GOOD AFTERNOON.
MY NAME IS ERIC GREEN, DIRECTOR. 
RESEARCH INSTITUTE.
IT'S A PLEASURE TO INTRODUCE
TODAY'S SPEAKER.
DR. DAVID GOLDSTEIN.
DAVID'S BIOLOGY INCLUDES EARNING
THE Ph.D. FROM SANFORD,
UNIVERSITY IN 994 AND FOLLOWING
POSTDOCTORAL TRAINING, TOOK A
FACULTY TRAINING AT THE
UNIVERSITY COLLEGE LONDON AS
PROFESSOR OF GENETICS, A
POSITION HE HELD FROM 1999-2005.
THEN RECUTED TO DUKE UNIVERSITY
AS THE DIRECTOR FOR CENTER OF
GENOME VARIATION.
AND LAST YEAR HE BECAME THE
RICHARD AND PAT JOHNSON
DISTINGUISHED UNIVERSITY
PROFESSOR AT DUKE.
DAVID HAS MANY AWARDS, INCLUDING
ONE OF 7 ROYAL SOCIETY MERIT
AWARDS IN THE UNITED KINGDOM FOR
HIS WORK IN HUMAN POPULATION
GENETICS, MOST RECENTLY, ELECTED
CO-CHAIR FOR 2011 AND THE CHAIR
OF 2012.  GORDON RESEARCH
CONFERENCE ON HUMAN GENETICS AND
GENOMICS.
DAVID IS A VERY TALL WANTED,
INSIGHTFUL RESEARCHER.
WHICH IS WHY WE BROUGHT HIM
HERE.
HE IS WIDELY REGARDED AS A
LEADER, SPECIFICALLY ADVANCES IN
WHOLE GENOME SEQUENCE APPROACHES
ARE USED TO IDENTIFY THE GENETIC
BASES OF COMMON DISEASES AND
DRUG RESPONSES.
AS YOU KNOW, GENOME WIDE
ASSOCIATION STUDIES, KNOWN AT
GWAS, ALLOWS RESEARCHERS TO ZERO
IN ON REGIONS, BUT MANY
QUESTIONS REMAIN ABOUT WHETHER
STUDIES CONFINE ALL RISK
CONFERRING VARIANTS, AND HOW
THEY'RE GOING TO BE IDENTIFIED.
DAVID HAVE BEEN AT THE FOR AT
THE DEBATE AND INVESTIGATIONS,
AND HAS REALLY INVOLVED HIS OWN
WORK TO TRY TO DISSECT THIS
PROBLEM USING NEXT GENERATION
DNA SEQUENCING TECHNOLOGIES AND
NEW ANALYTICAL APPROACHES FOR
STUDYING IMPORTANT DISEASES.
GOING TO DESCRIBE SOME OF THIS
WORK TODAY IN HIS TALK ENTITLED
HUMAN GENETICS, THE NEXT
GENERATION, RARE VARIANTS AND
COMMON CURES.
DAVID.
[APPLAUSE]
>> WELL, THANK YOU VERY MUCH FOR
THAT INTRODUCTION.
AND IT REALLY IS A PRIVILEGE FOR
ME TO BE HERE TODAY.
I WILL BE TALKING ABOUT THE ROLE
OF BOTH RARE VARIANTS AND COMMON
VARIANTS IN TRAITS OF MEDICAL
INTEREST, AND MAKE COMMENTARY
ABOUT THE RESEARCH IN THE
FUTURE.
BUT BEFORE LAUNCHING INTO TOO
MUCH DETAIL ON THE RARE VARIANT
SIDE, I WANT TO KICKOFF WITH
SOME DISCUSSION OF SOME OF THE
WORK THAT WE HAVE BEEN INVOLVED
IN IN LOOKING AT THE ROLE OF
COMMON VARIANTS IN CONDITIONS OF
CLINICAL IMPORTANCE.
AND FOR ANYONE HERE TO SEE SORT
OF THE LATEST CHAPTER IN THE
GWAS WARS, I'M AFRAID YOU MIGHT
BE DISAPPOINTED.
I HAVE RATHER POSITIVE THINGS TO
SAY ABOUT GENOME WIDE
ASSOCIATION STUDIES AND RATHER
CONCERNING THINGS TO SAY ABOUT
SEQUENCING IN SOME REGARDS.
