drug thalassemia

richard wilson:thanks, larry. so it's a pleasure to be here on this occasion,the opening of the exhibit that salutes the work that's been done, the field of genomics,and the promise of genomics. and i think my job is to give you the perspective of somebodywhom, over the last 30 years, has gone from being one of the foot soldiers in the trenches,to trying to figure out how to actually do all this dna sequencing stuff, to about 10years ago, becoming the director of one of the three large-scale genome centers thatare funded by the nhgri. so about in the mid-to-late 1980s, i was apost-doctoral fellow at cal tech, working with a guy named leroy hood, and the groupthat i was working in was trying to develop

automated dna sequencing. we had been doingdna sequencing for several years. and we used lots of radioactivity and things like thatto actually get results. it was slow, manual, hard-to-do, and we didn't really get all thatmuch data out of every experiment that we did. so we set about developing a technologythat was much safer, much more automated, much more efficient, and, hopefully, costquite a bit less than it cost to do sequencing the way we did it in the mid '80s. and while i was working in the lab, my bosswould occasionally come back from these meetings where they discussed the possibility of sequencingthe human genome project, and what did we think about that? we said, "that's a greatidea. we think you should continue going to

these meetings." but we'd get together atour post-doc lunches and say, "this is crazy, but what the heck? sounds like a good idea.and, yeah, you know, we can kind of see what the benefits would be. so i think we shouldgo for it." in 1990 i moved to washington university inst. louis, with one of the first genome sequencing grants that came from what was then knownas the national center for human genome research. and our grant basically had two goals. thefirst was to begin to sequence the genome of a model organism called c. elegans, whichis a little round worm, that is interesting mainly because it's taught us a lot of importantbiology, including apoptosis, which is programmed cell death, which has become very importantin a number of human diseases, including cancer.

as well as to start to build the infrastructure,the methods, some of the software, et cetera, that would instrumental, indeed critical,for actually thinking about sequencing, even the first human chromosome, let alone thehuman genome. so this went well. back in those days we diddraw parallels to the space program, as eric touched on. and, you know, we made a goalas a nation to land a man on the moon and bring him safely back to the planet earth,only after we, you know, realized very modest success in launching a few rockets, of fewof which didn't explode, and the most recent at the time, of which actually carried a humaninto suborbital flight. and that's kind of the same situation thatwe sort of saw in front of us, so we had to

sequence three billion base pairs of genome.and we really needed to do about six times that. that's what we needed, we believed,to be very accurate, so about 18 billion bases of the genome. and at the time, in 1990, oneperson could go off, spend a day in the lab, and generate about 12,000 base pairs of dnasequence information. so, we figured we needed about a million and half person days, about4,000 years, to actually get the job done. and it was expensive sequencing. this is whywe just didn't get a thousand people and expect them each to spend four years getting thejob done. you know, the goal for the genome projectin the early '90s was to sequence a genome for less than a dollar a base. and i can remembergoing to the cold spring harbor meeting, i

think in about may of 1992, and reportingon the beginnings of the c. elegans genome sequencing project. and we had done abouta half a million bases, which was amazing accomplishment in those days, but richardgibbs, who is now the director of one of the other centers, raised his hand at the beginningof the q-and-a and said, "how much was the cost per base, rick?" and i said, "well, itwas $17 a base." now, there was a lot more than sequencing that had been done with thosefunds, a lot of development and so forth. but that's about the level that we were at,and we still had a long way to go. and if you think about computing, which wasanother challenge that we had back then, i actually had a portable computer in 1990.it didn't have a battery. you couldn't call

it a laptop. if you actually tried to usein it on your laptop, you might injure your leg. you could take it home and plug it in.but if you think about what we have, you know, 10 years later, and about the time the genomeproject was finished the first time, the advances in computing technology are incredible. but we did it. by 2000, we had a draft ofthe human genome sequence. by 2003, we had a largely finished encyclopedia of the humangenome that we could use. sequencing technology evolved rapidly. but large-scale centers werereally critical in this project. there were several large centers in the u.s. at the time,funded by either what had then become known as nhgri, or by the department of defense,but there were also centers in many other

countries around the world, including thesanger centre in the u.k. at the peak of our sequencing work in -- about the time genomeproject was finished, we had 135 of these expensive, complicated dna sequencers. as eric touched upon, one of the key thingsthat we had going for us were companies like applied bio systems, who was really the majormanufacturer of the key technology throughout the '90s. the companies could take technologythat was developed, sometimes in academia, sometimes at companies, and commercialize,and, most importantly, harden these technologies into boxes that we could use in the labs toget the job done. they were supported, and they were repaired by the companies, but thecenters really had a key role in sort of focusing

