Genomics and Personalized Medicine

Stanford University so it gives me great pleasure tonight to introduce actually one of our newest chairs at Stanford dr. Michael Snyder is the Stanford a Sherman professor and chair of genetics and he is the director of the center of genomics and personalized medicine he actually just joined the faculty after leaving Yale in July and so what a way to welcome him here he got his PhD at Caltech and then did his postdoc two here at Stanford and he's a world's leader in functional genomics I think you cannot walk around or drive around the Bay Area these days without seeing the blimp up there asking us all to get our genes tested and so tonight he's going to talk about genetics and personalized medicine so dr. Snyder ok well it's really a great pleasure to be back here after 23 years and I'm really having a blast once again so I'm going to talk about genomics and personalized medicine I haven't seen the blimp I'm probably the only person in Palo Alto who hasn't apparently and I will know I'll be looking for it this weekend right so this is an outline of the talk I will give a brief introduction on genes DNA cells these things I know you've had already I just want to make sure everybody's on the same page and if you're not you know raise your hand and ask questions and that sort of thing and then I want to talk about how we about sequencing genomes what's been going on and what we might expect and while we go through that you might just think a little bit about whether you want to get your genome sequenced and then I want to talk about the impact of genomics and genetics on personalized medicine and what we are seeing right now and what we might expect to come and I'll just say a little bit at the end Stanford is launching a new center for genomics and personalized medicine and we hope to be significant players in this whole area so just to give you a feel for what this talk is all about so I think all of you know this we start with from one cell and ultimately go into something that has 50 100 trillion cells and Arnold perhaps has more cells than the rest of us or else they're bigger I'm not sure which but anyway that's what we happens and along the way we make many many different cells and it's hard to define exactly what a different type of cell is but there's at least 200 basic types of cells and probably many many more in our body all carrying out very specialized functions as you see here we have nerve cells that help us think and fat cells that store things for metabolism and liver cells that do things and then of course every now and then some cells come out that go astray cancer cells is obviously one of the most significant ones and these often look closest to the cells they actually derived from but there they certainly are malignant excuse me and we'll talk about that a little bit so it's a DNA that's in us that determines what we are and determines what each of these cells are going to be for the most part nearly every cell has the exact same DNA there are a few exceptions and this programs these different cells and it also then ultimately programs us and we are different because of our DNA and as you probably know our DNA is housed in chromosomes and we have 23 copies and to cut right should say 46 total chromosomes 23 different kinds times to one format one set from our mother and one set from our father to give us 46 total chromosomes the DNA has you've heard from Gil and others is made of four bases or nucleotides as they're called these are distinct chemical entities and they're listed up there in big letters a T C and G and they string along in a sequence and form a very distinct sequence so a string of these nucleotides is called a sequence and that comes relevant as we talk about sequencing and what's shown here is one strand of DNA but DNA is normally double-stranded meaning there's a sequence on one strand and then you have it with called a complementary set of DNA DNA or bases or nucleotides on the other strand so there's a specific pairing such as or with T's on the other strand and T's or with A's and C's or with cheese and G's or Missy's so ultimately you get a two stranded structure such as the one shown here and it has a site pitch called a helix and that is in fact the structure of our double-stranded DNA that's in all of our cells so the human genome then is the collection of all of these DNA strands and we usually refer to one set of these chromosomes as comprising of the genome so the one copy if you will is three billion bases or three billion letters as is often the case and just to put this what this would be like if you were to put this in phone books this would be the equivalent of 200 phone books 1,000 pages thick so there's a lot of information here stored in our DNA okay and it's a fairly precise for the most part you'll see in a minute we our copies are relatively similar to not identical or we would all look the same but they're quite similar from one to another so not so long ago there was an effort launched the human genome project it started in 1985 and it was completed in 2003 and along the way there was a draft of the sequence made meaning they got a lot of it put together but it was still in pieces there was a massive effort it involved 2,000 people as you see here trying to get my arrow up here 2,000 people and the cost is it's a little difficult to estimate but it's somewhere on the order of 1/2 to 1 billion dollars the whole project and Counting technology development and things cost about 3 billion bases and it was often referred to as analogous to the moon project to send people to the moon this is the genetic equivalent the sequence one copy of the of the genome and this copy that was sequence is actually a pool of people they took a number of people pulled their DNA all together so that you would at the end identify one individual with this and the sequence they came up with is called the reference genome ok so that's the thing the first genome that got sequence out it's 3 billion bases that refers to one copy essentially of each of our chromosomes and it's called the reference genome and that was hot you know hailed as a huge milestone and it has been a huge milestone in completing this project before this time there were maybe a handful of genetic diseases that were mapped on the order of tens that were precisely mapped down to the exact gene as you'll see in a minute but after this project was completed it was a very useful framework to be able now we know hundreds of genes although there's still many more to know there hundreds of genes that cause diseases that have been mapped quite precisely so there's a big effort once the genome was sequenced to try and understand how do you get from one of these sequences again three billion bases of this to form a human being an equally important we went understand not only how you get from the DNA to form a person but what happens when something goes wrong and you don't get quite what you were looking for okay and so we want to try and understand the genetic basis of this and this is really clearly one of the biggest challenges in medicine today is to identify the genetic lesions ideally before they occur or before they manifest themselves as as problems and then try and set out corrective action that's clearly an ultimate goal of this whole area so that one of the critical parts of the genome as you might imagine are the genes you've all heard the term genes before there's only order of 20 to 25,000 genes in the human genome and you would think we know after we sequence the genome exactly where they all are and what they're all doing in such but we're nothing could be further from the truth we don't know exactly where all the genes are we know where a lot of them are we know it because we can compare our genome believe it or not with out of the mouse and the parts that are conserved is one of the good ways of finding finding where the genes are that's one of the ways in which you can do this or several but their genes are there they're scattered throughout our DNA and it's even more challenging you might think because these jeans they they're typically having 1500 of these bases or letters but they're they're actually broken into pieces that are fairly short and it turns out they only comprise a small fraction the parts that actually encode things that ultimately turn out to be the functional units which are or the the genes actually make something called RNA which in turn makes protein and those final functional units if you look at the genes that lead to those proteins it's a small fraction of our total DNA having trouble getting the arrows up here it's only one and a half percent of the DNA so there's a sea of these bases and our genes within those are a very small part and so finding those hasn't been so easy we found a lot of them but we don't always know where the ends are and things like that and that's one of the big challenges going on people are trying to define exactly where all the genes are in the human genome it's one of the things our lab works on in fact one reason you want to do this is because of genes actually make what genes are turned on determines what cells get made so in a nerve cell you'll have a set of genes turned on and make all the all the RNAs and the proteins you need for making a nerve cell and in the muscle cells you'll turn on some of them will be the same but a lot of them will be different you'll turn on things that specifically lead to making muscles and your fat cells make fat RNAs and proteins and so on and so forth so these 200 different cell types each make their own sets of each have a set of genes that get turned on to make specific proteins and so really one of the goals and is not only trying to identify the genes but to determine which ones are turned on in which cells and that's one of the biggest challenges ok any questions about any of this so far yep the DNA all the cell's is exactly the same upon all of the body only the difference between different sizes what is their nominal yes so the question is DNA is the same in all the cells and it's only a matter which genes get turned on or off determines which cells become which types and that's exactly right oh boy there's a whole bunch we'll start at the bottom and work our way up okay I wasn't going to because I want to spend a little more time talking about why we're differ from one another why you don't look like the person sitting next to you there are a set of proteins that I can give you a two-minute synopsis they're a set of proteins that will sit on in front of these genes and then they're responsible for bringing in the machinery that that turns on those genes and there are about 1500 of those proteins they're called transcription factors that are actually encoded in our human genome and it's really quite interesting different sets of those transcription factors are turned on in different cells and different combinations turn on different genes and so if you love combinatorial controls and thinking about nice complex problems this is the perfect one for you and there are groups like errors we have a huge project to actually try and figure out which of these controlling proteins is turning on every single gene in the human genome we've invented some methods for doing this and then it's a matter of slogging it through the problem is ideally you want to look at all 1,500 transcription factors in every single cell type in every single condition if you can it's a big challenge but there's a number of groups working on this so that's called the regulatory code to try and crack that so we now know the genetic code is a big push to try and decipher the regulatory code it's a great question yeah wanting you to treat one futures DNA yes say so what yeah I didn't explain that very well so there there are 23 chromosomes they actually have a different piece of DNA they're not all the same the 3 billion bases actually refers to the ensemble of DNA from all 23 chromosomes so that 3 billion is broken into 23 pieces that's correct it's 23 pieces yeah and they're as you might imagine they're quite large then they're huge by the way