BUT LET'S JUST LAUNCH INTO IT.
SO FIRST, I'D LIKE TO TELL YOU
ABOUT STUDIES THAT WE HAVE BEEN
DOING TO LOOK AT THE GENETIC
CONTRIBUTIONS OF VARIABLE
RESPONSE TO TREATMENT FOR
CHRONIC INFECTION WITH HEPATITIS
C.
AND THE BACKGROUND TO THIS WORK
IS REALLY CLEAR, I THINK, TO
ANYBODY INTERESTED IN THE
GENETIC CONTROL OF CLINICALLY
IMPORTANT TRAITS.
IF YOU LOOK AT WHAT HAPPENS WHEN
STANDARD OF CARE TREATMENT IS
USED FOR CHRONIC HEPATITIS C
INFECTION, YOU SEE THAT THE
STANDARD OF CARE TREATMENT IS
REALLY VERY FAR FROM OPTIMAL.
FIRST, FUNDAMENTALLY IT WORKS
ABOUT HALF THE TIME.
ABOUT HALF OF THE TIME THE
PATIENTS GO THROUGH THE FULL
COURSE OF TREATMENT.
THE VIRUS IS GONE FOR GOOD.
ONLY WORKS ABOUT HALF THE TIME.
ON TOP OF THAT, THE FULL COURSE
OF TREATMENT IS QUITE ARDUOUS,
ALWAYS UNPLEASANT AND SOMETIMES
DANGEROUS.
FINALLY, THERE HAS BEEN A STRONG
INDICATOR THAT THERE IS A
GENETIC CONTRIBUTION TO THIS
VARIABLE OUTCOME TO STANDARD OF
CARE AND THAT IS THAT ADVISE OF
EUROPEAN ANCESTRY ARE KNOWN TO
BE MUCH MORE LIKELY TO BE CURED
BY STANDARD OF CARE THAN
INDIVIDUALS OF FINANCE ANCESTRY.
AS I COMMENT MORE IN A LITTLE
BIT IF YOU'RE INTERESTED THE
GENETICS OF A TRAIT LIKE THIS,
THEN IT IS THE CASE, EVEN FOR
SOMEBODY LIKE ME, THAT THE FIRST
THING YOU DO FOR SURE IS CARRY
OUT A GENOME WIDE ASSOCIATION
STUDY WHICH IS WHAT WE DID.
AND WHEN WE CARRIES OUT A GENOME
WIDE ASSOCIATION STEADY WE FOUND
A VARIANT, IL28B.
IT ENCODES INTERFERON LAMPDA, A
PROTEIN RELATED TO INTERFERON
ALPHA WHICH IS USED IN THE
STANDARD OF CARE TREATMENT.
WE FOUND A VARIANT THAT HAS A
DRAMATIC IMPACT ON WHAT HAPPENS
WHEN PATIENTS ARE TREATED.
SO HERE YOU SEE IN THE DIFFERENT
RACIAL AND ETHIC GROUPS,
INDIVIDUALS THAT ARE GROWN UP BY
GENOTYPE.
HERE IS THE GOOD RESPONSE YOU
SEE, AND THE POOR RESPONSE, T.T.
WHAT YOU SEE ON THE Y AXIS IS
THE PROPORTION OF INDIVIDUALS IN
EACH OF THOSE GENOTYPIC GROUPS
WHO ARE CURED.
THE LINGO IS SUSTAINED
VERLOGICAL RESPONSE.
WHAT YOU SEE HERE IS THAT FOR
INDIVIDUALS THAT HAVE THIS HOMO
ZYGOTE TYPE, MORE THAN 80% OF
THEM WILL BE CURED BY THIS
TREATMENT.
WHEREAS ADVICE WITH THIS ONLY
JUST OVER 30% OF THEM WILL BE
CURED.
WHEN YOU TALK TO CLINICIANS THAT
ACTUALLY SEE THESE PATIENTS,
MOST OF THEM AGREE THIS IS A
DIFFERENCE THEY WOULD WANT TO
TALK TO THEIR PATIENTS ABOUT.
IF OTHER THING THAT'S STRIKING
ABOUT THIS VARIANT, THE EXISTCY
OF ITS -- CONSISTENCY.
YOU SEE THE HOMO ZYGOTES DO MUCH
BETTER AND THOSE OF AFRICAN
ANCESTRY DO MUCH WORSE.
