates as we learn more or as technology improves Does anyone have any questions? Again, it's quite introductory The data what I gave there was at least for the human was was data that we had either measured ourselves or found in the literature over the course of about seven or eight years of compiling that data So those are the concentrations that you will find So I didn't give you any specifics about the concentrations for plants So it varies tremendously with with plants You know my own recollection of what we've seen with Plants is that we'll see things in the millimolar high millimolar concentrations for certain tissues and things like glutamine and glucose And obviously the secondary metabolites some of them are well below nanomolar levels So the range is there It may vary with other organisms So in terms of Metabolomics and its importance, I probably don't have to convince most of you but I think it's a challenge for the rest of the world and it's something that I I guess been Pushing for for probably more than a decade Most people believe that the only important molecules are our genes and proteins and Certainly someone who went through biochemistry And who is still a structural biologist That's kind of the brainwashing that I went through as well as a student and as a researcher And I think it's a bit of struggle to try and get people Interested in metabolism partly because it's not well taught Also because it's an area where people feel like it just doesn't count or that the molecules aren't that important But maybe these statistics might Might be useful. These are statistics I hope you would try and convince your colleagues who look down upon the work that you might be doing as being irrelevant So basically it was a common clinical assays that are used today Besides the word common, but if you go into a doctor's office 95% of the assays that your doctor will run are testing for small molecules 90% of the known drugs are small molecules Even though the vast majority of R&D work in drug companies is focusing on proteins Still the vast majority of drugs that are both introduced and are known are small molecules Half of all the drugs that we know of actually are derived from Metabolites in fact hundreds of drugs actually are metabolites But they may be found from plants or exotic bacteria or microbes But the inspiration for those drugs is drawn largely from metabolites and the structure of metabolites a Third of the known genetic disorders genetic diseases are Directly of all small molecule metabolism And then of course most proteins and enzymes would not function without small molecules and we tend to forget that we typically draw Hema-globin without the heme or Subroxide dismutase without the zinc or the copper, but each of those proteins would not function without those cofactors Likewise, we tend to forget and it's usually not drawn in your keg pathways But small molecule metabolites are integral to signaling far more than we realize And unfortunately people still today when they get their list of metabolites immediately go to the keg pathways and try to explain what they're seeing based purely on metabolism When nine times out of ten, it's about signaling Metabolites when I do that pyramid are really the canaries of the genome And let me say canaries as the analogy is the canary in the coal mine coal miners used to have canaries and cages that they would put Either hanging up in the coal mine or wear them Under helmets and if a canary stopped singing, you knew that you're in danger of either a methane explosion or or gas explosion because Canaries typically die very quickly when they're exposed to high concentrations of CO2 or methane So that's the term canary in a coal mine It was an early warning system, but it was also a way of amplifying signals that you couldn't easily detect so in that regard metabolites Can be amplified by a factor of about ten thousand over just a single Gene mutation and this is the fundamental basis to why metabolites still are used in so many clinical assays Because it's kind of like looking for a needle in a haystack if you're trying to find a single base mutation that might be causing a gene Dysfunction or a metabolic disorder where it's often pretty easy to measure Certain metabolites, especially as they're amplified by a factor of ten thousand over the normal I mentioned before about the sensitivity of metabolomics particularly to what we eat to drink or breathe And this is just illustrating the effect of the metabolome on what someone's eating Within seconds, you'll see profound changes in the metabolome due to various physiological and metabolic responses And these are a consequence of enzymes acting on things consequence of organs changing And this will go up and down radically for Hours after you've eaten a meal in terms of your proteome whether it's plant or animal barely changes In the case of humans, it's just a few proteins and as I said before the genome does not change over time It's I mean it's supposed to be very static. It's critical for that Thanks to a lot of work that was done last century largely in the 30s 40s and 50s a fair bit of metabolism is understood And thanks, you know, we now have these cake pathways So we know about the catabolism and anabolism of of various molecules We don't know very much about the signaling, but we do know about the breakdown and formation of many many molecules And in fact the most complete sets of pathways still belong to metabolism Tablohms are connected to the genomes the proteome And the fact the metabolome is connected makes it particularly suitable for systems biology And it's a typical Paradigm, I think I've seen over and over that people doing metabolomics typically have the broadest understanding or largest systems understanding of biological system because in order to Measure the metabolome they also have to appreciate the proteome and the genome whereas people who study the genome are Largely oblivious to the metabolome and only marginally aware of the proteome and the same with people doing proteomics so just to reiterate on that small molecules whether it's the Nucleosides and nucleoside phosphates that are Obviously constituents of every piece of DNA and every piece of RNA Every protein is made up of 20 amino acids The lipids are absolutely crucial to cellular structure or tissue structure We also know that small molecules are the source of essentially all the cellular energy. It's not the proteins. It's not the DNA and as I said before the small molecules play a key role in both Keeping enzymes functioning and then you could argue that really enzymes only evolved to help hold certain small molecules in proximity So they could function more efficiently And then the signaling is critical for many metabolites So this last statement that the genome and proteome largely evolved to catalyze the chemistry of life and the chemistry of small molecules I think is true Living systems or life-like chemistry can occur in the test tube Those of you who are old enough to remember various reactions Mars Probes that they sent in the 1970s and 80s They weren't sure whether life existed on Mars because they were actually getting positive reactions that seemed to suggest life existed so, you know production and conversion of CO2 and H2O to sugars sugars to CO2 and H2O The replication of certain molecules these do happen spontaneously Those of that that's the chemistry of life, but to make it happen efficiently you need genes and proteins so Metabolomics because it has that connectivity and because it explores that connection between the genome and the proteome Helps enable systems biology and it helps people I think rather than thinking of Metabolomics and proteomics and genomics as separate