one of the world's greatest storiesis the doubling of human life expectancy over the last 150 years. that progresshas mostly been the result of the control of childhoodinfectious diseases, and in fact all of us in this roomare direct beneficiaries of that extraordinary progress. a child born today has an 85% chanceof reaching the age of 65. forty-two percent of those childrenwill celebrate their 85th birthday.
many of them will go on to bemore than a hundred years of age. but as is so often the case, extraordinary successreveals new challenges. and in this case we havea new and profound challenge for the continued extensionof human longevity. so, at the present time, we're faced withcontinued infectious disease problems but the primary killer of peoplehas shifted remarkably. we're now faced with an epidemicof age-related chronic diseases. heart disease, cancer, diabetes,alzheimer's disease.
these are the diseasesthat we're spending three quarters of our investmentin health care on. these are the challenges of the future,and let me suggest to you that new approaches are going to be requiredto overcome these challenges and define the next frontierin acceleration of human longevity. all great challenges need great heroes. you've heard from one of thoseheroes this week, peter diamandis, who's been a speakerwith you in a number of sessions, he is a co-founderof human longevity.
i want to introduce youto another hero, craig venter is arguablythe most important contributor in creating a unique opportunityfor all of us. and that is the beginningof understanding the software of life; our dna code. craig was the first personto fully sequence a bacteria twenty years ago. 15 ago, he was the first personto sequence the full human genome. since then, he's gone on to haveother extraordinary accomplishments,
including defining the microbiome. it turns out that we have more bacterialiving in and on our body than we actually have human cells,and all of that genetic material is interacting with our ownhuman cells' genetic material to impact our health. and most importantly, in 2010 craig did something that wasextraordinarily important; he created synthetic lifefor the first time in human history. he started at the computer,
wrote the software codein the form of a dna sequence, built the dna material, inserted it into a membrane and for the first time createda self-replicating form of life. let me give youa brief review on genomics. a lot of people talk aboutwhole-genome sequencing when they're actuallyjust talking about looking at a fragmentof the whole genome, and what's possible now is tolook at all 3.2 billion base pairs
across the human chromosome. this is an extraordinary accomplishment,but it's not complete, and, in fact, we're working very hardto look at what's called the diploid human genome; the full six point four billionbase pairs. this will be the full characterizationof the software of life. we think that this isa very important moment where a number of technologieshave converged to create this unique opportunity.
the first is that the costof dna sequencing has declined dramatically. when craig sequencedthe first full human genome in 2000, the cost to do a single genomewas one hundred million dollars. it took nine months, and it required the constructionof a computer, a custom computer, that at that time,was the third largest in the world. today, the cost of fullhuman genome sequencing is down to around $1300,
and we expect that priceto continue to decline. but that's not the only essentialprogress that's occurred in the last 15 years. the availability of reasonably pricedcloud-based compute power has also made the storageand manipulation of these vast data, for the first time, possible. machine learningis also a critical technology, and althoughthis is not a new technology, it has demonstratedthe potential to transform industries
where rich data are available. and lastly, as a result of the epidemicof age-related chronic diseases, we are in desperate need for movingfrom a volume-based healthcare system to one that delivers health value and continued extensionof healthy human longevity. let me give you a briefreminder of some biology that's relevant to definingthe software of life. so, this is the beginning of you, and more importantly than birthdays,this is the conception day.
and this is when yoursoftware is written. basically these two cells,these specialized cells, one from your father,one from your mother, that bring 23 chromosomesfrom your father in a very unique way, 23 chromosomesfrom your mother together, and this is the first timethat your software is written and starts to work on your behalf. that cell begins to divide and creates what's knownas a blastocyst.
and at this point,that software starts programming all the unique cells that are necessaryfor the function of your body. these include your muscle cells, these include the cellsthat you need to see, the cells that help you feel thingsand think - your neurons, the cells the color your eyes, the cells that help you absorbfood in your intestines. these cells and hundreds of other cellsare coded for in that dna software. and they come together in waysthat are unimaginably complex
to create one of these, ok? so, it is this software of lifethat it contains all the instructions that build, operateand reproduce our bodies. but here's the challenge, we don't know howto read that software. we've been lookingat small fragments but we actually don't knowhow to read the software, and that's really the nextprofound challenge that we think is goingto provide continued progress
in the single disease model for the practice of medicinethat we currently have, but most importantly it's alsogoing to fuel future progress in addressingthe root causes of aging, so that we'd deferthese age-related chronic diseases and we'd lead longer, healthier lives. so what can your genetic codepredict about you? we believe it's going to be...the answer to this question is gonna be quite surprising to people,
and that your genetic codecan predict much more about you than you currently believe. this is basically the game planthat were using that human longevity to hack this software of life. we're building integrated health recordsthat include the whole genome sequence and we're combining thatwith extensive phenotype data, this includes environmental data,clinical data, health history. we're doing thaton a cloud-based platform. we are currently oneof amazon web services'
the largest cloud customers, even though we've only done about 25,000whole genome sequences at this point. we expose those data to machinelearning at very large scale, and out of that, we believewe will gain profound insights that will drive a next generationof pharmaceuticals and medical interventions, but also the basis for a fundamentallynew practice of medicine, one that begins to shift fromwaiting until people get sick before we intervene,
to acting in a fundamentally proactive,preventative medicine format. how many of youhave used google translate? good, good. and i hope this translationis a good translation, maybe you can tell me. this is an important analogy. we are embarkingat human longevity on the translation of the language of biology,in the form of sequence data, into the languageof health and disease.