WHAT HAPPENS IF YOU LOOK AT
INDIVIDUALS WHO HAVE THE GOOD
RESPONSE, HOMO SIGHINGIS
GENOTYPE AND COMPARE THEM TO
INDIVIDUALS OF EUROPEAN ANCESTRY
THAT HAVE THE POOR RESPONSE?
THESE INDIVIDUALS DO ON AVERAGE
MUCH BETTER THAN THESE
INDIVIDUALS.
SO THIS REALLY STRONGLY
EMPHASIZES THE IMPORTANCE OF
ZEROING IN ON THE UNDERLYING
GENETIC CAUSE OF THESE AVERAGE
DIFFERENCES BETWEEN THE RACIAL
AND ETHNIC GROUPS.
MUCH MORE INFORMATIVE.
AND AS PEOPLE WHO DON'T KNOW THE
STORY MAY BE GUESSING, THE GOOD
RESPONSE IS THE DOMINANT ALLELE
IN INDIVIDUALS OF EUROPEAN
ANCESTRY, THE MINORITY ALLELE IN
INDIVIDUALS OF AFRICAN ANCESTRY.
THAT DIFFERENCE, IN FACT,
EXPLAINS MORE THAN HALF OF THE
DIFFERENCE IN AVERAGE RESPONSE
RATES TO AN INDIVIDUAL OF
AFRICAN AND EUROPEAN ANCESTRY.
SO THIS CAME OUT AFTER GENOME
WIDE ASSOCIATION STUDY, CLEARLY
THE RIGHT TOOL FOR THAT PROBLEM,
IDENTIFYING A VARIANT THAT HAS
AMBASSADOR BIGIOUS CLINICAL --
AMBIGUOUS CLINICAL RELEVANCE.
I WOULD LIKE TO EMPHASIZE THAT
THERE IS NOT ALGORITHMIC CLARITY
IN WHITE YOU OUGHT TO DO.
I THINK ALL YOU CAN REALLY SAY
IS THAT THIS INFORMATION GETS
INCORPORATED INTO A KIND OF
OVERALL -- IF YOU HAVE A POOR
RESPONSE GENOTYPE YOU YOU MIGHT
DELAY TREATMENT FOR NEWER
OPTIONS TO COME ONLINE.
I DON'T KNOW WHAT THAT WARNING
MEANS.
BUT IT MEANS I CAN'T ADVANCE.
I HAVE A DISCLOSURE, DUKE SO
DUKE AND INVESTIGATORS ARE
BENEFICIARIES OF A PATENT
ASSOCIATED WITH THE DISCOVERY.
I WANT TO MAKE THAT DISCLOSURE.
BUT I DO WANT TO POINT OUT THAT
IT'S NOT JUST MY PERCEPTION THAT
THERE IS CLINICAL RELEVANCE T
TEST IS OFFERED BY LAB CORE AND
THERE WERE 1,000 OR 1200 TESTS
BEING ORDERED A MONTH, SO THIS
IS ACTUALLY BEING USED IN THE
CLINIC.
I HOPE IN THE END WILL PROVE TO
HAVE BEEN BENEFICIAL IN THIS
SETTING.
WE CARRIED OUT ANOTHER GENOME
WIDE ASSOCIATION STUDY USING THE
SAME DATA.
THE DATASET CAME FROM A TRIAL
THAT WAS RUN BY SHARON CALLED
THE IDEAL TRIAL.
THIS WAS FOCUSED ON ONE OF THE
IMPORTANT ADVERSE EVENTS
ASSOCIATED WITH THIS TREATMENT,
THAT IS, ANEMIA.
A MAJORITY OF PATIENTS THAT
UNDERGO THIS THERAPY HAS
SIGNIFICANT DECLINE IN HEMO
GLOBIN LEVELS, DUE PRINCIPALLY
TO THE RIBAVIRIN COMPONENTS.
SO WE ASKED WHETHER THERE WERE
GENE VARIANTS THAT MEDIATED THE
EXTENT TO WHICH INDIVIDUALS
SUFFERED HEMOGLOBIN DECLINE.
WE SIMPLY TOOK THE DECLINE IN
HEMOGLOBIN OVER THE FIRST 4
WEEKS OF THERAPY, ASSIGNED THAT
AS A QUANTITATIVE SCORE TO
INDIVIDUALS AND CARRIED OUT A
GENOME WIDE ASSOCIATION STUDY.