entities to think of them as the same And the way that we're able to do that is through things like bio-traumatics And this is largely what this course of this workshop will focus on Obviously, we'd like to be able to teach people hands-on how to use mass back or NMR and GCMS But that usually takes about a week and we'd only be able to have about a third of you Taking that it would probably cost about ten times more than what you're paying for this So this is a way of at least giving you some flavor or taste of what it's metabolomics is applications Your introductions. I think a lot of people became aware or talked a little bit about how they're using metabolomics and their research There's lots of other applications Metabolomics is used in toxicology testing in a very large way People are trying to develop biomarkers in monitoring drugs and clinical trials phase three and phase four It's also used When we ferment things for wine and beer metabolomics is used quality test for foods and beverages also performed Nutraceuticals quality control assurance metabolomics people will phenotype Responders and non-responders using metabolomics drug field Water quality or environmental metabolomics is picking up One of the first applications of metabolomics is actually looking at oil From the petrochemical industry as an organic mixture produced by organisms The most widespread clinical application of metabolomics is actually an inborn error is a metabolism testing It is a true application of metabolomics and there are dozens of metabolomic biomarkers that are tested every day hundreds of thousands of times a day in North America and Europe Nutrient and nutritional analysis metabolomics Blood and urine analysis cholesterol testing that's all examples of metabolomics drug compliance monitoring transplant monitoring and a variety of imaging applications So those are just a few the list goes on it grows every year So there's a tremendous number of applications people may not use the word metabolomics and metabolomics But in reality it is So I'm going to switch gears and we're going to talk about the methods of technologies and metabolomics But before doing that again, were there any questions that that people had with respect to some of the background So this is a standard metabolomics workflow So it can start with cells it can start with tissues plants animals whatever So it may be a solid sample In most cases although not at all most cases people take those solid samples and will extract Something so the idea is to fluidize the sample Now in some cases you can actually work with a solid sample and That's fine. It's a little more difficult in other cases You may not try and perform the extraction. You may simply just take fluids that are excreted So in the case of mammals, it's urine and blood But you can also get excreted fluids from from plants and even from microbes And the reason why we like to work with fluids is because it's a lot easier to do the analytical chemistry on fluids We can run them through columns and put them into or inject them into mass spec So from the bio fluid, which may have been extracted from tissue. We'd start doing our chemical analysis And so it's typically HPLC or LC mass spec GCMS NMR And what we produce is typically a spectrum This is an NMR spectrum. It could be a GC, LC If I didn't have the scale here, most of us couldn't even tell what it was a chromatogram for an NMR or mass spec or GC or LC The data analysis is where we're going to be focusing on for most of today But for the next 20 or 30 minutes, I'll be talking about this chemical analysis approach so In the field of metabolomics this pyramid rears It's like we had a game But it's something that I think is is relevant to maybe one reason why people look down on their noses at metabolomics And this has to do with the quality and extent of coverage With genomics, whether it's humans or other mammals or even plants We can routinely sequence and measure all of the genes And it's something on the order of about 20,000 genes that we can find in most complex organisms We can also through transcriptomics measure the expression of those same 20,000 genes We go up to proteomics and we look at the same organism. Let's just say it's humans for state of argument a Good proteomics experiment can identify and semi quantify about 5,000 proteins Now with proteome human proteome, maybe a hundred thousand proteins So we're technically only measuring you know about 5% of it, but 5,000 is a big number. It's not as big as 20,000 On the other hand a very good metabolomics experiment will typically only identify about 200 compounds And yet I've been telling you at least in the case of human metabolomes It's perhaps several hundred thousand compounds that are there And and so 200 out of 200,000 is pretty abysmal So the completeness in coverage as I go from genomics to proteomics and metabolomics Scales as I've shown here so it gets progressively smaller even though The pyramid itself is essentially inverted in terms of the number of known components So that's a weakness Why is it an issue? Well problem is that It's an issue of complexity or diversity in genomics We can sequence genes because we just have to worry about the chemistry of four different molecules Which are largely the same and so we've worked out through enzymology, although you can actually do chemical sequencing Methods to characterize DNA very easily very routinely. It's a chemical-largely process Proteomics also is relatively easy because we're just dealing with 20 different types of amino acids And again, it's just one chemical class. And so we know the chemistry of how to break up Amino acids we can sequence by mass spec. We can also do chemical sequencing throughout in degradation So the detection the manipulation the processing of Genes and proteins is actually pretty simple, but the diversity and complexity of the metabolome is Many many times greater than it is for the proteome of the genome. And so that's why it's so tough So rather than one instrument which we typically use for DNA sequencing or one instrument that we use for transcriptomics Or maybe one instrument that we use for proteomics Metabolomics requires many instruments We use separation technologies completely electrophoresis liquid chromatography HPLC We use a variety of mass spec instruments high resolution low resolution. We use NMR. We even use crystallography We use gas chromatography, which they never use in proteomics or genomics We use a variety of detection methods including fluorescence and ultraviolet so just about every Every analytical tool has to be employed to be able to measure that chemical diversity So talk about some of them and maybe we'll begin with chromatography And again, many of you are probably familiar with it so dive into it a whole lot of detail but chromatography is obviously a separation method and And typically we'll be using these same terms a few more times, but there's a mobile phase, which things are dissolved in and They the mobile phase contains the metabolites of interest and then it passes through a stationary phase This is the the white stuff you'll find in most columns Could be silica or could be a sephidex or some other compound But essentially there's differential partitioning you're separating based on the fact that the the small molecules on the mobile phase interact with the stationary phase and They will partition or flow down at different rates So we have things like column chromatography and thin layer chromatography like liquid and gas and affinity chromatography There is a variety of ion exchange and size exclusion gravity and high pressure and ultra high pressure methods So the varieties are considerable and they depend on both their cost and