on the methods, the applications, the softwaredevelopment; all the things that really made those sequencing instruments a powerhouse. so by 2003, we could sequence a human genome,as eric sort of said, for something in the $15 million to $20 million range. i wouldargue that it probably took a couple of years, at least, rather than a few months. but assequencing technology has continued to develop, in the years since the human genome projectwas finished, it's been amazing. the biggest breakthroughs came in the mid 2000s, as westarted to get our hands on what is known as next-generation sequencing technology.and today, in 2013, we can actually sequence an individual human genome and analyze thatsequence for a cost of around $10,000, and

we can do that in two to three months -- sorry,two to three weeks. and i'll talk a little bit more about this in a minute. well, since 2003, we've really focused onseveral crucial next steps that have really built upon the human genome project. we continuedto refine the reference human genome sequence. it's the gold standard by which genomic sequencingand medicine is currently performed. it's what we use to understand the biology of genomes,and of human cells and animal cells as well. it was about 98 percent complete in 2003 whenwe declared victory. some of the regions that remained unfinished were essentially impossibleto resolve with the technology and the methods that we had then. and we knew that some ofthese regions contained important human disease

genes. several of them contained the kindof variation that eric touched upon in his talk, which we referred to as structural variation.and those regions are characterized by very repetitive elements, and are just almost impossibleto sort out with the technologies that we were using even six or seven years ago. well, we've resolved many of these throughthe aegis [spelled phonetically] of the genome reference consortium. my lab in st. louisis part of this, as is the sanger institute, the national center for biotechnology information,and the european bioinformatics institute. and by bringing in some of the new technologythat's come available over the last couple of years, we've laid in plans to sort of takethis gold standard to a platinum standard.

the other key thing about the reference humangenome sequence is that it largely comes from a single individual. as you just heard ericvery nicely illustrate, there's a tremendous amount of variation within the human populationaround the globe, and it's really critical that we start to capture more and more ofthat. 1,000 genomes project was a great start, but we need to get deeper into key ethnicgroups, and really build sort of that utility into the reference. this is going to be absolutelycritical as we move ahead and start to bring genomics into the clinic. with projects like encode, we've come a longway in understanding all that non-coding dna. there are regions where, as eric mentioned,they code for little bits of rna that have

their own sort of atypical function. thereare places where important proteins bind with the genome to turn genes on and off. and iwould argue that, you know, an understanding in diseases like cancer of what secondaryfunctions turn genes on and off at the time, or in the right, or even more concernedly,in the wrong cell type is perhaps more critical than the mutations that we find in the genesthemselves. and with projects like hapmap and 1,000 genomes, we've really made a hugeleap forward in starting to understand the diversity of the human population. it's goingto be absolutely critical to better understand this so we can start to understand why somepopulation groups are more susceptible to certain types of diseases.

and by sequencing animal, and pathogen, andplant genomes, and understanding their biology, and comparing their genome sequences and theirbiology to our own genome sequence and our own biology, we've come to better understand,not only ourselves, but the other living things on the planet that might help or harm us.again, nhgri and the large-scale centers have really been key in driving this work forward. when the human genome project was proposedback in the mid-to-late 1980s, one reason to do this was that we would better understanddisease. we would accelerate our ability to diagnose disease effectively, develop newtreatments, and eventually cure many human diseases. i'm proud to have played a rolein two of the signature projects that are

both represented in the exhibit down on thesecond floor. and these had made very, very fundamental use of the human genome referencesequence, and the methods and technologies that we developed to get that to that point,and in the years since. so the two i'll just give you quick exampleof are tcga and hmp, known by acronisms -- [laughs] known by acronyms, as many things at nih are.you heard about tcga, the cancer genome atlas, which was a project -- a joint project betweennhgri and the national cancer institute. and basically, what's gone on in that projectis that the large-scale centers, sequencing centers, have collaborated with a number ofcancer biology labs to start to build a very comprehensive, genetic catalogue of severaldifferent types of cancer. it's been amazing