it's very interesting problem how you pack what's literally 2 meters of DNA into tiny little chromosomes so they're really super packed to be able to do that yeah they're I'm sorry yeah all right I'll do that going forward sorry guys I got a backtrack too much yeah well they all weren't sequencing themselves I'm happy to say they were actually they were working there's this pool of DNA that was all put in the same tube and then now from about 50 different people they mix that into one tube but then they made I wasn't going to give in details of this there was a two-step process but they basically broke it into tiny little bits it's it's an interesting the first step is to break it into 1000 piece bits and sequence those and sequence them at random and sequence in many many of these fragments and then how many you guys are computer types here Oh Lots so you'll love this so what you do is you take those thousand bases and you see how they overlap and your mission then is to assemble them all and in a perfect world if these were all unique sequences you could just put them all together and you have the whole human genome sequence the problem is it's not a perfect world we have repeats in our genome and so means assembling this thing is imagine a puzzle with lots of identical blue pieces that you can't tell apart in a sea of things that you can tell part when you hit these blue pieces you're kind of stuck so you need ways of actually assembling in I won't get into that and it's a great question it takes a bit of time so that's how a lot of it was put together and then there were ways of actually stitching things were put together the larger chunks and those chunks were stuck together yeah chromosomes given the genes DNA become proteins which become everything else and chromosomes are different collections of different genes what's the role of chromosomes role of chromosomes so chromosomes are both DNA plus protein and they basically keep things packaged properly so ultimately your cells divide and I think the Gil tell you about cells dividing and stuff I would have assumed so oh I forgot to repeat the questions all right all right so the question was what's the role of chromosomes and it really is it's a packaging thing just like you go in the supermarket you don't want the contents all spilled out could you imagine if we didn't have things in packages that's just like a cell you don't want you want to put things in their proper compartments yeah the other side comprises chromosomes there's a little discussion to 90 research being done yeah so the question is what about this other 98.5% so I said 1.5 in codes parts that we'll be coding for proteins and there's a lot of other stuff and that's a big mystery we we think a lot of it is this regulatory stuff so we remember what I said these transcription factors bind they actually bind to regulatory elements and they're very hard to find and pick out just by computer algorithms and things you really need specific experimental tests to find these things and so we think a lot of it is doing that but a lot of it nobody knows what it does and some people like to say well it's just there and we can evolve that stuff that is we can get new genes coming out of that so maybe many centuries from now there'll be some new genes popping up and such but the bottom line is we don't know what most of it does and some of it's thought to maybe assist in packaging and things like that you guys are asking great questions boy this could take a while here I guess haha alright well we'll do a circle here will come to you laughs okay yeah go ahead so finally genes is our transcriptor always like where they're located to find the transcription um most people think of it as the other way around there's a gene and then the transcription machinery comes in and expresses this gene that's the term expresses off or not the question is transcriptase I think I learned by now want you so the question is is is where the he called it the transcript where this transcription machinery is occurring is that where the gene is brought most people think it's the other way around where the gene is the transcription machinery is brought in that it makes this RNA and then actually the RNA goes this is all done inside something called a nucleus and then the RNA goes outside and gets turned and gets interpreted in the protein that way and something a process called translation okay yeah when we talk about sequences that refer to the helix and the nucleotides which come first in the combinations and going up and down the various rows and bases yeah let me rephrase this if I can so yeah what do we mean by sequence sequences really it's a distinct set of bases in a very particular order because those bases get interpreted so those 3 billion bases for the most part are in a very specific order again there are differences and that's what makes us different we'll talk about that in a minute but for the most part you know most the bases are in a very specific order and then as they get interpreted to make RNA and later protein that's very very critical so that they're making the same protein so the word sequence refers to that specific arrangement that's correct yeah so that was the hard part figuring out this 3 billion bases in what order they were in everybody knew there were four of them that was easy how do you put them together into those uh you know 200 phone books that was the hard part okay one more and then maybe I'll move on and we'll see if we can catch a fluid yes allied 23 chromosome it's a great question and you may know that different organisms have different chromosomes so you know Mouse and kangaroos and things and there's some organisms that will have lots of little chromosomes and nobody really knows why some of it may be evolutionary artifacts that is these chromosomes have broken up and then they stay that way their reasons once they get set up a certain way it's hard for them to change very quickly it's because it's very very important ultimately that chromosomes when we make our kids one set goes into you know the sperm and and same with the egg that you get one and only one copy of each chromosome if you start getting miss what's called miss segregation events if if an extra copy gets segregated improperly then the consequences can be quite deleterious either that individual will not grow if you have an extra chromosome or Down syndrome is the classic case where you have an extra copy of chromosome 21 so we don't know why there's that many as I say may just be an evolutionary artifact but it's very you can't have too many screw-ups as you go through evolution okay good well great questions I think that's a good sign we draw all on board all right so we'll start here identical twins that come from the same egg so an egg that has somehow split they don't always happen by the way at the first division it can happen later generally come out looking pretty much the same I think we all know that and that's because they have the exact same DNA but what about two different people so you again sitting Nick how are you different from the person sitting next to you and until not so long ago it was always assumed that there were just the single base differences called single nucleotide polymorphisms I'll spell this out on the next slide that there would be rare differences between one individual and that would make them different from the other okay and so here's a picture downloaded from the web this is some guy named Steve next to the basketball player Yao Ming and if I really asked you a few years ago if you asked any scientists what's the difference between Steve and Yao everyone would have said well it's these single base differences and the frequency of those is roughly one in every 1200 that's roughly how many differences there are but what that means is because we have three billion bases now some of them are repeat you'll see the math isn't exactly right but it's form there's about four million bases then that are different between you and the person sitting next to you sometimes these sit in these coding regions I mentioned in the genes but more often than not they're outside the genes and they could still be affecting regulation and such so this is until recently thought the major form of variation in in people recently though several groups I'd say lavich Cold Spring Harbor named Michael wiggler Charles Lee in Boston and more recently our lab has actually found these other kinds of events going on this is a big surprise there were hints of this before from a number of groups but I think the extent of us turned out to be a huge surprise and these are called structural variations so these aren't these single based differences these are big chunks of DNA to have a that either deleted duplicated or gotten rearranged typically inverted in the human genome so if this is that so-called reference genome I mentioned you have an order of DNA segments here ABC you can get a deletion whereas a B got removed you can get it usually it's a duplication an extra copy although you could have a novel sequence that does happen so you may get an extra copy say of B here and then lastly you can have an inversion in which two segments are flipped around relative to one another so our lab I won't get into the details of this but we invented methods for being able to actually find this in large numbers of these in the in the human genome and from that we found that there's about a thousand of these that are different between any individual and the reference genome and if you add up the number these how many bases are affected it's more bases than those single base difference as I mentioned before so in fact we think this is the major cause of variation in humans not everyone agrees with that by the way the people who love snips will tell you they're the most important but those of us who just like math say well there's 10 times as many bases affected by this so there's going to be 10 times greater effect from these think common sense is likely to rule out in the end but I'd better watch it this is on YouTube – I'm in trouble okay anyway we think that this will have a major form of variation yeah so it's not that snips aren't important don't get me wrong they are important but we think this is going to be at least just as important yeah well that's the thing that typically in a definition we call oh yeah the question is so the size of these things how big are these are they smaller than a gene are they bigger than G by definition we call them a thousand bases are larger and a gene is on the order of about 1500 bases well it's broken into bits so these are often at least smut may be smaller in size but they will hit genes you'll see in a minute but they can be incredibly large some of them are mega bases in size over a million bases that will differ between individuals so it's amazing that they're these big chunks of DNA that are out there that are different okay I have one in the back yeah so since you get half your DNA from your mother and half your DNA from your father as you might imagine you share half of that with your siblings then right because it's a random assortment of chromosomes you don't get all of your chromosomes from the other you get half same with your father so again you'll typically share half with your siblings no then you would have a half of that generally it's a good question yes sequences is like the median because those two people had four million differences and you're you having a thousand with the reference is a reference taken in such a way that most people fit in without so the question is is I'm not sure I understand the questions try to try me one more time here I put any good sequence no reference is a real sequence it's a sequence that came out of sequencing these 50 people okay and that's what your ultimate Lee comparing against so that's the thing we have sequence an alumn and this is important you're actually touching on the part I was going to get to later but when we actually start seek when you sequence genomes it turns out you don't just seek when somebody's genome brand-new you're always comparing it against the reference