availability The one that we mostly use in metabolomics is high pressure or high performance liquid chromatography Technology was done largely in the early 70s Pressures we work with our thousands of pounds per square inch And we use rather than what you have with gravity feed, which are particles that might be measured in Millimeters, they're typically working with microns or even sub micron particles Detection capacity actually with chromatography is quite impressive. This is why we can get down to potentially picomolar And chromatography can be used to separate both polar and non polar compound Most of us use reverse phase Chromatography, but that's largely restricted working with non polar molecules Hydrophobic ones lipids if you want it's characterized where we have a very Hydrophobic stationary phase and we use a fairly polar organic mobile phase Very few people use normal phase HPLC, but that was the very first form of HPLC that was developed It's not as good. It's separating In that it's partly because of the choice of a polar stationary phase and often a non polar mobile phase However, what's been picking up in recent times is a form of Normal phase chromatography called hyalic So it's a hydrophilic interaction liquid chromatography and instead of non polar molecules It's really good at separating polar molecules things like organic acids amino acids And it has a typically polar stationary phase and a kind of a mixed polar non polar mobile phase HPLC columns can be made up of different materials the very first ones are made out of glass Most of them now are made out of stainless steel material called peak Small columns are typically used For analytical work large columns for preparative work Most of what we do in metabolomics is done with analytical columns And you can see that the internal diameters can be as small as one millimeter and they can be measured from 20 millimeters to half a meter So really big columns really long columns Are typically the preparative ones So the particles those two or three micron size particles typically round made up of sort of porous silica so this is a The microscope image of those particles very very spherical But on them they're decorated with a variety of organic molecules. And so this is a Octodecanoic or octodecanoate or octodecane stuck to the outside of Silica beads, but you can also have shorter forms. And so this is what we call a C18 18 carbons C4 4 carbons, which is less hydrophobic there can be also Benzal groups that are attached again depending on the chemistry you can change What the retention times are what's attracted what the separation is the size and width the column the The size of the beads has a lot to do with how efficient your separations are So short columns are not so good for separating things long columns are But you'd like to use short columns if you have very little time because this takes a long time to do the separations You can still stick with very short columns But if you change the bead size say from five microns to something like one and a half microns or even sub micron You can greatly improve so you can improve the separation efficiency or resolution by going to small small beads Or long columns with HPLC the process is typically to take a solvent And this could be whether it's gravity feed or HPLC But typically with high pressure you'll have your solvent and then you have a pump Which is applying the solvent at very high pressures It's pushing things in and then mid-stream. You'll inject your sample Sample then goes into the column Things are separated and then you'll have a detector could be an ultraviolet could be a fluorescence detector It could be something like evaporative light scattering From there you'll measure it on the chromatogram and Process your data We can also use gradient HPLC This case we use two solvents some cases three solvents and if you're looking at metabolomics for Lipid separation often a three solvent gradient is preferred For each solvent you need both pumps and mixers Same process in terms of the analysis same sort of detection So the result is we get a chromatogram The intensity often corresponds to the amount material and the position has to do with essentially the affinity to the column matrix So stuff that comes off late at 30 40 and 50 minutes strongly attached to the column material Stuff that comes off at the beginning flows easily through and is typically more polar if we're looking at reverse phase HPLC Now in the case of UPLC Which some of you use I how many people actually use UPLC MS in their metabolomics? How many people have used HPLC in their metabolomics? So UPLC the the idea there is to work at much much higher pressures, but also much smaller Bead sizes and essentially as you shrink the beads you have to apply more pressure. So that's why it's called ultra high pressure ultra high performance liquid chromatography But allows you to get separations on the short columns in short times that you would have normally got using HPLC on long times long columns So now talked about liquid chromatography. There's another form called gas chromatography. This is what we'll talk about now It's another way of separating molecules And it's been used for a long time Decades, how many people use gas chromatography about half of you How many to use liquid chromatography in general? So the gas chromatography is not suitable for for proteins or DNA. It's it's ideal for small molecules And it's ideal for molecules that vaporize So you have to deal with typically either volatile molecules or molecules that you can convert to volatile forms So the separation isn't involving a fluid. That's not your carrier. It's actually or mobile phase It's actually a gas usually something like helium It's got to be fairly inert. So that's why we go for these noble gases The column itself is lined with some kind of polymer that's and then they these Molecules volatilized molecules move through and are partially adsorbed to the surface The columns are very very long often 10 meters in length, whereas in HPLC we're talking about things that are 50 millimeters or less Internal diameter very small very narrow a couple of millimeters The way that we get molecules to be volatile is we derivatize them with trimethylsilane and That allows them to stay in the gas phase so Trimethylsilane is done through sort of a two-step process. There's We'll take some compounds. Let's say it's a sugar and we'll do some methanol is this Methoxyne addition, but we'll try and get Some of these groups Methylated sort of stabilizes them And then we'll react them with trimethylsilane And we'll produce in this case for TMS or molecule containing for TMS components You can also end up where the reaction is incomplete and we may only have one with Say this TMS is not there or this TMS derivative is not there So the derivatization process is both an extra step that you don't see with HPLC and A complicating step because reactions are never a hundred percent complete So for a single molecule you may actually end up seeing four or five different peaks And that's an important issue you need to remember when you're looking at these DCMS Once you've derivatized and volatiles the sample then you're essentially injected essentially largely as well initially as a liquid and then it vaporizes very quickly and You push it through with helium instead of acetonitrile as you might with LCMS and you get the same kind of separation things that are Stick to the column come off more slowly things that bounce through the column quickly come off very quickly and this is a type of GC chromatogram and One thing about GCMS is you get a much higher if you want plate count resolution is much better than LC HPLC or UPLC So that's that's a very