to see the results that have come out of that,and i'll touch on a few of those in a minute. but basically, for several different typesof cancer, like breast cancer, like lung cancer, we've been able to genetically dissect severaldifferent forms of those cancers. so, for example, in both lung cancer and breastcancer, there are several subtypes of disease, which really weren't even known very wellfive, 10 years ago. so now we can understand that some of these patients come to the clinicwith a completely different disease. and in the case of lung cancer patient, quite oftenlung cancer patients who have no smoking history are often expected to have a mutation in agene called egfr, epidermal growth factor receptor. and we can now test for that. andif they have a mutation in their egfr gene,

instead of giving them the very nasty typesof radiation and chemotherapy, we think -- we can give them a new class of drugs calledtyrosine kinase inhibitor, with very few side effects, and, in some patients, a very dramaticresponse. in some patients, the tumors just essentially melt away. but not all patientsrespond the same way, and so, again, we need to dig a little deeper and better understandwhat populations are going to be more susceptible to this particular subtype of cancer, whichtypes of patients are going to be more apt to respond to these new classes of drugs. the other project that i want to mention wascalled hmp, the human microbiome project. and here we've been able to use all of thislarge-scale sequencing technology to start

to catalogue the population of microbes thatare present in our gi tract, and the insides of our mouth and nasal passages, that arecarried around with us all the time. and there's pretty good hypothesis that as our healthstatus changes, if we become ill, or maybe as a cause of becoming ill, there's a changein these microbial populations. and so this is sort of one of the new areas over the lastfew years of genomics, and again, the technology, the methods, the software, the infrastructurethat were developed during the genome project have made these kind of things possible. well, to talk a little bit more about cancer,in 2008, using this next-generation sequencing technology that i mentioned a little earlier,my lab at washington university was able to

publish the first report of the genome sequencingof a cancer patient. and this was a -- the patient was a woman who lived in st. louisand had been diagnosed with acute myeloid leukemia. we actually sequenced two genomesfrom that patient: her normal genome, which came from some skin cells, which were takenat the time that she had her first bone barrow biopsy; and her tumor dna, which actuallycame from her bone marrow biopsy itself. so it blew our mind that we could actuallymake this work. we used this new technology. it actually took us a couple of years, notso much to do all the sequencing for that patient, but to develop the software toolsand to try understand how to deal with the enormous amount of data that we had generatedfrom this new technology. but it worked. in

her tumor genome, we found 10 mutations ingenes. and we now know, several years later, after sequencing a couple hundred more patientswith acute myeloid leukemia, we know the two genes that we saw mutations in that ultimatelycaused her disease. so just in our lab alone, since that firstcancer patient's genome, we've sequenced the genomes of over 1,500 cancer patients. this,again, is something that i would have had a time believing around 2000, 2003. but we'vedone it, again because of all these new developments. this number includes almost 1,000 genomesfrom pediatric cancer patients. and combined with the work that we've done as part of tcga,all of this work has really led to some amazing new insights into the biology of cancer.

so i could give you examples. one of the thingsthat we learned in our pediatric cancer genome project is we learned how to look at the genomeand begin to differentiate in children that have acute lymphocytic leukemia, the onesthat have a fairly standard subtype of that disease, and probably will do very well onstandard chemotherapy, from a smaller group of kids who have a very severe and aggressiveform of the disease called etpaol, who typically don't fare well unless they're treatment isaccelerated and very aggressive. so now we can start to think about how we pinpoint thosekids very early on in the diagnostic process. in an adult form of a brain cancer calledglioblastoma, one of things that we learned was that in some cases of this brain cancer,there are major mutations made to the mechanism

that we all have in ourselves to repair dnadamage. and if you didn't know about that in a particular patient ahead of time, youmight give them one of the most common chemotherapeutic drugs, but what we learned is that in thesepatients, that particular drug does more harm than good. in fact, just continues to ravagetheir genomes and cause even more mutation. so now we can think about how we check forthat before we start treatment in those glioblastoma patients. and finally, in the disease i first mentioned,adult acute myeloid leukemia, or aml, we now have what we call a genetic playbook thatsort of gives us some direction as to which patients probably will fare best with specifictreatments. and as we go forward and we get

a better association with the things thatwe found in the genome with what actually happens as those patients are treated, we'regoing to be able to pull this together and really use it in clinical practice. as you heard eric mention, a number of placesaround the u.s. have now started to have some very interesting and early experiences withactually trying to do genome sequencing of patients who are currently in the clinic,and may have a particularly bad form of cancer, or children that have a rare genetic disease.and we've had several cases of these. it's still very hard to do. as i said, the work-up-- the turnaround time is about a month for a cancer patient, and you'd like to turn theanswer back to the oncologist much more quickly