and it has to do with technical considerations most people don't realize this if I sequence your genome I'm not going to determine your AC GS and T's from start to finish all from scratch I'm going to actually determine bits of sequence I'm going to compare it against this reference sequence and I look for differences and that's how these are done so these thousand bases that were math they were mapped because their map is different in a person with some technologies and I'll show you they're mapped as being different from the reference sequence okay but remember that in some sense the reference sequence it is true it doesn't exist because it's a ensemble of 50 people did I answer your question there are four billion differences at single base there's only 20,000 genes yeah okay yeah cool nucleotide polymorphism I can see how that would maybe tweak a protein structure slightly and you have some sort of roughly equivalent for 18 and from one person to another but if you're sitting here flipping like giant chunks of a gene and how are you getting a final protein out that whatever the gene yeah so the question is he can see how single bases might be little tweaking mutations but these big structural variants are like sledgehammers if you will they should be clobbering genes every time they land in them and so that is correct and we've done that there wait we've done statistics on this to show that there are less of these on average in a gene and there are in other regions of the genome so that tells you you are selecting against them but these things do happen in genes probably very often they are deleterious and so those if it happens that person is not born and so they're selected against but I'd be a little bit careful you know a single base change inside a coding region can still destroy the function of a protein pretty easily so it's a bit of tricky business we happen to be believers that most variation actually affects gene regulation for the reasons you're sort of alluding to that is it's a lot easier to tweak regulation and have people live actually than it is to start tweaking the genes themselves because that can kill the gene and then therefore kill the person again not in the person itself but in their next generation typically well see if you will move to this one and I'll show you an example in a minute so the question is what's that doing to the genes so these things are all over the genome there are hotspots and some of these hotspots do map to diseased regions and some of these structural variants have been mapped at things like psoriasis and Crohn's disease and things like that so there have been some of these affiliated with diseases but because they've been relatively new in their discovery people haven't we don't know for way all the ways in which they affect human differences and disease so the way a geneticist by the way looks at disease disease if you think about an extreme end of differences between people we're all different but diseases are course all the way at one end of the spectrum there where often harmful situations come so a number of them do affect genes as I say it's it turns out at 17% and sometimes they'll just lop off a chunk of a gene or maybe lop off the regulatory region or flip it around and we have some new data that actually says they in fact can do just that but sometimes they'll actually affect the genes themselves and I'll this is a very interesting example what I'd show – this is a case of one of these structural variations that affects olfactory receptor genes so we have a remarkable ability to be able to smell an amazing number of different odorants and they trans do most mammals and other well virtually all organisms Oh animals and this is one of these examples here we found the case when we were studying these in which their two factor receptor genes we don't know exactly what they do they lie near one another and it turns out in some people this region in pink is deleted and these two genes are fused into one that's a structural variation a big chunk of DNA got deleted so there are people running around I guarantee they're sitting in this room there are people who have two of these genes just like the bottom part shows here and then there are other people who are fused into one so we actually followed this up by looking at all of the olfactory receptor genes they're 851 of these some of them are thought to be non functioning and that's a whole separate topic and we looked amongst 25 people we compared them all to this last individual and we asked this is these are the genes every one of these little bars is a different gene we asked whether there's an extra copy a loss of a copy or no difference across all these 25 people here and the genes arranged according to where they sit on a chromosome for chromosome 1 down to there's chromosome 22 and the sex chromosomes the X and the y and we looked across these individuals and what you're supposed to appreciate there are no two identical people in this panel and I suspect that that's true sitting in this audience that maybe there are a few but you get the idea we're all different in all factor receptor genes and so next time your dinner in your person next to you doesn't like the main course well there may be a genetic basis for that maybe didn't smell it anywhere as you think it's terrific or something so I do think this is and it's very clear there's Studies from a lot of people that people very much detecto Durance very differently from one another so we think there's likely a very strong genetic difference for this genetic basis for this yes the these structural variants as far as we know so the question is what's a timing for these structural variants to have an effect so if the question is when are these changes occurring these are mostly germline changes we think that is so they're occurring in our germline and passing them on to progeny there's an open question how often this happens what's called somatically in our tissues as we grow and develop and it does happen but it's much rarer so the difference is between two individuals will be much much greater than the difference between say your skin so and your liver cell there will be some differences we think we're trying to actually figure out exactly how many now it's something that's a great question something we're working on and we think they're there but they're not as common so these are mostly germline events okay yes what a saying means it's gained relative to this individual number 25 okay so individual number 25 the way we did this experiment is the reference in this particular X and so we see so there's a gene for example up here where there's an extra copy and individual number one and an individual number five there's one less copy of that relative the number 25 yep okay so I told you the original genome cost Elmo somewhere from half to a billion dollars a sequence we had machines that would basically process 384 samples at once there were they were the automated sequencers of the time and this wasn't so long ago remember this finish to 2003 what's happened is a very insightful there's a major funder of Health Research I'm sure you've heard of it called the NIH National Institutes of Health and there's a group in there that actually put out a challenge it said we want to sequence a genome a human genome for $1,000 and a group of company stepped up to the challenge nobody's hit that yet so don't get me wrong but what's happened is some remarkably inventive technologies have popped up that are actually starting to push this and we're now talking about sequencing things for much much less over two logs less than they were just a few years ago and these are just some of the companies that are out there and there's a whole series more in the pipeline that are going to even leapfrog the ones I'm showing you here but these this is the first of them four or five for then there's some of these code Illumina Life Technologies HeLa codes and here's the latest one to hit the scene and they all make claims and it's hard for me to tell exactly what the real numbers are but but I'll show you some numbers in a minute there I mean we can get a pretty good idea what the ballpark's are so what's different about these this will make now about 450 base 450 nucleotide long sequence reads these amazingly enough are very short when they first came out luminal Life Technologies they're only making 36 or so bases now they make 50 aluminum's making somewhat longer but these are relatively short reads they're only 100 bases right you've got to get three billion bases to ciphered so how do you do it well the way you do it is the way I mentioned before you get what's remarkable up these technologies they make very short reads but they make millions and millions and millions of them so a single run will in fact make somewhere on the order of a hundred and what is 160 million say fifty base or 100 base reads it depends on the instrument and technology but you can get very very large numbers and even though they're short and they often have high errors but you can get so many of them that you can find these single nucleotide differences pretty well the way you do that is again you don't try and sequence a genome de novo you match it to the reference and so when you get your sequence determined at the end of the day you're not really getting your sequence determined you're finding the differences between you and what they can match in this reference genome and that works really really well for these single nucleotide polymorphisms it's a lot harder you can get some of it for these structural variants but you can't capture all of it there's no good technology effort capturing all of that but you can capture a certain amount so we still have a ways to go to really be able to properly sequence human genomes and again there's some very interesting technologies that are coming in the pipeline that will get make things more and more accurate but nonetheless we can get a pretty good draft of a person sequence and then there's a big argument well gee does that count as a sequence or doesn't that count I don't want to go there but the bottom line is we can get a lot of information genetic information with perfect from a person even if it's not perfect yes well it's even worse than that because it got if you if you want the complete story there's a few people who dominated the pool and they happen to be so they're sequin is over-represented relative to others and as it turns out it's Caucasians so the reference genome has a lot more Caucasian information than that of other ethnic groups or geographical groups but if you want to get even the more complete story there's a big race on when this was all happening you may be familiar with this or was a public group that all it would mean ultimately made the reference genome that everybody's using but there was a private group led by a guy named craig Venter who marched ahead and was doing this separately and in the end he sequenced himself so the private group was Craig sequencing Craig's group sequencing Craig and the public group sequencing this pool so and that actually would then was in fact the first true individual that would seek the first personal genome that was sequence was Craig Venters ok that was there another question down is there any work being done to augment so the question is is there any work being done to augment the reference genome and the answer is yes in several ways first of all even when it was called done it wasn't really done and it's still not done there's about 300 I believe gaps maybe 2 or 200 or so I don't know the final number right now but so there still are a few holes there and the other is it's not done because it's not a real sequence if you start comparing with other people I'll get to this in a minute but there are other variations out there so there are projects that capture all the variation out there and it's also not done because we don't understand all the information that's in the genome the regulatory information we talked about earlier and again exactly where all the genes are and what all the genes are and certainly what they're doing we're