appealing feature of GCMS Peaks are very narrow and like LC the peak intensity roughly corresponds to the amount of material that's there and In GC the columns don't have to be straight They can be coiled because you're basically dealing with a gas flow and also because you're dealing with columns that are very long The columns themselves are lined with material. So they're not packed like they are with HPLC or UPLC So they are lined with a pile of polysiloxane polymer And that's what the molecules are interacting with as they bounce down through the column So there's a surface adsorption phenomena that you're detecting What you do in gas chromatography Because it is actually very reproducible far more reproducible than liquid chromatography is that you can use a thing called retention time or retention index To figure out what a molecule is So the retention time is just that time where you see the peak coming off So it's the time taken for an analyte to pass through a column So a typical GC at a run might be 30 minutes So you're seeing peaks coming off that you know 1.2 minutes 7.63 minutes 18.42 minutes. Those are your retention times So the retention time varies with the compound with the column with the flow rate and pressure But if you use the same columns same flow rate and pressure Same temperature and these are all described in every GC experiment You can from one day to the next day to the next week to the next year largely get the same retention times Now you can further improve the time by coming up with what we call a retention index And that's more general because columns have different lengths Sometimes there's a bit of variation So if you pass through a set of Elcane's six or seven Elcane's into your column you can calibrate your retention times and convert them to a retention index And that is largely universal so the retention index for a given column under standard conditions, which are usually reported and pretty consistent You can identify a compound based on its retention index alone Now that doesn't happen all the time and people usually insist on having more than just the retention index Provided, but it's a very powerful method. It's unique to gas chromatography so here's an example where we send something in we've got two peaks and Do this the next time with another sample? Here might be our calibrant standard, which we tells us how much was there And if we know it was the same that was injected and we can see here Obviously the area under this peak coming off at exactly the same time is much greater than this we can quite easily quantify how much is there and Use GCMS in a quantitative or GC in a quantitative way So retention time retention index tell us what it is the area under the curve tells us how much So gain just another shot of bio fluid GCMS If you compare that to the HPLC one much better overall much finer resolution the only disadvantage is that Many of these peaks actually represent alternate derivatives of the same compound so you have to then worry about separating or identifying those those extra derivatives So key to GC and critical to LC although not absolutely critical For LC is to have a detector So in LC we can actually use UV or fluorescence and in fact there are many many good Metabolomic assays that are truly fluorescent LC fluorescent based on athletes But for GC and obviously many of you Working on LC the detector of choice is the mass spectrometer So it's a very sensitive detector and allows you to further distinguish and characterize molecules based on their atomic or molecular weight So mass back how many of you use mass spectrometry in your metabolomics experiments So in the case of mass spectrometry, this is a typical one. This is a time of flight I Think most of you would know that through mass spec we identified compounds by their mass And many compounds can be uniquely identified by their mass if we have sufficient mass resolution With the highest resolution Instruments like FTMS we can measure below one part per million Which is point zero zero zero one percent of the mass With large proteins we can measure to within one Dalton for 40 kilodalton protein One atomic mass unit gain, which is usually sufficient to uniquely identify proteins So mass spectrometry is used in metabolomics and proteomics. It's coupled to gas chromatography It's coupled to liquid chromatography and then we also will use mass spectrometry coupled to itself So that's tandem mass spectrometry and that too is a very powerful approach for identifying Compounds by their fragment times So in the case of mass spec in the early days, and it's not so far and long ago even just 10 or 15 years ago Resolution of mass spectrometers was was not particularly great The ideal situation is you want to be able to measure the mono isotopic mass and this is an example of a high resolution mass spectrum of some High molecular weight, I don't know lipid And so we can see for this this is a corresponding mono isotopic mass This is the mass is the most abundant isotopes. So carbon 12 Proton H not deuterium Sulfur 32 These are all of the Most abundant ones and that's the mass but what you're seeing here are the isotopomers Which correspond to the addition of one proton roughly? Changing mass so that might be a C13 or N15 variant In the old days all of these peaks actually would be merged with your mass spec because the resolution was so poor And so what you were most interested in was the average mass Because that's all you could measure So you'll still see Basically to handle lower resolution Instruments people quoting both a mono isotopic mass and an average mass for given molecules So those extra peaks that you will now see with higher resolution mass spec as I say are a function of the existence of the abundant isotopes proton carbon C chlorine 35 Sulfur 32, whatever, but then there are the rare isotopes deuterium C13 chlorine 37 So if we look at something like chlorobenzene We would calculate its monoisotopic mass is a hundred and twelve point zero zero eight Dalton's and This is the monoisotopic mass that we'd see peaks roughly one Dalton apart Because there are either C13 derivatives or Deuterium derivatives or two Dalton's apart because we may have two carbon so 13s Two deuteriums or one chlorine 37 and then you can do this as well for the variations with other isotopes and other combinations The intensity of those peaks corresponds to the isotopic abundance so chlorine 37 is not a rare isotope Whereas deuterium is very rare Carbon is relative carbon 13 is relatively abundant. So that's probably the most common isotope that we typically see with Or isotopamer that we will see when we look at mass spectrum of metabolites So result is that you will actually get a distribution of peaks in this case for the chlorobenzene of about six different peaks With different intensities and different masses, but it is all the same molecule And so this is an issue when you are analyzing MS data with high resolution instruments where you have to essentially not view these as six different compounds Which people unfortunately do all too often you have to then de isotope The spectrum and make sure that you're merging or eliminating those peaks So that you can work correctly Quantify or perform relative quantification mass spectrometers are Basically composed of three components an ionizer a mass analyzer and a detector Typically when we look at a mass spectrum, this is a real mass spectrum. This is for aspirin we will see Peaks it looks like a chromatogram. It's just the peaks are much much narrower And rather than measuring say time We measure mass to charge ratio and we have an intensity Which is what we saw before with the chromatograms from GC and LC So this just sort of puts to words