WE FOUND, AGAIN, A GENOME WIDE
SIGNIFICANCE ASSOCIATION, THIS
TIME ON CHROMOSOME 20.
HERE THERE WERE A SET OF
VARIANTS THAT WENT UP TOWARD 10
TO THE NICHE 50.
AGAIN -- NEGATIVE 50.
NO QUESTION AT ALL THAT WE HAVE
A REAL ASSOCIATION OF WHO
SUFFERS ANEMIA.
HERE IS A COMMENT ABOUT
INTERPRETING GWAS.
ALL THE VARIANTS ON THE CHIP
THAT SHOWED ASSOCIATION WERE
SITTING IN THIS OPEN READING
FRAME.
HERE ARE THE VARIANTS THAT HAVE
THESE STRONG ASSOCIATIONS.
YOU DON'T SEE ASSOCIATIONS
NEARLY SO STRONG ANYWHERE ELSE.
WE KNOW FOR SURE THAT THE OPEN
READING FRAME IS INNOCENT OF ANY
EFFECT WHATSOEVER ON ANEMIA.
WHAT'S HAPPENING IS THAT THESE
VARIANTS ARE MARKERS WE WERE
ABLE TO WORK OUT FOR 2
FUNCTIONAL VARIANTS IN A
NEIGHBORING GENE, ITPA, ENCODES
[INDISCERNIBLE] AND THERE ARE 2
FUNCTIONAL VARIANTS IN THE GENE
THAT ARE LOW ACTIVITY VARIANTS.
IT TURNS OUT THAT THOSE 2
FUNCTIONAL LOW HANGIVITY
VARIANTS ARE ASSOCIATED WITH THE
MORE COMMON DISCOVERY VARIANTS.
THEY'RE NOT ON THE GENE CHIP. 
SO ALL THAT'S HAPPENING, THESE
MORE COMMON VARIANTS ARE
REFLECTING THE EFFECT OF THOSE
RARE FUNCTIONAL VARIANT.
WE WERE ABLE TO PROVE THAT
STATISTICALLY BY DOING THE
OBVIOUS THING, TAKING THOSE 2
FUNCTIONAL VARIANT, GENOTYPING
THEM, AND THEN ASKING ONCE YOU
ACCOUNT FOR THOSE 2 FUNCTIONAL
VARIANT, DOES THAT MAKE 2
INITIAL GENOME WIDE ASSOCIATION
SIGNAL GO AWAY COMPLETELY?
AND THE ANSWER IS YES.
IN EVERY RACIAL AND ETHNIC
GROUP.
SO STATISTICAL GENETIC
PERSPECTIVE, THOSE ARE SHOWN TO
BE RESPONSIBLE.
WE HAVE BEEN ABLE TO WORK OUT
THE BIOLOGY BEHIND HOW THIS IS
HAPPENING.
I'M NOT GOING TO GO INTO THAT AT
ALL, IT WOULD BE LENGTHY.
IT HAS TO DO WITH HOW LOW
ACTIVITY OF THIS ENZYME ALLOWS
THE MAINTENANCE OF ATP LEVEL IN
RED BLOOD CELLS GOOD SPITE THE
EFFECT OF RIBAVIRIN ON KNOCKING
DOWN ATP LEVELS.
BASICALLY, THE ITP BUILDUP IS
CAUSED BY A REDUCTION IN THIS
ENZYME, ALLOWS THE BIOSYNTHESIS
OF ATP TO CONTINUE DESPITE THE
PRESENCE OF RIBAVIRIN.
ANYONE INTERESTED, CAN SEE THE
WORK JUST COMING OUT BY A
POSTDOC, [INDISCERNIBLE], NAME
IS INDICATED THERE.
I THINK THIS IS A REALLY, TO ME,
ENCOURAGING EXAMPLE OF HOW YOU
CAN GET A POINTER OUT OF A
GENETIC ASSOCIATION STUDY THAT
REALLY LEADS YOU TO A BIOLOGICAL
CHARACTERIZATION OF WHAT'S
HAPPENING.