a long ways from done there so there's there's much to be done ok ok so these new technologies that have come along as I say dramatically drop the price of sequencing genomes and in fact this red line then is the cost of sequencing as a function of time on the x-axis and you can see it's kind of marched along pretty good until these new technologies have emerged now I imagine many of you have heard of Moore's law right yeah does anyone so who can tell me what Moore's law is yeah so capacity doubles every 18 months so that certainly wasn't happening with sequencing but if you look what's happened with these new technologies they're actually increasing about tenfold every 18 months it's much much more dramatic than Moore's law ok so it's really going incredibly well what's happened then is the cost of computing storing the information and actually analyzing the data that's expected will be higher than that of actually sequencing the DNA pretty soon if not already so there's so you can see that the challenges are shifting from what they used to be and with the technologies now it's hard to get a firm estimate on this because everybody compares things a little bit differently but it's roughly on the order of about $50,000 to get your genome sequence with the limitations I just mentioned you get lots of these short reads you can figure out most of the snips and a lot of these structural variants but you won't figure them all out but you'll get a lot of information okay and that would be a Personal Genome so there are projects as I just mentioned a minute ago to actually now expand on this information so there's a project that's been launched to try and understand human variation across the globe and this is called a thousand genome project so there's a it's a consortium valving many hundreds of people to sequence a thousand genomes using these new technologies and trying to understand or at least capture all the variation that's out there and this is not really to necessarily understand disease just yet it's just to trying to understand variation that exists around the globe as I said before there's a big push on to sequence a human genome for $1,000 now I said not so long ago I'd probably get my genome sequence when the price hit ten thousand dollars but the way things are moving so quickly I think I'll wait a few years and I think they it won't take long before it's only a few thousand dollars I'm sure so but this is something you can think about and we'll get to this in a bit so with this then this now it's the blue line that's showing the costs in sequencing server for the color change with this dramatic drop in sequencing people's genomes are just starting to get sequence so that's supposed to be zero first individual genome sequence not the reference sequence around here and so on and so forth so now there's eight genomes that have been sequenced and I think everybody's expecting this to skyrocket and on top of this there's a number of diseased genomes that are getting sequencers a big project out there to sequence cancer genomes and sequence large numbers of these with the goal of trying to understand the genetic basis ease of lots of different kinds of cancers that are out there that's another large consortium project that exists that and some of these first cancer genomes have just been released and we don't know for sure the disease causing genes and those but there's some very interesting candidates if you will and the hope is that they'll lead to a platform for discovering new camp genes involved in cancer yes yeah this must be Craig huh so the question is once we get to a few thousand will that change the reference genome and not really but it does raise the issue you're hitting on a really really good point what should be the reference genome because right now it is a conglomerate it's a it's a one copy genome – it's not even two copies right and there's no really good answer to this what should be used as the reference genome and I've heard interesting proposals out there and it just hasn't been resolved and at this point the reference genome may just be the one we have even if it's imperfect and and it's in a sense it's a virtual genome because it is a conglomerate of individual sequences so I think because of the way everybody is mapping things but I think the main point is that we make sure we capture all the variation that's out there so we understand what each person's sequences at each position because at the end of the day what I would like to see us do and I think this may know these are the three well I get that and I mean we have many many sequences out there and I think this will be a very very powerful data set if we know the information associated with each individual along with the DNA sequence now there are other issues that are involved with what I just said as well that we'll probably touch on later in the talk yes expensive we get criminal activities where somebody's accusing the rapes winston's they check the DNA compare with some residual on the victim what are they really checking that was doing the see if it's not a sequence yeah so he's asking about criminal events when when during a rape for example what are they actually checking when they check DNA on on a victim and the answers are actually checking markers specific markers in the DNA so regions of DNA that do vary between people they're checking that and they're basically they check enough markers it is like a fingerprint if you will something that's characteristic that by probability can only be associated with a particular individual and not other people so if you're checking up markers nucleotides and destruction when you go it is yeah so at the end of the day those would also the question is is that part of the nucleotide structure and the answer is yes it's it's part of these nucleotide differences that I mentioned earlier it's not the whole genome sequence it's just little snippets but enough to tell you that one person is different from another yes you look for could you look to defy statistics or on each of the differences and ultimately look at those statistics to come up with a more probable reference genome yeah the question is can you use statistics to some knowing what's going on in each position and use that somehow to call reference genome I mean I think what you're talking about then is a consensus genome that is trying to say here's there since most people have an a at this position let's call that an A and since most people have this but to be honest that's also virtual genome as well because nobody's going to have the consensus that every one of these positions and then also I'm not sure what that means because again it's going to be based on whatever population you're sequencing as well so if we sequence a bunch of Europeans they're all going to have a certain sequence and that will be different from a bunch of people from Yoruba for example from China who will have a different one I guess if you want to do it straight all numbers maybe you would say the Chinese I suppose anyway let's yeah so you see the issue it's not that at the end I think the main thing is just to have something that you can compare against and what's most important is the functional information that's that's deduced from these sequences which is not that easy by the way we'll get to that in a minute so anyway these are some of the genomes that first got sequence we have Jim Watson craig Venter and we even have one at Stanford Steve quake just sequence his own genome with one of the technologies that I mentioned out there so you probably may have seen some press releases on that so these are three of the eight individuals that are there so I'm a believer that there will be infected billion human genome sequences done and this is something for you to think about virtually everyone in this room will have the opportunity to get their genome sequence if they want now you may not want that's a separate issue we'll get into in a minute but I think cost-wise this will be affordable for most people and then you might ask well why should we be sequencing genomes maybe I should have let off with this and there's a lot of reasons for wanting to sequence genomes one is getting back to that fundamental question we'd like to understand how a genome comes to life how do you get from DNA to a person what are all the parts how do they work together to be able to build a living organism and why is it our genome builds a human being and a chimpanzee builds a chimpanzee genome how is it they're very very similar to one another yet the organisms at some level are fairly different and so I mentioned before we'd like to understand why we're different from one another we look and I mentioned the species one just now and for this talk a big issue is we'd like to understand the basis of disease and try and use information from the genome to try and improve human health in various ways there through diagnostics and therapeutics and I'll touch on some of these in the remaining time here one thing I can definitely say is a genome I said it had huge impact on discovering genes that cause disease as I said earlier there's been a handful of genes that were identified that responsible for causing human disease you've heard of them sickle cell anemia cystic fibrosis fairly common diseases but after the genome got sequence and after we found a number of markers at the these single nucleotide polymorphism types of markers it was suddenly possible to map lots of different diseases and in fact there's now something on the order of over 200 disease genes that have been identified just in the last few years and this number keeps skyrocketing because as now the technology is so much simpler based on the information that's come out of this so this has been really remarkable for helping us understand and identify disease genes doesn't necessarily cure that helped us for a cure for those diseases but it certainly helps us understand the underlying genetic cases and then gives us angles we hope to be able at least at some point cure disease or at least diagnose disease so let's touch on this issue do you want to get your genome sequenced here's why you might you might want to because you might ultimately want to know your sensitivity to medications we'll go through some of this in a minute you might want to know if you're likely to get certain diseases but the number one reason if you ask people who have gotten themselves either genotyped meaning had markers with one of these services that you've seen the balloon flying around on say 23 Mir Navigenics most people do it because they're just curious they want to see what's in their genome and if you get your genome sequence you'll get even a lot more information why you might not well you might not want to know if you likely to get certain diseases and there are many incurable diseases that are out there that you can diagnose and do absolutely nothing about Huntington's disease is a good example and some people would not like to have that information and other people would and that there's no right or wrong to this it's just the way it is but it's up to you to decide if you're interested this one I'm going to touch on more in a minute most diseases are complex so even if you get your genome sequence you might think you're going to understand a lot when you get your genome sequence but the answer is you probably won't because we only know the genetic basis of say 200 or so genes that you can actually do anything about or even understand and most of those are incredibly rare so the odds of you having those are pretty low there are a few common ones that are that are relevant but there's not enough information I'll get back to this a little bit later to know how to interpret a lot of our genome information I don't think that will be true forever but it's true right now and lastly I'll also touch on this issue you worry about genetic discrimination from various entities and we'll touch on that a little bit in a minute are they only doing the tool energies that go or they doing that much those companies so the question is what about 23 me and other companies are they doing