what I was saying before The height of the peak unfortunately in mass spec means almost nothing It has nothing to do with quantity it has a lot to do with How an ion flies And this has made mass spectrum This is the Achilles heeled a mass spectrometry and that it's not terribly quantitative When we talk about a mass spectrum we talk about resolution resolving power So the width of the peak is a measure of resolution In the way we measure resolving power is the line width of the peak or delta m over the mass so it could be the something we might call from Difference between two masses that we can separate so if we have two masses closely spaced Could we separate those two? so This is a peak you'll see in a mass spec if you zoom in as I say the peaks are very very narrow But they will have a width and we can look at that delta m And we can see that the two peaks here are resolved at say a 10 percent half height Most of us use a 50 percent half height and this is used historically both in NMR and in mass spec and in optics Which is where the word resolution first appeared and so we use a half height line width to resolve two peaks So this is an example of a resolution for both a high resolution and a low resolution instrument a toff is a high resolution Instrument and what we can see here. This is for a single compound same compound pure We're seeing about six seven different peaks. These are the isotopomers. These are the carbon-13 deuterium nitrogen 15 oxygen isotopes for this particular compound This is a low resolution instrument. This is an iron trap a Linear iron trap and this is providing you just with one big mound Where we essentially get the average mass of this compound whereas with this we get essentially the Mono isotopic mass which is this one down here and then all of the isotopomers so if we had the average mass Here 3484 here is the mono isotopic mass and these are just essentially examples based on a blue curve low resolution Where you have a resolution of a thousand red gain not very good resolution of three thousand so that's characteristic of quadrupole iron traps But then as you start getting into toff instruments or you start getting up into orbit traps or FTMS You get this remarkable resolution very very fine Details where all of the isotopomers can be detected So this is a more complex view of the mass spec where we talked about ionization mass analyzer detector But this is just you know, we separate we've talked about that We have vacuum systems which are critical for mass spec, but there are different ways of ionizing small molecules So Ionization methods are critical for metabolomics. They're also critical for things like proteomics in In metabolomics we have several choices unlike in proteomics One of the favorite approaches is to use what's called electron ionization. This is historically been used for 50 years or more and It's ideal for small molecules and it's coupled to most GCMS techniques Also chemical ionization methods gain also use for small molecules. They can also be used in GCMS and even Some LCMS applications Electrospray ionization, which is a soft ionization method, which is what most people use Because many people used to be in proteomics and shifted into metabolomics. And so they're reusing those up to those instruments And another soft ionization technique called Maldi matrix assisted laser desorption ionization And it can be used for small molecules But because the matrix that's used as often contains many small molecules in it You typically can't see things below about four or five hundred Dalton's so Maldi is somewhat limiting in terms of of doing Metabolomics So EI or electron impact ionization as it says the preferred method for GCMS And it's a very standard method that uses a standard 70 volt voltage differential so it has a No cathode setup Which draws electrons off of a plate the electrons come flying off at the 70 Electron volts and they collide with the analytes that are coming off the gas Chromatogram those electrons strip off the electrons of the analyte producing these positively charged molecules, but they also Crack up or break up the molecule in somewhat definable components So samples introduced by gas chromatography gas phase molecules as they come in Is that filament which might have rhenium or tungsten, but it's the it's the electrons that are being torn off those metals come flying into The molecules the gas phase The electron the energy of the electrons is much greater than the energy holding Covalent bonds of these molecules so you get these mass fragments and then they're sent out Through a repeller which pushes the positive ions out Through the port and that's where we pick up with the mass analyzer So if we looked at methanol thanks to electron ionization or electronic impact Methanol can be turned into a positively charged parent ion, but it can also be stripped of a hydrogen it can be fragmented into a Methane ion and a hydroxyl ion. It can also be fragmented further into this triply Triple bonded molecule and stripped off with another hydrogen So these are fragments that are known to occur in methanol And they have characteristic masses and this is what you would measure With a GCMS of say methanol and these this fragmentation is somewhat predictable There are people who they've been doing EI For a long time and I'm not one of them, but there are people I know who can kind of look at both a molecule or even a spectrum and figure out what the components are So they do the fragmentation in their head their programs computer programs that also can do some of this But this is an example of that if you wanted an EI mass spec Fragmentations of methanol where we can see the parent ion, but it's usually not the most intense It's it's all the other ones that you're fragmented Soft ionization methods. Did you have a question? It depends on the sensitivity and scanning rate of the instrument. So there's a few issues related to You know what you can see what you can't see The So, yeah, it's variable the So in terms of the soft ionization methods There's the Maldi approach Which is to put your metabolites or your tissue into a Siamic, oops, I haven't had to see cinematic acid matrix There are no matrix free approaches, but the idea is you blast your sample with a laser it explodes So it's hard to call it soft, but it actually is because the molecular ions actually remain intact or Electrospray Electrospray is the most common one in metabolomics in this case Typically we have our solvent coming from an HPLC or UPLC system. It flows through a very very small Capillary and then around the capillary you have essentially metal sheath Which essentially has Fairly high voltage difference between these two sides The result is that this high voltage and then this flow of liquid Causes essentially a spraying effect an aerosol. So if you've ever sprayed something from an aerosol can this is essentially what this is doing And it causes a cone to form Spraying out things in tiny tiny droplets which are then sent through a series of skimmers of different voltage Gates to accelerate the ions, but it's also a series of vacuums will then dry the solvent particles to very very tiny particles So here's your capillary Solvent comes out. There's this high voltage which sprays the Droplets all over the place they fly through a vacuum and they evaporate very very rapidly, but they don't evaporate into nothing They still have to have they are highly charged and as they actually evaporate further the charge density increases which causes them to sort of explode even further. So essentially get these tiny tiny ions hopefully Things with single charges mostly in the case of positive ion mode positive charges As I say just this is putting to words