SO I THINK THAT THAT REALLY
MAKES CLEAR, AND I DON'T MEAN TO
SAY THE EXAMPLES ONLY COME FROM
OUR WORK.  THERE ARE MANY
EXAMPLES WHERE GENOME WIDE
ASSOCIATION STUDIES HAVE FOUND
CLINICAL USEFUL FINDINGS AND
ALSO INSIGHTS TO BIOLOGY.
I THINK IT'S FAIR TO SAY, AND AS
I HAVE ARKED BEFORE, THERE WERE
MANY TREATS WHERE THE VARIANTS
RESPONSIBLE ARE NOT SUITABLE TO
BE DEWE CANNOTED BY GENOME WIDE
ASSOCIATION, BECAUSE THEY'RE TOO
RARE TO BE EASILY DETECTED.
NOW, FOR SOME TRAITS, AND I
THINK THEY HAVE A PATTERN TO
THEM, THE TRAITS THAT ARE MORE
SUBJECT TO COMMON VARIATION, FOR
SOME TRAITS, GENOME WIDE
ASSOCIATION APPROACHES DESCRIBE
A LOT OF WHAT'S HAPPENING.
FOR A LOT OF TRAITS, THEY WON'T.
THAT HAS LED TO A TREMENDOUS
AMOUNT OF INTEREST IN DEVELOPING
THE BASIC MACHINERY THAT WILL
ENABLE US TO USE SEQUENCING
STRATEGIES TO IDENTIFY RARER
VARIANTS TO CONTINUE TO TRAITS
OF INTEREST.
I WOULD LIKE TO SPEND THE BULK
OF THE TALK, TALKING ABOUT HOW
WE'RE APPROACHING THAT RIGHT NOW
AND WHERE I SEE THAT WORK GOING
IN THE COMING YEARS.
THE WORK THAT I'M GOING TO TALK
ABOUT DEPENDS UPON THE OUTPUT OF
A SMALL GENOME FACILITY COMPARED
TO GENOME CENTERS THAT WE
OPERATE AT DUKE, THE GENOME
FACILITY IS RUN BY KEVIN,
CURRENTLY CONSISTS OF 11 GENOME
ANALYZER 2s WHICH ARE BEING
PHASED OUT AND 6 [INDISCERNIBLE]
JUST BEING BROUGHT INTO
OPERATION.
SO I WOULD SAY THAT'S RELATIVELY
LARGE COMPARED TO PLACES THAT
AREN'T GENOME CENTERS.
I WOULD SAY MY INSTITUTION HAS
BEEN VERY SURPRISED AT THE COST
OF TRYING TO STORE THE DATA THAT
WERE GENERATED.
THERE ARE ONGOING DISCUSSIONS
ABOUT HOW WE DO THAT.
THE DATA COME OUT OF THIS
FACILITY AND WE HAVE PUT A LOT
OF ENERGY, SPECIFICALLY, INTO 2
BIOINFORMATIONIC AREAS IN ORDER
TO WORK WITH THESE DATA.
ONE THING THAT [INDISCERNIBLE]
HAS DONE OVER THE PAST FEW YEARS
IS TRIED TO DEVELOP WHAT YOU
MIGHT CONSIDER A APPLE MACINTOSH
TYPE APPROACH TO LOOKING AT THE
VARIANTS THAT EMERGE FROM
SEQUENCING DATA.
SO I THINK THERE ARE A LOT OF
WAYS THAT PEOPLE ARE THINKING
ABOUT ANNOTATING THE VARIANTS
THAT ARE PRETTY CONSISTENT.
WE HAVE FOUND IT'S REALLY
IMPORTANT FOR USERS TO BE ABLE
TO INTERACT WITH THEIR DATA.
TO BE ABLE TO LOOK UP THINGS IN
GENES OF INTEREST, PATHWAYS OF
INTEREST AND TO INTERACT WITH A
REASONABLE TIMEFRAME.
[INDISCERNIBLE] HAS SET UP WHAT
I THINK IS AN EXTREMELY EASY TO
USE USER INTERFACE THAT HAS A
KIND OF BROWSER LIKE STRUCTURE
TO IT FOR LOOKING AT THE
VARIANTS THAT ARE IDENTIFIED IN
EACH OF THE GENOMES THAT YOU'VE
SEQUENCED AND MAKING COMPARISONS
BETWEEN TYPES OF VARIANTS.
IF ANYONE IS INTERESTED IT'S
AVAILABLE AT THE WEBSITE THAT'S
INDICATED.