but are they doing a limited amount of sequencing another doing something different they're doing something called genotyping they're basically identifying genetic markers so instead of sequencing the whole genome they'll they'll identify for 23 Me's case will identify 600,000 markers which are these single based differences throughout the human genome and based on these markers they can say if sometimes these markers are linked to diseases like diabetes and things like that so they make estimates based on that what the probability is I think they're fairly careful about saying that these are not you know these are you might have a two-fold elevation for something for some particular disease over something else but they'll certainly that doesn't mean you're going to get this disease a twofold you know instead of maybe instead of being 1 in 10,000 now you're 1 and 5,000 chance or something like that you see it and none of the data or I should say there's some there's some of these markers that are fairly predictive but the bulk of them are not you know not strongly predictive shall we say so the information is it is somewhat limited although there is something interesting information but in the long run this information will be very valuable but right now it's fairly limited yes the question is how do you tell where you come from it's because they're markers which you can get either using these genetic tests or from your genome sequence that different ethnic groups share and it's pretty amazing there's somebody who is going to be joining the Stanford faculty soon who actually looked at this in detail by looking very very carefully around the world not through genome sequencing but doing fairly high density markers you can actually pinpoint to where your family comes from within about I think it's 170 kilometers 170 miles something like that within Europe and it raises this issue about when you get your genome sequence are you really are you going to somehow be identifiable okay because they'll certainly know where your family comes from obviously people move around so it's not exactly where you're living this minute but it's where your family you know has been living your larger families been living for some time so your ethnic groups have certain markers some markers are fairly recent and some are very old and that's been able that's been very very valuable for being able to trace human evolution and human migration patterns as we've moved out of Africa and populated the rest of the world yes this is genome sequencing pick up BRCA mutations and things like that any answers absolutely yes it also picks up well if you sequence your genome you will identify all that information and that I would call is incredibly valuable information you want to have because you if you do have mutations in those genes you will do you'll get yourself checked regularly or you should so that's something you do something about which is a good thing there are other cases for example a Bowie which is implicated in Alzheimer's some people don't want to know so Jim Watson when got his genome sequence said you can publish my sequence but not my a Bowie because he doesn't want people to know if he's going to get Alzheimer's now he's like 85 years old or something so I think we probably know the situation already how many far as I can tell he's still quite bright so yes yeah the question is is there some evidence we've had a little limited number of ancestors and yeah humans have gone through bottlenecks in our evolution there's some incredibly good groups at Stanford Marcus Feldman who's looked at human migration patterns and these sorts of things and he would certainly be a good person to talk to about this but the answer is yes and yeah type so yeah oh yeah so the question is discrimination and I'm going to touch on some aspect to that on the issue about releasing information this is a very serious issue that is out there so there's been a lot of different ways to approach uh some groups that have been involved in human genome sequencing like George church's only takes volunteers who were willing to put their DNA sequence out in public meaning they'll get it sequenced and then they'll put it in the public domain and they're not worried about it okay and he's got a I've forgotten what the fine the latest number is but he's got at least over a thousand people signed up to do this so then there's no issue for those individuals for others this could be an issue about consent issues and getting yourself if your DNA or your cells are used it's believed you should get proper consent and if you're giving up your samples for something and you don't want your genome sequenced out there you might make sure you see if you're giving up these samples about what's going to happen to them if you don't want your sequence out there just I mean these are issues because you only if they took one of your samples and sequence it's possible you could get identified from that and you that may not bother you but it may bother you and you don't want other people to know so these are issues there's other very interesting issues too because if you decide I will I'll get my sequence my genome sequence well you just released half the sequence of your siblings without their permission perhaps so there are issues like that or your parents get themselves sequence where your parents got sequenced and you didn't give your permission well if your parents get sequenced you just if both parents got sequence well your sequence is in there somewhere you don't know the exact combination but it is in there so you see the issues about this and my own view is that I I hope we can get to a culture where this is not not an issue that people aren't afraid to get their genome sequence and that we can share this that's a very utopian sort of view and I can imagine not everyone shares I will talk a little bit more about this as we go along all right so I wanted to spend some time oh we're in great shape talking about the impact of genomics on medicine and this is really when emphasized in it's incredibly nascent stage this is incredibly early days and there's much to learn and do ahead but I'll tell you where we are now which is not very far but we're trying to give you a sense of where we might be as we go forward so the biggest impacts on genomics are the first two listed here predisposition testing what are you potentially at risk for can't cancer certainly one of the most obvious and the other is something called pharmacogenomics using your genetic information help derive treatment or predict outcomes than that sort of thing and lastly there's this issue that we just raised about how one might use personal genetic information certainly the biggest impact in genetic testing has been bracket 1 bracket 2 which are fairly common mutations in women and you almost automatically know there's something there because there's a family history of breast cancers house first pen but now it's very clear Braca at least bracket 1 I believe bracket 2 have been implicated in ovarian cancer as well so if you have these lesions if it's running in your family it's almost always recommended you get tested for these and as you should because what will happen is if you are tested and you are found to be carrying these mutations you need to get yourself monitored very very frequently so that you can catch things cancer especially for solid cancers if you catch it early it's virtually always curable not always nearly always but if you catch it late it's almost never curable so the key is catch it early and so if it's clear you have a genetic disposition for this and you're telling these mutations you should get yourself tested quite a bit so you can catch it or and for some kinds of cancers breast cancer it's a lot easier to catch earlier ovarian cancer as you might imagine is really hard because it's not like you're going around getting your ovaries tested all the time so but if you know your abraca mutant you will get scanned and more often than not and therefore can catch things much earlier than you would if you did not know this so this is a clear case where there's a genetic test and it does have an actionable outcome that is you will you should be behaving differently than if you did not get that genetic test and if you know if your don't carry the mutation then you don't have to worry you should probably see it tested periodically but not at the same frequent frequency okay so it's a very clear case there are others that are interesting these are not very common this one I think is quite interesting there's something called long QT syndrome which is a series of it's a series of different genes that are out there that excuse me affects a your your heart rate and such and individuals with this can have arrhythmia issues and one particular type of this disease is actually if you get startled in your sleep these individuals will go into cardiac arrest so obviously you're tearing this you shouldn't startle those people so and also there are drugs that can actually work on these individuals and stuff too so this is a very very rare disease I don't know if I want to ask anyone to disclose what they have it but it's again it's unlikely but there it's not a common disease and then there are others coming along such as type 2 diabetes and such where there have been genetic tests as well so these are usually diabetes is almost always discovered to glucose levels and stuff first but you can subtype these at least in this particular case there's one here and there's getting to be more and more these most of these are rare again the brakus are the most common there's a new one coming out for breast cancer that's not quite as frequent as the BRAF mutations but has just been implicated as well so you'll see more of these as we go along and the expectation is again this list will get longer and longer so that's predisposition testing are you likely to get a disease other things you could get tested for are Huntington's Alzheimer's there's a number of Parkinson's there are certain genetic diseases there and some people again choose to get tested for those and some of these are incredibly strongly predictive meaning if you are positive there's very very high probability you will get the disease and others are less so so but anyway these are and and again some people choose to do this there's nothing right now you can do about those so other people don't want to get tested yeah there's a other question is is there a genetic predisposition for Alzheimer's so at least for one test the answer is yes these are often not single genes that are involved in these sorts of things that is so for some individuals yes you can predict be fairly strongly predictive of Alzheimer's with certain a bowi alleles a bowi I should say mutations or Maui's mutations are there called colon variants there was a time not long ago when medical profession believed you couldn't ascertain some they died Alzheimer's without a autopsy right yeah so because point was raised you can tell somebody died of Alzheimer's unless you had an autopsy and by the way so they're not everybody who gets Alzheimer's has these apoE so there are many other types of Alzheimer's that can come along with this just like not all breast cancers involve bracket 1 or bracket 2 if you follow so there are some genes that are strongly predicted but there's lots of other things that can contribute to these as well now timers is one of those so I I've forgotten what fraction of the population is assignable to apoE but it's not the majority I don't think I'd have to look that up so you can email me if you like okay pharmacogenomics so this slide is really mentally the landscape for the issue and then I'll talk about some specific examples in the next one so as you probably appreciate and sure you've seen this either for yourself or family members people often take medication and nothing happens and these are just some numbers here for some very very common situations hypertension depression diabetes for each of these these are you know large numbers of people were talking about you can see there's a significant fraction of