But as I say typically with ESI you want to have a fluid where you don't have salts That messes up ESI Whereas Maldi you can work with salty substances You push it through a metal capillary very low flow rates strong voltage With the gas that's coming around it causes you to get an aerosol and then you get this evaporation So if you have a fairly polar Solvent like water with strong hydrogen bonds You have to apply a very very high voltage 3000 volts to get a spray if you add a lot of acetyl nitrile, which is sort of the preferred solvent for a lot of work in in HBLC you'll start to see a spray much earlier if you're dealing with a gradient you may have to adjust your Voltage as the gradient moves through to get the spray that you want You can change the flow rate so tens of microliters down to less than a microlitre that's nano spray You can get away with very small amounts of material with mass back as I said things like salts and detergents mess it up Depending on your carrier solvent you can get things into a positive ion mode for say formic acid or negative So these are the different approaches for generating Your ions to measure your mass to charge ratio, and then you have different mass analyzers, which are used for separating those ions It's the original mass specs use Matt magnetic sector analysis MSA But the ones that most of us use these days are either quadruples or triple quads time of flight Orbitraps or FTMS systems, and they have different resolution We saw some of those pictures of those red resolving power for the different types of instrument This just puts it to text and words. This is resolution mass accuracy So your best is with the FTMS Second best is with Orbitraps and the cost of these instruments is basically proportional to their resolution or mass accuracy So that the cheapest instruments are these ones typically The most expensive ones are these ones These are rarely made anymore But certainly you'll see all of these and these ones So mass resolution obviously is important. We'll get to this later on in our discussions Whether it's GC or LC we get what we'll call a gas or mass chromatogram Things are being separated over time. We're detecting things so You're going to see a essentially what we'll call a total ion current chromatogram Which is essentially that intensity where we might measure over time the appearance of Peaks that were detecting by our mass spec So the GC-MS spectra I was showing you actually were the GC spectra I was showing you were actually GC-MS spectra because the detector wasn't UV It was it was MS and so each of the peaks we were detecting were masses That were coming off so we had time and then we had intensity So same with LC-MS we can have total ion current or tick chromatograms We can Simplify those to base peak chromatograms We're just displaying the most intensity or we can have extracted ion chromatograms And so these are different terms You'll see when people are talking about or displaying their chromatograms and these are sort of what you'll see in terms of what they associate with those So if we're looking at say an ESI chromatogram This might be for tomato or Arabidopsis and what's showing here is More like the base peak chromatogram So we're seeing the retention time 14 minutes and we're seeing the mass of the most intense peak 478 Dalton's Here's another one which is that 17 minutes and then the mass of the most intense peak is 385 Dalton's Now in many cases there may be several masses coming out from under these peaks and Often want to look at those to determine them because the resolution of the chromatogram is Often not sufficient to separate every single pure compound But by knowing something about the time and the mass we can often figure out what the molecule is or Perhaps which class it is So I'm going to switch gears from mass spectrometry to NMR spectroscopy. So how many of you use NMR? You had a question the resolving power. They're not the most of the most sensitive ones. So they're able to Allow to measure high mass accuracy Down to parts per million, but they're not necessarily the most sensitive Sensitivity has to do sometimes with the configuration of the instrument with the detector system Yeah, well, I think it's often defined by availability The I Know of no lab that's equipped with all of the mass specs or of GCMS Options Tens of millions of dollars so Yeah, it's largely defined by availability What what's available to you if you're happy or it's a situated where you can choose between any of these instruments? basically Orbitrop is the one that most people prefer for metabolomics but In order to do complete coverage and I think maybe the message for today is that you have to use all the different modalities so even Orbitrop LCMS Misses a lot of compounds that you can easily detect by NMR It also misses compounds. You can easily detect by GCMS and it misses a lot of compounds. You can easily detect by HPLC fluorescence detection And that's just sort of the nature of the thing If compounds are too abundant or too polar or your separation system doesn't do a good job You don't see it if the ions don't fly particularly things like sugars don't see it So things that don't ionize efficiently Like sugars like certain other kinds of molecules or glycosylated molecules just never see them by mass spec Or LCM aspect, but you can very easily see them Sometimes by NMR by GCMS HPLC so The coverage you get depends on the molecule depends on the sensitivity of the instrument And as I say complete metabolomics study typically means you should try and use all available platforms any other questions Okay, I guess the question is how many people actually have used or do use NMR spectroscopy in their metabolomics Okay So we'll get into this because typically people who do mass spec never use NMR and people who do NMR never use mass spec And so well a number of you were falling asleep while I was talking about mass spec maybe the people who know about NMR can start falling asleep, but NMR produces a spectrum not unlike what we see with HPLC, UPC, GCMS. It's a whole bunch of narrow peaks at different times What you do in NMR is you put samples under a strong magnetic field and then you send a perturbing radio frequency into that sample and the radio waves are essentially that's the electromagnetic radiation you're sending in are differentially or preferentially absorbed So it's a similar actually to if you've ever used UV spectroscopy It's just that we're using a different wavelength of light not not UV vis but radio waves and the only way that we can get the absorption phenomena happen is Putting it under a magnetic field if we don't put the sample under magnetic field the absorption phenomena doesn't happen And we scan and we collect a spectrum just like we would collect a UV spectrum And we get peaks that absorb over the length of wavelengths that we typically see So you can think of absorbance happening in red and in the green and in the blue and just like in UV vis and the result Is that an absorbent spectrum? So that's that's the analogy for NMR It's an absorbance of radio waves, but you have to condition the sample So put it under a strong magnetic field so the absorption will actually happen and more technically NMR is looking at nuclear magnetism So changes in the nucleus, so it's not a radioactive phenomenon. It's just looking at magnetism in the nucleus And we're measuring light radio waves that are and that absorption happens due to changes in Nuclear magnetism, which is result of nucleons having a spin as I said, it only happens when you're things putting things under a strong magnetic field so Protons is usually what we look at Which are found in hydrogen Which is what most NMR measures these days Have a