THE OTHER AREA WHERE WE HAVE
BEEN PUSHING DEVELOPMENT A
LITTLE BIT AS OPPOSED TO
DEPLOYING THE METHODS THAT
OTHERS HAVE DEVELOPED IS IN
CALLING COPY NUMBER VARIANTS
USING THE DATA EMERGING FROM THE
NEXT GENERATION SEQUENCING.
AND WHAT WE HAVE FOUND AS ALSO
USING A QUITE DIFFERENT
APPROACH, IS THAT WHOLE GENOME
SEQUENCE DATA DOES AFFORD A
TRULY STUNNING RESOLUTION FOR
DEFINING THE COPY NUMBER
VARIANTS LANDSCAPE.
IF YOU'RE TALKING ABOUT DUPLIES
OR DELETIONS OF GREATER THAN,
SAY, 2 KILLA BASES IN SIZE,
THOSE CAN BE HARD DOWN TO THAT
KIND OF SIZE TO PICK UP
ALTERNATIVE METHODS BUT TO READ
THAT DATA GIVES YOU REALLY GOOD
RESOLUTION ON THOSE COPY NUMBER
VARIANTS.
THE WAY WE HAVE SET THIS UP, WE
MAKE USE OF THE ALIGNMENTS THAT
COME FROM BWA THAT YOU USE FOR
OTHER VARIANT CALLING.
[INDISCERNIBLE] APPROACHES TO
USE AN ALTERNATIVE SET OF
ALIGNMENTS.
WE HAVE AN APPROACH THAT USES
THE BWA ALIGNMENTS THAT AFFORD
VERY GOOD RESOLUTION,
INCORPORATES A NOVELTY WHICH IS
IT MAKES USE OF INFORMATION
ABOUT WHERE HETEROZYGOTE SITES
ARE CALLED IN THE GENOME,
PRECLUDES THOSE BEING CALLED
DELETED, MAKES IT PERFORM WE
THINK BETTER.
BUT IT'S QUITE CLEAR THAT YOU
CAN GET VERY GOOD RESOLUTION
ABOUT COPY NUMBER VARIATION IN
THE GENOMES AND THAT REALLY, I
THINK, LONG TERM IS GOING TO
CONSTITUTE A VERY STRONG
MOTIVATION FOR MOST OF THE
SEQUENCING BEING DONE BY WHOLE
GENOME AS OPPOSED TO WHOLE
EXOME.
IT'S NOT GOING TO BE AS HIGH
RESOLUTION, SO AS THE
SEQUENCING, WE'LL BE TURNING
THAT DIRECTION.
THE FIRST EFFORT WE MADE IN
TRYING TO SEE WHAT KINDS OF
THINGS COME OUT OF LARGE SCALE
SEQUENCING STUDIES WAS REPORTED
RELATIVELY RECENTLY IN PLUS
GENETICS.
I WANT TO EMPHASIZE ONE THING
FROM THIS STUDY.
THAT IS JUST THE SHEER MAGNITUDE
OF GENETIC VARIATION THAT
EMERGES FROM THESE STUDIES.
I'VE GOT 2 PLOTS HERE THAT
ILLUSTRATE THAT.
WHAT I'M SHOWING HERE IS THE
NUMBER OF GENOMES WE CONSIDER.
ON THE Y AXIS, THE NUMBER OF
NOVEL, SINGLE NEW CLEO TIED
VARIANTS.
FOR THE FIRST GENOME, IT'S A
VARIANCE THAT ISN'T IN
DATABASES.
THIS IS AT THE TIME WE DID THE
ANALYSIS, IT WOULD BE DIFFERENT
NOW.
A VARIANT THAT'S NOT IN ANY OF
THE DATABASES.
THE NEXT ONE IS A VARIANT THAT
IS IN THIS NEXT GENOME, NOT IN
ANY OF THE DATABASES, NOT IN THE
PREVIOUS FELLOW THAT YOU
SEQUENCED, AND SO ON.
THE REAL STRIKING THING IS THAT
AFTER YOU'VE ALREADY SEQUENCED
20 GENOMES, EVERY TIME YOU
SEQUENCE A GENOME, YOU'RE SEEING
MORE THAN 100,000 SINGLE
NUCLEOTIDE VARIANTS THAT YOU
HAVEN'T SEEN BEFORE.