people with any of these three diseases who do not respond to treatment okay we don't know why they respond to treatment there can be lots of different reasons one is maybe they're not metabolizing the drugs the same as other people and so on and so forth and we'd like to be able to understand those who are going to respond if you think about it to a treatment from those who don't you've saved a lot of time a lot of money in certain cases any lives if you knew who was going to respond to a drug and who wasn't so this is ultimately got very very important for the genetics of human health okay and so these are some of the factors they've written at the bottom here in which genetics can can influence treatments drug concentration and whether a drug is active against a particular target I'll explain this in these that so there are really two ways one is dosing drugs and the others will drug be effective against the target they're not exactly the same you know that's explainable here so we have proteins in our body that are encoded by particular genes called the cytochrome p450s these are actually very important molecules for they make our hormones they make they detoxify things in the environment and we use them for all kinds of different things and they are involved in drug metabolism there are many genes there 88 e d cytochrome p450s and actually they show differences in expression between men and women so if you and your mate man a woman don't respond the same it's because they may have real physiological differences these these are not expressed the same and men and women so there's now several cases the most famous as this one warfarin it's an anticoagulant used quite a bit and this is one of these cytochrome p450 genes of cyp2c9 sorry these genetics has weird names i tried to avoid that in this talk but this one does exist this is not a cytochrome p450 this vko are c1 it's a different gene it's a reductase that also modifies compounds and basically what these do depending there there's what are called two different alleles meaning two different versions running around in the human population was more than two but for each of these genes there's at least two different versions around one version makes people metabolize this drug a lot faster than the other version so if you metabolize this drug faster that means you need more of it for it to be effective so what they now have there's a commercial test that we'll test which of these variants you have and they know how to better dose you for the drug based on that test that makes sense now this is one example we don't know what most of these eighty do so you can see that ultimately we'd like to match every single drug up with every single metabolite ensign and then in an ideal world be able to match that that hasn't been done but that would be an ultimate goal of genetics to be able to match drug doses with the particular variants of these different alleles of these different teams that are out there so that's a drug dosing example there's another one not as common that exists as well how about targeting specific diseases well breast cancer that you hear this lot for cancer people throughout names breast cancer colon cancer leukemia these are rarely single diseases they're they're very heterogeneous there'll be several types they all look like breast cancer but this is a classic case breast cancer it can have something that are called her2 positive meaning they're expressing this protein it's a growth receptor called her2 and it there is about it depends on the population but somewhere on the order of about a quarter of people with breast cancer are her2 positive okay if you're her2 positive the best therapy is an antibody against her – okay makes sense but of course the other 75% aren't her2 positive it would be worthless to throw her two antibodies against those people it's very expensive first of all but second it would have little or no effect and so therefore you match the drug to the particular disease the exact disease you have now there aren't huge numbers of examples of these although quite frankly there are plenty of breast cancer patients around so actually it does affect many individuals in this particular case but there are many different kinds of examples I should say out there but it's the early days and so the expectation is that we will get better and better at matching these drugs here's another amazing one that I have a personal experience with this one ERISA it's used to treat non-small cell lung carcinoma and only a small fraction respond 10% but that small fraction turns out just about get cured it's amazing their tumor just regresses completely ultimately they made developers but at least for several years I should say this is usually diagnosed late so when people get this they usually have several months live so my cousin actually got this one he was 42 or so and so fairly young and it was diagnosed late and you know they had a certain window which to treat him unfortunately he was a non responder so they did go through an arrest a treatment he did not respond and he ultimately died but during that window you can see now they know that we didn't know at the time why it worked it's now clear that's very specific mutations this is a kind of diagram of a protein this won't mean much to but probably 95% of the people here I suppose but this is a diagram of a protein but the drug only binds to this protein if you have a certain mutation okay and certain kind of mutation so only the few people have those mutations these 10% will respond to the drug and they respond an amazing way it binds this compound and it basically affects this whole pathway this protein works in and cures the individual but the other 90% don't bind the thing so it's completely worthless and so now there's a genetic test if you have this they will test you to see whether you have one of these mutations if you do you take the drug if you don't you do something else so you can see the power of matching drugs to the disease why waste time on something you should be doing and again this is a matter of life or death this is not something you want to mess with this is you want to do it you want to do it right and so there are very many examples but this is what you know geneticist hope is the future we want to be able to match drugs to the exact disease and be as effective as possible yes in this say so the disease I didn't get into this but MIT's the the party line is that many forms of cancer most cancers arrive by say five mutations or so usually they can be structural rearrangements they can be point mutations in this particular case they're point mutations I forgot to repeat the question so the question is what do you mean by these mutations and what I'm saying is these particular ones are point mutations but cancer can arise by any of these things they can arise by structural types of variations and they ten arise by point mutation and every cancer usually is multiple mutations not always but usually what you guys asked great questions all right so you're not really supposed to read this but the point is you know it looks better here than on my screen all right so the point is there's now getting to be a lot of different genes for which either some sort of response has been measured it's still a fairly small list if you think about we have 20,000 genes and and the list at most it's not even 200 but or maybe it's 200 but it's not huge and most of these are often rare kid situations but we are starting to be able to match drugs and treatments with with diseases but again it's incredibly limited so far and I've shown you the best examples that are out there – her – and the rest and there's a few others like that so there was an interesting case yeah I think we probably have time for this but you know what one issue that's come up is can you match drugs to ethnic groups and this was controversial at the time it came out in 2005 there's a particular it's a call to combination therapy which means you had two things together to get an effect and it's called by dill and it's a combination of two things first it was tried in a general population it wasn't found to be successful so then they try to in an african-american population and did seem to have a positive response when compared to a placebo said here so has been used for a while it hasn't been a remarkably successful drug it also goes to show how you deal the biggest issue is that african-americans are an incredibly heterogeneous population and anyone the way this is used it's all by self reporting meaning the person's self identifies themself as an african-american or not and since African Americans uh are quite mixed often it's not the best way to actually be doing this and so it hasn't turned out to be quite as effective so it may not be the best way it shows how you probably shouldn't stratify a population you should really probably have good genetic test to see if you could really match this up up that was the lesson I guess I was trying to say so we'll see what happens that drug isn't being used a whole lot so just to recap what are the advantages of pharmacogenomics well as I've implied already you really it's targeted therapies you only want to give drugs to patients who will respond I didn't talk about this but an ideal world you want to give drugs to patients who will not have adverse side effects I'm sure you're probably aware of that there are a lot of drugs out there that are incredibly successful at some level but our problems for a select number of individuals and certainly Vioxx is a classic case right Vioxx has been associated with heart valve defects and there's no question there is that correlation but Vioxx actually helps a lot of people so in an ideal world you'd like to be able to stratify away those who have the adverse side effects from those who don't and I'd say you can do that with Vioxx maybe ten maybe can't but drugs get killed because they will have an adverse side effect on a small population and yet they can be incredibly beneficial to others and so there are a lot of people who are upset about certain drugs when they're pulled from the market because it's the only thing they'll respond to so you hear about all the negatives but there are positives that can be an issue I mention the issue about well I guess I just touched on item three you could therefore rescue drugs that may fail if you can figure out why they're failing on a set of PAH set of individuals and I met number two is a matter of giving the right doses talked about that already what are the problems with pharmacogenomics well probably the number one is economics if you think about a drug company doesn't want to stratify its market it doesn't want to make a diagnostic test to go along with every drug it's expensive doesn't always work to pain in the butt so what do you do about that well right now nothing I mean it's better for them to take it's it's a lot easier for them I should say to make a drug that everyone would take them to try and stratify the population I think as the competition gets higher though in certain areas I think you probably will have to stratify the population the second thing is that sometimes these studies if you think about they're typically done in one particular population it may not be generalizable across all populations that is a trial may be done in one area where there's one ethnic group that predominates for example maybe it works in European people of European descent but it may or may not work in those of Asian descent or in African descent and such and the trials may not have been geared to actually see that they may be heading towards one population another so these are things that you have to look at then probably one of the most important thing is something I touched on earlier I want to spend a few minutes on now is that most effects we think you're out there are not due to most genetic effects are not single gene low sigh so the Cystic Fibrosis you've heard about which runs very purely a family is a single cell locus sickle cell anemia but there's a lot of drugs out there that are not easy you can't assign a single gene to them very