spin so you can think of protons as little marble or sphere spinning And they can spin one way or the other way so we define it as a Spin up or a spin down and this is because protons have a surface charge and depending on the rotation of their spin Will produce a little magnetic field So if they have a spin up the north pole points up if the spin is down the north pole points down so again and think of Protons or any nucleon is being just a miniature magnet so when we send in Radio waves or or electromagnetic radiation And we have a sample in a strong magnetic field the radio waves will cause or will interact with the spins And will cause some of them that we're saying pointing spin down to go to spin up So this is a low energy state. It's a it's a bold spin equilibrium So not everything is pointing down some are randomly pointing up But you put in energy into the system things go to a higher energy state They briefly go to that higher energy state, and then they relax over time to their low energy state So in NMR we use big magnets and the bigger the magnet stronger the field essentially The higher the frequency we can look at So there are also different types of isotopes that we look at and so some Have a low ratio So even though we put them in a big magnet they are only the equivalent to looking at things in a lower field strength But increasing a field strength allows us to look at higher frequencies Which also gives us better resolution so modern NMR we typically will have samples these days often they have Flow flow injection systems or robotic systems. You can Take a sample out put it into this big magnet And start collecting data on the sample a big radio wave Transmitter receiver that's called a transceiver sends signals in takes the signals out and that information is then processed into computers You can get your spectra Modern NMR you can feed a sample in every five to ten minutes and you can do it day in day out Looking at hundreds of samples a day Characteristic of NMR is this big magnets refrigerator sized device They are measured in things like 400 megahertz 500 megahertz 800 megahertz gigahertz Magnets cost anywhere from half a million dollars to 10 million dollars for the very largest ones They are big containers containing Helium liquid helium and liquid nitrogen. They are super conducting magnets Magnets function measured at sort of 10 12 and 20 Tesla. It's a very powerful. They could easily pick up a city bus But they are super conducting magnets. They are not electromagnets. They are permanent magnets But you have to keep them cold if you don't keep them cold and they lose their magnet magnetism so if you're looking at a cut cut away view of a NMR system, you'll see that typically There's a liquid nitrogen bath. It's on the inside and the outside and then the very middle is a liquid helium bath The liquid nitrogen keeps the liquid helium cold And the liquid helium coats Niobium-10 superconducting magnet, which is made up of thousand hundreds of kilometers of wire So very powerful magnet and in the inside the magnet which might have a bore size of about Maybe 10 or 20 millimeters you will stick a probe the probe is what has all the electronics that allows you to send in radio waves So this is your probe Electronics wires going in and in the center of the probe is what's called a saddle coil It looks like two wires One half-shape like one saddle the other half-shape like another saddle put the two saddles together and allows you to generate an orthogonal magnetic field called B1 Inside the saddle coil you drop your NMR tube, which is typically about the size of a pencil very thin tube and The NMR sample which is maybe 500 microliters sits in that saddle coil and you pulse pulse is essentially creating a weak magnetic field which effectively to the Sample looks like a radio frequency electromagnetic field which Excites or provides energy into the system and that allows you also to measure The signal coming out from the tent sample to so this is both your transmitter and your receiver That detects or generates the radio waves that go in and detects the radio waves that are absorbed inside the sample So that data then is processed and it produces an NMR spectrum And the NMR spectrum has chemical shifts Should the positions here just like m over z values for mass or Retention times that we use parts per million. We also have a characteristic. So this is one compound, but it has five peaks whereas with You know LC you'd usually hope one compound one peak Or I'll see a mass There's certain splitting patterns couplings. We have chemical shifts Measured in parts per million and we have peak intensities the intensities. Tell us how much Splitting patterns that come from split spin coupling and the frequencies as I said are chemical shifts Tell us what type of molecule and molecular fragment is there? So NMR is different than mass back in the NMR Uniquely allows you to determine the three-dimensional structure of molecules You cannot determine the structure of molecules by mass spectrometry alone also Quantity in terms of quantification NMR is far more quantitative than mass back So these are two important advantages of NMR over mass spectrometry Reason why we can use NMR to determine structures is because of chemical shifts chemical shifts Tell us the composition because they're very characteristic of certain features and electronegativity of the atoms in a molecule So each compound has a unique fingerprint for chemical shifts But we can also use and predict chemical shifts with remarkable precision It allows us through the hot and through coupling patterns to figure out what is actually inside a brand-new or unknown molecule So chemical shifts for hydrogen or proton NMR range from around zero to 13 parts per million And these are the various groups that we can see in many organic molecules from aromatic groups to Carboxylic acid groups to double doubly charged or methylene groups or saturated unsaturated or amines and amides and These are the positions so many NMR spectroscopists have this sort of table memorized in their head And they can look at the positions of various peaks and figure out which component is which spectrum the coupling patterns also are Uniquely informative So in this case There's a spin spin coupling and I won't get into the theory behind that but it's essentially you get this splitting You get a one to two to one intensity So there are three hydrogens here The peak area of the integrated area is equal to three here The integrated area under this is two corresponding the number of hydrogens But because there are two hydrogens on this molecule they cause the hydrogens chemically equivalent hydrogens here to split into three peaks Because there's three hydrogens here They cause the chemically equivalent hydrogens on this one to split into four peaks So the N plus one rule and so again knowing this and looking for these very characteristic one to three to three to one or one to two to one patterns We can determine which hydrogens are where whether they're Methylenes or Methyne's and so on so these are all things that NMR spectroscopists can do based on chemical shifts and coupling patterns to determine what's there So in NMR we assign spectra We assign each of the hydrogen atoms in this case to different compositions So hydrogen atoms here around this ring all have chemical shifts around seven parts per million These two hydrogens here have chemical shifts around two and a half parts per million These ones have chemical shifts around one part per million and then there's this reference Which in organic chemistry is called tetramethylsilane Not unlike the trimethylsilane we used in in GCMS or use DSS or TSP Which are water soluble equivalents of TMS in