than that. but that's the best we can do now,although we see a great promise, and i'll talk about that in a moment. but for several of these cancer patients whosegenomes we've sequenced and be able to give that information back to the treating physicians,this has already saved lives. and perhaps, the most dramatic example of this, at leastin our own experience, was written up in the new york times by gina kolata last spring.a case of a physician who was actually part of our group who was diagnosed with a secondrelapse of acute lymphocytic leukemia, wasn't given a very good prognosis, but in sequencinghis genome, we found a mutation that could be targeted with a drug that was approvedfor kidney cancer, and that was effective.

within 12 days he was in a complete remission,and he just accepted a position to join the full-time faculty on the first of july ofthis year. so this has been exciting, and as eric touchedupon as well, cancer is really one of the first places that we've seen a big impactof the new genomics technology. one of the things about cancer that sort of puts it rightin the wheelhouse of modern genomics is that every patient essentially has a built-in controlexperiment, their normal genome. so we can sequence their normal genome, we can sequencetheir tumor genome, and we can compare and look in the tumor genome for the mutationsthat have arisen, and probably get clues to what, actually, not only caused their disease,but how we might be able to effectively kill

it. and we're now, in the large-scale centers,starting to target these more complex diseases, as eric mentioned, that include alzheimer'sdisease. the large-scale approach is really critical for these, because there is no built-incontrol for every patient. so what we need to do is to sequence lots of people, cases,and controls. cases are folks that have a particular disease, controls are the folksthat don't have the disease, but because each person is already 3 million base pairs differentwithin their genome, it's difficult to say, "is that particular variance related to thedisease or simply due to the fact that they're not the same individual whose genome we sequencedearlier?" so there's a real need for scale

here. we have to look at thousands of peoplewithin and without disease. so we've come a long way since those often-contentiousdiscussions about whether we should or shouldn't sequence the human genome, or whether we wouldlearn much from doing it. we've learned a tremendous amount. we did it. we did it well.and through the human genome project, and in the 10 years since, we've developed incrediblypowerful methods, technologies, software tools, as well as infrastructure and the abilityto manage huge amounts of data. we've learned an amazing amount of relevant human biology.we've also learned an amazing amount of -- and useful amount of information about animaland pathogen biology. we've learned valuable and applicable things about human disease,and we can start to see how we can start to

move those into the clinic. so for somebody like me, who's part of allthis since the beginning -- i can't even remember, 20, 25, 30 years ago, it's exciting and it'ssatisfying, but i have to say that i'm really most excited about the next 10 years. youknow, i told you earlier that we can now sequence and analyze an individual human genome forabout $10,000 in a couple of weeks. but since i sort of sit in this position where i'm ableto have a pretty clear understanding of the trajectory for the advancement of sequencingtechnology, i'm ultimately looking forward to having one of these before 2023. so you guys are sitting there saying, "whatthe hell is he talking about? he already has

one of those. it's an iphone." well, you'reonly partially right. this is a prototype iseq, and the really cool thing about theiseq is what you can do, is there's a really small attachment that hooks on here at thedata port. i can pipe that in a drop of blood. i wait a little while. data is generated.it hits the cloud, uses the knowledge base that we develop over the next 10 years; that'swhy it doesn't work now. and in a few minutes to a couple of hours i get an answer, right?so when the technology gets there, and it's fast, and it's inexpensive, and it's ableto take a benefit of all of the work from projects like tcga, hmp, 1,000 genomes, etcetera, every cancer patient, every child that comes into the hospital with a rare geneticdisease will have their genome sequenced,

effectively, accurately, and at a low cost. so, of course, to become reality, this willrequire more hard work, additional large-scale projects to build the underlying knowledgebase. and the question earlier about how we do this in the time of ever-decreasing budgetsis an excellent one. but we have to be optimistic, and we have to continue to work hard. andi think just having seen what's happened over the last 25 years with this technology -- itold you about my portable computer in 1990 -- i have no doubt that that vision can becomereality. so i'll stop there and be happy to answerquestions. [applause]

larry thompson:so, i was going to hold all questions for maybe the next three of you together, so maybe-- so don't go too far, though. what i'd like to do is ask greg lucier, the chairman andceo of life technology, to come up and give us a little sense of the state of the industry,and tell rick how long it'll be until you guys make the iseq phone and attachment thathe would like.