easily I'll go through this in a minute most diseases so I guess to recap if you low SCI are strictly predictive of disease where you have this you're going to get the disease when they're there you see them again but they're not that common most common most common diseases diabetes which is on the rise schizophrenia autism even asthma which is probably not and say as much these are very very common diseases right in the population amazingly common it's very very hard to find a Z the disease-causing genes let me give you one example diabetes they just did a recent study where they looked at 30,000 disease cases and 30,000 non diabetics and look for the markers to map getting details but to map all the genetic disease is responsible we know there is genetics causing diabetes so they looked at this huge number of people tried to map all the genes they found 18 genes involved in diabetes you add them all up and you can account for 7% of the genetic contribution of diabetes so a huge study can only find a tiny little bit and they're spread over 18 different genes so what that says is there's not just one dominant gene out there that's causing diabetes that tells you there's a combination effects we don't fully understand us but there's probably two major likely scenarios one is it's a combination of different genes you get the wrong combination you get diabetes and the other is that they call that gene interaction and the others that you might have genes and certain environmental situations so the genes plus a particular environmental situation these diabetes those are the two most popular thoughts for for diabetes and this is true of other things the easiest one to explain to you though is this one I listed up here about height the best way height we all know is genetics right you'll see your parents and such everybody knows belgians of the tallest and southern Europeans are shorter if you look at a European map very very obvious the best way to tell height is to look at parents people have gone in and looked at large numbers of individuals to map to genetics and again a few low-side been found and they account for this tiny little bit of the genetic contribution of height we can't explain height by genetics well and so looking at a person's parents is the best way to tell height okay fight of all the genetics and such then I hope that won't be true in the long run that is to say I hope as we understand things better we'll say oh there's 30 different genes involved in height you have this particular combination and all you who raised your hands about you know your computational skills we need you for these kinds of problems we see the right combination leads to whether you're tall or short okay yes I don't know the answer that my guess is no because it's hard to do that and the way the study was structured I think the answer would be no you'd have to somehow know that there are sets of individuals that don't get diabetes and it's somehow running in that family it's not you know because it's not that calm and it's not that easy to do that kind of study that there are diabetes as so as far as I know nobody's done anything like that there are other diseases that that has been done for where you look for people protected including for heart disease for example there's certain populations that seem to be resistant for heart disease amazing right so I forgot to repeat the question but sorry and I'm assuming your discussion of diabetes you were talking about type 2 not in and yes yes yeah thank you yeah so the question was type 2 I was talking about type 2 diabetes the answer is yes okay sure Kevin say anything about a genetic predisposition toward is there such so I haven't said anything about genetic predisposition to chemicals and addiction and is there such and the answer is yes there have been genetics associated with all kinds of things like addiction dyslexia so various learning things it's amazing what's out there and they're very very hard things to map a gene has been linked to dyslexia by the way and I know people are working on addiction I haven't heard of one yet yes is it possible that this is kind of all genetic position is actually only a very minor part of it that what's actually happening here is a systemic interaction with the environment chemicals and that it's almost random which you mention numbers those diseases in those cases you only talk about a very small percent right so the question is is it possible that's really the system such as interact the environment things like data that are contributing to that and the answer is yes but again remember there is a genetic basis to these diseases there is you can follow diabetes in families it's not always a hundred percent penetrant meaning it doesn't happen a hundred percent of the time perhaps and for some kinds of diseases but you do see it in families you do see it it's greater than random so there is a genetic component what I'm saying is when they map that genetic component they can only find some of it there still is more genetic component to be found but you're the point you're raising is exactly right it's not we think it's not just the genetic component it's either the interaction of a bunch of genetic components or the environment as you just pointed out can still be a strong factor and there's no question the environment has a very strong effect for many many types of diseases so but but there is a genetic component to this and again all that's saying is that this isn't simple that means there it's going to be combinations and it's our mission to figure out these combinations so that we can do something about either avoid those combinations or make sure we're prepared to be ready with therapies and things okay what we're careful monitoring of the disease in affected individuals okay we'll start back in yeah for you look in the genes for disease counts and you're finding little correlation many patients and a half percent of DNA well yeah great question so the question is because we're only looking at we're not looking at the other 98 and a half percent maybe we're missing stuff there and the answer is we actually are looking at a law that 98.5 percent that is a say the markers are all over the genome so you're still looking for this linkage and it's getting a little bit technical but the point is we act they are actually looked for so you would find it if it were they're the answer they're just small contributions either yeah I'll leave it at that so the bottom line is that we're not ignoring the 98.5% now there are times when you're trying to zoom in you find a region that's causing a disease and you try and zoom in and find the gene responsible so you find a big region a big chunk of DNA and in there's a couple of genes and you're looking through these genes to try and find a mutation and often you find it's not in the gene then it's often in that you know that other non gene part that you should be looking at and more and more people are learning to look that way so your point is very well taken did you have a question as well as we do ratings and the person things for this DNA structure change yeah so so the question is after this is to if after something is found and then a person takes drugs does DNA lesion change and the answer is it can so in the arrest case I mentioned what can happen is they individuals start developing resistance to that drug that is suddenly a mutation will pop up that will no longer bind that and then suddenly the tumour will start growing again so that can happen and into these resists and things do happen and unfortunately a lot of things they use to treat cancers and such reaction mutagens themselves so you do have to worry it does make people so often these are nasty chemicals and they can make people sick depending on the particular cancer in the particular treatment so that that kind of thing can happen they can other complications can arise you know certainly radiation that you know the goal is the questions is radiation these other things cause cancer and such so it certainly damages the cells and the goal is to kill them but it definitely has got to cause problems as well yeah it's banging on things you know with a sledgehammer again it's it's it's not at all targeted therapy like you'd like to have mutations into someone that didn't have it already and then cause to get cancer can blood transfusions bring things into people and cause cancer and the answer is that it would most commonly happen by an inadvertent virus being transferred such as hepatitis or something I can't think of a situation I guess theoretically it's possible I haven't heard of that but that doesn't mean it doesn't happen as a general you'd have to have these cells go back and populate the bone marrow or something any I worry more about a viral infection type of thing rather than some sort of genetic disease transfer right then transfer until you pass that on we get passed on to your children now these are almost always so the question is if you get a genetic mutation from an environmental situation would you pass that on to your kid and the answer is almost I would assume nearly in every case no because these are usually what are called somatic mutations they're happening in your more adult tissues and your germ line is usually protected from that okay so it would be very rare that that would pass into the germ line okay yes only on that there's a lot of discussion about average so actually passed and prosper so the issue is can epigenomic changes be passed down to offspring and again if they arose semantically the answer is no if they rose in your adult tissues but there's no question they're epigenetic effects that do get passed on to offspring and so identical twins are not exactly identical because of epigenetic effects is we clearly have a ringer in the audience here okay so yes black and white where gene stops and stars or and your so it's a black and white where a gene starts and stops and the answer is sort of so we're the coding sequences are the things that code for RNA that lead to protein most of that we can figure out where it starts and stops now those start and stop parts can change in different cell types and one cell you might use one start and another cell you might use a different one and another's same with the stop it could it can vary from cell type to cell type but you can map those who you can actually figure out where they start and stop at different types the part that's hard is these regulatory elements that I mentioned earlier they haven't been well mapped those are controlling the gene and so that's the part that's been very very difficult to pin down and they can off a lie large distances away from the actual structural part of the gene the part that encodes the RNA so it's a great question though yes is there any progress facing defense no is there progress on fixing a defect in the genome itself so the answers for certain diseases there has been good success gene therapy has been a real roller coaster ride for the sciences and for all kinds of reasons there's been some amazing successes something called ad a particular deficiency that has had looks like some pretty good success there was a case where they thought they had success for people who had lost their immune system because they're missing a particular gene and they added this gene back this is one of the biggest hallmarks of gene therapy only to discover when the added back the gene that preferentially landed in many of these kids were like eleven kids I think tried r15 something like that and four of them developed the gene the the corrective gene landed in a spot that actually caused cancer so they got cured of the immune thing but then they got cancer and nobody could have foreseen that at the time of them now I think there should be a strategy to avoid it but you have to go all the way back to the beginning there's been other situations where some of the vectors they use for transferring G weren't as safe as people thought and that's actually led to people dying and things not being managed as properly and those have been huge setback for the field so it's been this kind of thing