metabolomics And these are the reference value that allows you to say this is the zero point and Just like retention index it allows you to give a universal chemical shift that everyone can understand Regardless of the make or manufacture of the instrument regardless of the field that you're working at So chemical shifts reported in PPM give you this universal number a mile post that everyone can understand NMR spectra are not perfect when you get them They have to be fixed up straightened up And this is an example of an NMR spectrum when you get it initially and you can see peaks But it's kind of messed up This is the water peak. This is from urea. You can see things dropping below the baseline You can see shifts that are kind of way off the scale What you need to do and you'll be doing this in the next Section is you have to reference Adjust things so that they're all Calibrated you have to phase so things don't have these unusual sliding Changes you have to get rid of the water peak. That's called water suppression You have to shim things so that the shape of the peak is perfectly Larensian and then you have to make sure that this side is at the same height that this side is which is called baseline correction So these are things that you have to do to manipulate or fix NMR spectra So this just Puts in words to what I was saying here, which is trying to normalize the chemical shifts just like we did with GCMS shimming face fixing the line shapes phasing to make the spectra or the peaks look Absorptive rather than dispersive To get rid of big peaks like water which mess things up and make the spectra look flat so this is an NMR spectrum of a biological mixture and If you took away the the scale whether you showed me a GCMS LCMS NMR spectra HPLC to my eyes. They all look the same It's just the scale difference and All you're trying to do in metabolomics is try and figure out which peaks are which and how much is there So they're different technologies as we talked about we've covered initially GC and GCMS We looked at LC and LCMS. We've looked at NMR NMR is the least sensitive method measures concentrations from Few micro molar to molar levels GCMS is slightly more Sensitive often you can get into sub micro molar levels, especially with the GC TOF Most sensitive method is LCMS So in terms of number of compounds you can measure by NMR It's maybe 50 to 100 GCMS maybe 150 to 200 LCMS the number of features that people can identify is sometimes reported in the thousands Getting flagged that we'll have to wrap up soon. I don't speed up or I've just got a few more slides So the The issue here is that when you are measuring in this range Most of the things you're measuring you actually know what they are When you're up in this range, most of what you're measuring you don't know what's there And this is a big problem with metabolomics. So it's the trade-off between sensitivity and what you know and Generally, you can't publish what you don't know. It makes it very hard to publish some of the things that we do by GC or LCMS, especially with the high sensitivity your most sensitive bones So that's a challenge. It's something that plagues a lot of us in metabolomics Here's the sort of comparison of the different techniques very quickly NMR allows you to measure mostly water soluble things all kinds of different tissues and samples Sample volume is generally large GCMS you can generally work with smaller sample volumes again The preference is for more soluble things but also for volatile molecules things that you smell and then MS Preference generally for hydrophobic molecules, but you can work with very very small volumes, which is very appealing Sample prep time varies for different samples They're set examples for DIMS can be as short as a few minutes per sample and a mar depending on if you do it manually 90 minutes automatically 10 to 20 minutes GCMS run time is also relatively slow because the columns are long limit of detection, you know micro molar hundreds of nanomolar low nanomolar range metabolites varies depending on the samples but generally more with MS less with these techniques But there isn't a whole lot of overlap So what you can easily measure by NMR you can't measure by MS what you can easily by measure by GC You can't measure by NMR and vice versa So using multiple platforms is critical to getting comprehensive coverage So in the case of NMR Depending on the fluid Looking at blood or cerebral spinal fluid or cell extracts you can measure about 50 metabolites in urine You can measure up to 200 metabolites and you can quantify them and identify them GCMS Typically when people are reporting both quantitation and identification the range is somewhere between 50 to 150 compounds Gain depends on the bio fluid most people don't quantify in GCMS. That's unfortunate There are kit based systems for doing mass spec that allow you to do quantification and Identification, but that's targeted mass spectrometry. You can get into the low nanomolar range LCMS methods most of the methods that are used can report up to three or four hundred metabolites being identified at the very best Most are not quantified, which is a problem Lipidomics there are techniques that allow you to measure up to three thousand left bits to And to semi quantify them So lipidomics is metabolomics of lipids And then if you're looking at that, you know other secondary metabolites often you have to use a combination of sometimes HPLC based methods Sometimes specific assays sometimes specific derivatization methods Many of them are LC based LCMS based So to wrap up here basically there are two routes to metabolomics One route is called a targeted or quantitative metabolomics I prefer the word quantitative because typically it means taking a spectrum mass spectrum LC spectrum or chromatogram and NMR spectrum and identifying every single peak and quantifying what's there The other approach is called non-targeted or profiling or chemometric methods where typically you don't really care what's in here at least and not initially but you compare this spectra and You group the spectra together And then you use chemometric methods to compare the spectra and you look to see how the spectra are different and Then from the regions in the spectra that are different then you'll focus on identifying what those peaks might be So untargeted typically take samples Collect many many spectra for many many samples you perform a data reduction principle component analysis say to identify the clusters and Then once you've seen if there is a difference and where those differences might be Then you go back and do perhaps your metabolite identification to figure out what caused those differences The other approach is the quantitative one whereas every spectrum you analyze for every sample you identify and quantify You have a list of metabolites use the list of metabolites to do your principle component analysis And then you go in directly to the biological interpretation So in either regard you eventually go from spectra to lists and When you have your lists of compounds and relative or absolute concentrations, then you go from those lists to pathways And so what we're going to spend the next Day and a half is going how we go from Spectra to lists and from lists to pathways, and that's the bioinformatics that we're going to focus on We'll talk a little bit of how we can actually go from those pathways perhaps even to models and also to biomarkers Which might be of interest to some of you as well So there's challenges there the challenge of getting to lists. We have to worry about data analysis alignment normalization data reduction To go to lists and pathways we'll have to talk about pathway maps and identification biological interpretation