SOME OF THESE ARE NOT REAL BUT
THE VAST MAJORITY ARE, SO IT'S A
TREMENDOUS AMOUNT OF VARIATION
EVERY TIME.
BUT MORE THAN THAT, THERE IS A
TREMENDOUS AMOUNT OF FUNCTIONAL
VARIATION.
WE TOOK ALL THE VARIANTS THAT
WERE PREDICTED TO BE PROTEIN
TRUNCATING AND HOMO HOMOZYGOUS,
AND WE ASKED HOW MANY NEW GENES
DO YOU LEARN ABOUT THAT APPEAR
TO -- SO YOU PREDICT THE PROTEIN
IS NOT MADE.
WHAT YOU SEE IS THAT EVEN AFTER
YOU'VE DONE 20 GENOMES, YOU
STILL GET 3, 4, 5 SUCH GENES
IDENTIFIED.
I WANT TO EMPHASIZE THAT WE
FOLLOWED UP SOME OF THESE.
A PROPOSE OF THESE ARE NOT REAL.
THE PROPOSE OF THESE VARIANTS IS
HIGHER THAN A RANDOM VARIANT.
IT'S UNLIKELY, EVEN MORE
ENRICHED.
THIS SHOWS THAT THERE IS A
TREMENDOUS AMOUNT OF FUNCTIONAL
VARIATION IN EVERY GENOME THAT
YOU SEQUENCE.
THAT IS GOING TO BE ONE OF OUR
CHALLENGES IN TRYING TO IDENTIFY
PATHOGENIC PARIENTS.
YOU CAN'T JUST LOOK AND SAY THIS
DOES SOMETHING QUITE TRACK, THAT
IS NOT GOING -- DRAMATIC.
THAT IS NOT GOING TO WORK.
WILL SEQUENCING WORK?
I WANT TO GIVE A COUPLE OF HE
CANS THAT I HOPE WILL BE ON THE
ENCOURAGING SIDE, MAKE SOME
CAUTIONARY COMMENTS AND BACK TO
I HOPE BEING ENCOURAGING.
SO OUR OWN MOST RECENT STUDY
EVALUATING WHETHER SEQUENCING
CAN WORK IN THE SENSE OF FINDING
PATHOGENIC MUTATIONS THAT YOU
HAVE NO INFORMATION ABOUT
BEFORE, COMES FROM A STUDY WE
DID OF MICROCEPHALY.
SO THIS IS WORK THAT WAS DONE IN
COLLABORATION WITH COLLEAGUES
FROM ISRAEL, [INDISCERNIBLE] IN
PARTICULAR, RUN BY ELIZABETH IN
MY GROUP.
AND THE START OF THIS WAS THAT
IN TELEEVEN, THEY HAD 2 FAMILIES
THAT BOTH HAD WITH A
MICROCEPHALY SYNDROME THAT THE
CLINICIAN FELT WAS ANN NOT
DESCRIBED BEFORE AND THE SAME.
SO NATURALLY WE WANTED TO KNOW
IN THAT SITUATION CAN YOU
SEQUENCE AND FIND THE MUTATION
THAT'S RESPONSIBLE FOR THE
PRESUMABLY SAME MICROCEPHALY
SYNDROME?
HERE IS THE FIRST FAMILY.
SO YOU CAN SEE THAT IT LOOKS
LIKE IT'S AUTOSOMAL RECESSIVE.
THERE ARE 2 EFFECTED THAT WE
SEQUENCED.
THAT'S A FETUS THAT WAS ABORTED
ON THE PRESUMPTION THAT THE
FETUS WOULD HAVE THE
MICROCEPHALY.
WHEN WE SEQUENCED BY WHOLE EXOME
SEQUENCING, THESE 2 INDIVIDUALS,
AND THEN LOOKED AT ALL THE
VARIANTS WE FOUND AND COMPARED
THEM TO CONTROL GENOMES, WE ONLY
FOUND 3 VARIANTS THAT WERE HOMO
SIGHING THE, FUNCTIONAL, AND
HOMO ZYGOTE IN BOTH OF THE
EFFECTIVE, ONLY ONLY.
SO IN THAT FAMILY, WE HAD IT
DOWN FROM THE EXOME SEQUENCING
TO 3 CANDIDATES.
THEN WE LOOKED IN THE OTHER
FAMILY.