and we believe that it's so fundamentallyanalogous to language translation that we actually hired franczak,who built google translate at google, and he's currently leadingour machine learning efforts. the reason we were ableto compel him to take on this challengeis because we think all of us have the world's most importanttranslation problem right now, and that is understandingthis software of life. let me suggest something to youthat may be surprising, one of the thingsthat we've done early on
is enroll a studyof about a thousand people where we obtained three-dimensionalpictures of their face, to see if we could createphotograph-quality images of their face from the genome. and i'm happy to let you knowthat we can actually do this in a fairly extraordinary way, where we predict people's facefrom their genetic material. and this slide showsactual photographs of an individual,
and then the renderingsof that person that were developed, both skin tone and compositionof the face, directly from the genome. this may sound like a parlor trick;i assure you that it's not and it's a very good representationof the power of machine learning when appliedto whole genome sequence data, and it will have profound implicationsfor the way that we begin to connect medical imaging, for example,to whole genome sequencing data. we're also demonstrating our abilityto identify the fundamental causes of human aging through machine learning
by measuring the telomere length; these are the ends of the chromosome thatbegin to degrade during the aging process, that are going to allow us to comparephysiologic age to chronological age and begin to build direct medicalinterventions for aging with the notionthat this will prevent and defer the consequencesof age-related diseases. this is a surprising findingof our early work, both men and womenbegin to lose their sex chromosomes as they age,and at least in the case of men
this is associatedwith the risk of cancer, again, a new target to begin toaddress fundamental causes of aging. we can actually predict remarkable thingseven from a recording of your voice. not only your sex, but also yourweight, also your height, just from the sound of your voice. and here's our first target,our first medical target for use of these data. eight now in the united statesthe risk of death between the ages of 50 to 74 years of age is 30%in men, and it's 20% in women.
again, largely the result of theseage-related chronic diseases. we have builtfor the first time a platform that brings togetherat very large scale the use of whole genome sequencing,microbiome characterization, and metabolomics,with traditional healthcare data as well as advancedresearch-based imaging. and we're doing this to builda new way to practice medicine, one that's fundamentallydata-based, preventative, and proactive. these are some of the teststhat we're using,
we're relying a great deal on thecombination of whole-genome sequencing with magnetic resonance imagingor mri, mostly because the qualityof these images is now extraordinary and there's no risk of radiationand no use of contrast. this is our mri scannerat human longevity. these are some of the scansthat we're using. another important transitionis that we're moving to quantitate all of the datathat come from the mri so that it can be usedin a machine learning environment.
we can now lookat the precise volumes of twenty different neuroanatomicalstructures in the brain compare those to normaland follow those longitudinally. this is going to give usextraordinary diagnostic and risk prediction for diseasessuch as alzheimer's, where the hippocampal volume,for example, is highly predictive, particularly in the context of genomicsof diseases like alzheimer's. we can now quantitate muscleand fat throughout the body, and this is going to give usan extraordinary new understanding
of diseases like diabetes type 2, which is probably a common phenotype that results from manydifferent pathways to what we're measuring currentlyas a single disease. the traditional medical record is about three and a halfgigabytes right now. the platform that we've createdhas taken that up to about a 150 gb, for the first time beginning to definethe truly digital characterization of a human beingin ways that we think
are going to profoundly changethe practice of medicine. let me end my comment by addressing some of the policyimplications of this work, and they're numerous. but there is an increasing recognitionthat we must begin to shift our model from a single disease focus to one that addresses basic causesfor human aging. and that doing thatcan result in extraordinary progress, and in particular, the compressionof morbidity or frailty or sickness
at the end of life, as well asextending the length of life. and as you might expect,here on the left, the cost of doing thiscan be extraordinary, not because individual cost go up but because there areso many more older people as a result of these interventions. but the graph on the rightshows that those increased costs can be addressedthrough policy changes that relate to extended work life
as well as defermentof entitlement health care spending. so, let me end my commentsby suggesting two important things to all of you. we now have the scienceand technology to begin to define the next frontier in human longevity. the biggest barrierto doing this is complacency, ok? satisfaction with the status quo. and so the need,in fact a plead for me, is for leadershipto get us to this next frontier
and for all of usto collectively benefit the way we havefrom control of infectious diseases. leadership is required, and all of youhave an extraordinary opportunity to positively impactthe lives of billions of people that live in this region. so, let me welcome you to thenext frontier of health and medicine. thanks very much.
No comments:
Post a Comment