so gene therapy just I guess I'll leave you with one comment it's a therapy of last resort if the therapy you use if you can't do anything else okay okay one last question because I want to cover a really fun story I think you guys want to hear go ahead twins and as they went through puberty that six months different and I they're they their fiber ologist so the test of their own DNA so they know they're identical but as a result one is little taller than the other and six months later okay so the common is she has identical twins that aren't exactly identical one went through puberty six months before the other one is taller and same individual weighs more than the other one yes they're not it I'm sure there are plenty of scientists who would love the study that your kids they're okay so I okay let me move on to I thought I'd throw up this issue it's a very interesting story about genetic discrimination and this has been a huge concern about there's the issue about genetic privacy so yes you may want to get your genome sequence but you may not want it out there maybe you want to store it in your own bank to me it's obvious that the person who owns your sequence is you and if you want to disclose it and put it out there that's up to you and if you don't that should also be up to you but there is there are big issues here associated so these are some of the issues about so discrimination I mentioned that the misinterpretation that usually is these complex markers so I want to tell you about a very interesting story about genetic discrimination so the most visible one I know of and it starts with there were a series of basketball players you know basketballs very stressful sport and there's been over the years at least three individuals who have died of a disease called hypertrophic cardiomyopathy perhaps the most famous as a guy named Reggie Lewis who played with the Celtics he was just you know starting to blossom and become a star for the Celtics and he died in a practice on a basketball court and afterwards they tested him but he did have heart problems before that I think they did clear him I don't know all the details but basically this condition is such that if you have hyper cardiomyopathy and you do Exec certain physical activity its potentially fatal and that was true for this particular individual so along came a another budding star named Eddy Curry who played for the Chicago Bulls and he developed a rythme ax and the Bulls basically said you either get this tester 8cm or you're benched okay so Eddy Curry decided not to take the test so the Bulls traded them to the Knicks and I knew he was a second leading scorer on the next big rebounder too and he obviously refused to take the test because if he would test the positive he felt it would have ended his career so this is certainly genetic discrimination right if he doesn't take the test he's off the team and that did happen so the big issue is then should an employer be allowed to require an employee to take a genetic test and sure that Iker you have the test done so you can make your own judgment here my own view I've you know it's interesting I presented this to an audience in New Haven to see what people thought I say how many think it a teacher you should have the test done most people raised their hand and said yes and there are few people said no way basketball is his life man that you can't do that to my own view as Eddy Curry should have had tests done should employee be allowed to require an employee an employer we allow but I'd be required to have an employee required to take the test and and it's an interesting question right so you have to think about it a minute to me the answer is that again you shouldn't require but they probably should have you know why they did it of course they didn't want him to go out on the basket port Paul Court and die and therefore be liable so to me the obvious thing would have been to have them sign a waiver or something like that but I had to think about that I didn't sit there like you guys try and do it off-the-cuff so but you can see so I think most issues about genetic discrimination are obvious but there are a few Gray's you know how about airline pilots and narcolepsy and things like that so you could envision a few scenarios that might be the case so but for the most part it's pretty obvious I think you know employers and my mind should be allowed to do this so but nonetheless say it so one is the employer then the other is this health insurance and so recently a law has been passed in Congress that you can no longer discriminate based on employment or based on health insurance and so that is the law so that just passed recently now does that mean it doesn't happen I don't know right so just because something's a law doesn't mean it doesn't happen but one thing that was not covered is in fact life insurance so if you think about it that's pretty relevant if a life insurance company were to ask you whether you had certain conditions and you had disclosed in the public that you in fact had these well you could actually perhaps not be considered so there so it's not the kind of non-discrimination Act that I would like to see but it's a big step forward anyway oh good we have five minutes so I do want to just let you know that you know it's Stanford and one of the reasons I came here is that we really want to be leaders in this whole area in virtually all aspects I've described and so I'll give a little selling point for what we're trying to do here and feel free to make comments either as we talk or maybe at the end they can certainly wait outside or something so we are launching a new Center for genomics and personalized medicine because I am a believer that anyone as I said before anyone who wants to get their genome sequence will and I can envision a time I'm not saying it's going to happen but I can envision a time where kids are sequenced at birth you'd certainly get genetic tests at birth now for PKU and certain things and if they're worth discrimination I could have vision the time when that might happen and if you think about there are actually ways in which you could get a genome sequenced from a kid before birth because some of the fetal cells are circulating and the mom you could actually fish those out and in fact sequence genomes and this isn't science fiction this is hasn't been done but it's not so our way you can do genetic tests in the mom for the kid okay so this will be coming down the line it raises a whole series of issues but these are issues that we need to face just because they're serious doesn't mean we should ignore them it means we have to deal with them and decide what responsible policy is so this is coming as I like to say the train has left the station you there Pete lots of people were thinking well how do we manage this let's put everything on hold it's not going to be on hold people going to charge ahead you know individual liberties are strong in this country if who's to say I can't get my genome sequenced right that would be discrimination I don't want somebody to tell me I can't get my genome sequence so those are issues so what do we want to do at the center we want to Stanford has a lot of amazing strengths it's got very strong Technology Development we have people who know how to apply this we want to make very key discoveries of genes sorts of things and therapeutics and we want to move them into the clinic as quickly as possible and I think vision this is a major goal of the Center to go right from discovery to the clinic and keep developing new technologies to push the envelope we have a lot as they say strong faculty to do it we're uniquely suited because there's a lot of really very strong and innovative companies in the whole Bay Area many whom are eager to work with us to try and achieve this goal so I think this would be an incredible thing to try and accomplish here and it's something as I say that's drawn me here to be able to pursue this there are going to be many aspects of the center there's going to be genome sequencing there'll be issues about how to interface it with the clinic there'll be issues about the ethics and there will be issues about education our doctors of tomorrow have to be able to deal with everything I just told you I mean again people are going to come up and ask about genetic tests and if our doctors don't know how to do this we haven't trained the right people or we haven't trained them right I should say would be a better comment so the point is we have to change the way we educate future doctors in this country so all of them have to have a reasonable amount of gin competency and going forward so one big issue one big thing we're doing is launching the sequencing center it's already moving these machines were we're bought some of these alumina machines and we have discussions underway with a number of other companies including 454 and this is not cheap stuff these machines I didn't say but they cost a half a million two million bucks okay so and then you pay 50,000 to get a sequence on top of that so it's very very expensive things but we think the payoffs are huge and the prices are going to drop and they're more companies coming down the pipeline Pacific bioscience I've listed here I haven't mentioned it before they're clearly one of the promising technologies that will be out very soon the other thing is that gets back to this issue our genetic predictive value is quite poor in many cases and we want to try and bring other kinds of information in for example maybe understanding the sequence of a person when combined with other sorts of information like blood tests I think the blood tests we do now are not at all going to be like the blood tests in the future we look at a few things tens of things we should be looking at thousands of things to really understand what's going on and we should deal with something that looks like a cell phone okay so you can do tests and maybe you'll want to do it yourself that would be the real future so that you can quickly with the combination of genetics along with really comprehensive tests try and understand what's going on so if you're not feeling well that day you run one of these tests and look at 5,000 proteins and say gosh darn it I've got a swine flu you get the idea so I think we can go buy in other kinds of information the way we're doing it is it's just it's caveman compared to what it will be I think 10 years from now and so the goal at the end of the day will be to combine different kinds of information genomics I didn't indicate up here proteomics where then we've gotten into this but where you look at thousands of proteins and try and predict so-called normal from disease states here and I'll just leave because we're out of time we want to look at lots of different disease areas as part of the center certainly all the major diseases and this is by no means a comprehensive list this is just a subset of things we'd like to look at I'm sure you can probably think of your own favorite disease and those people here who are very interested in it so I guess thank you very much for having me here so I think you can see why we recruited dr. Snyder here and he has graciously said that he's going to stay after and answer some more questions next week you're going to have dr. Julie thereƶ she's going to be coming into the micro world please read in your text so you're ready for that and we'll see you next week for more please visit us at


  1. Regarding the discussion at 1:24:00 on Type II diabetes, Prf Sapolsky notes in one of his lectures that a "wasteful" metabolism can be a preventative element. That is to say that people whose genetic history is "thriftier" are at greater risk because their bodies don't flush out the crap as quickly as Westernized folks do. This is seen especially when traditional cultures encounter Western diets. Hypertension, obesity and diabetes go way up in those cultures.

  2. check out a short witty video defining personalized medicine made by PhD students called "The greatest drug in the world " on the genomics education youtube channel

Leave a Reply

(*) Required, Your email will not be published