AND SURE ENOUGH, ONE OF THOSE 3
VARIANTS IS HOMO GIG THE.
WE, THEN, WENT ON TO TYPE
ANOTHER 1,000 INDIVIDUALS FOR
THAT VARIANT AND DID NOT SEE THE
VARIANCE.
WHAT I FIND PARTICULARLY
ENCOURAGING ABOUT THIS EXAMPLE
IS THAT WE HAD IT DOWN TO 3
CANDIDATES IN THE FIRST FAMILY.
AND I THINK THIS MAKES A REAL
STRONG CASE THAT EVEN IN
ISOLATED GENETIC CONDITIONS, YOU
HAVE SOME HOPE OF FINDING THE
GENETIC CAUSE.
I THINK THAT'S A REALLY
IMPORTANT OBSERVATION.
NOW, OF COURSE, WE HAD TO HAVE
THIS FAMILY TO PROVE IT.
BUT IT IS DOWN TO 3 CANDIDATES
IN THIS FIRST FAMILY.
AND BEING ABLE TO DO THAT REALLY
DID DEPEND CRITICALLY ON THE
FILTER OF ABSENCE IN CONTROL
GENOMES.
AND THAT'S ONE OF THE REASONS
THAT THIS FRAMEWORK CAN WORK FOR
HIGH PEN TRANSPLANT MUTATIONS,
WHEN -- OPENETRATIONS WHEN YOU
CAN FILTER IT THAT WAY IT'S
EASIER TO FIND WHAT'S
PATHOGENIC.
THE GENE THAT HAS THE MUTATION
HAS NOT BEEN IMPLICATED IN ANY
MICROCEPHALY BEFORE.
IT'S A NEW INSIGHT.
IT'S IN THE AS PAR RAGING
SYNTHETASE GENE.
THE. -- OUR FIRST THOUGHT THE
MUTATION WOULD EFFECT
DESAMEMATIC ACTIVITY.
WE SET UP AN EXPERIMENTAL ASSAY
TO TEST THAT.
BUT THE POSTDOC FOUND QUICKLY
THAT THE MUTATION CAUSES A
PROFOUND INSTABILITY OF THE
PROTEIN AND SO WE COULDN'T GET
ENOUGH PROTEIN TO BE ABLE TO
CARRY OUT THE ESSAY, SO WE THINK
WHAT CLEARLY IS HAPPENING IS YOU
JUST DON'T MAKE VERY MUCH OF
THIS PROTEIN WHEN YOU CARRY THE
MUTATION.
THAT'S WHAT CAUSES THE
MICROCEPHALY SYNDROME.
NOW IT WILL BE VERY IMPORTANT TO
LOOK AND SEE WHAT THE EFFECT OF
THIS IS IN A MOUSE KNOCKOUT.
I HAVE WE ARE PERSPECTIVE ON
ASKING THE QUESTION OF WHETHER
SEQUENCING OUGHT TO WORK.
AND THIS IS NOW NOT A REAL
STUDY, BUT THIS IS AN
EXPERIMENT.
LET'S RETURN FOR A MINUTE TO --
LET'S RETURN FOR A MINUTE TO THE
ANEMIA EXAMPLE, WHERE WE LOOKED
AT THE QUANTITATIVE CHANGE IN
HEMOGLOBIN FOR EACH.
SUBJECTS.
EACH HAVE A SCORE OF THE CHANGE
IN HEMOGLOBIN.
WE CAN ASK WOULD WE HAVE BEEN
ABLE TO EASILY FIND THOSE
MUTATIONS IF THEY HAD IS NOT
SHOWN UP IN THE J WAS.
WE GOT VERY LUCKY.
THEY'RE NOT ON THE CHIPS.
WE ONLY FOUND THE MARKER BECAUSE
2 VARIANTS HAPPENED TO ASSOCIATE
JUST RIGHT.
MAYBE WE HADN'T GOTTEN 0 SO
LUCKY, WE COULD HAVE TRIED BY
SEQUENCING.
WOULD WE FIND IT?
LET'S SET A SIGNIFICANCE
THRESHOLD.
LET'S AT THAT WE'RE GOING TO
TAKE ALL THE NON SYNONYMOUS,
TESTING FOR ALL THE VARIANTS.
WE HAVE TO CORRECT FOR THAT.
CALL THE SIGNIFICANCE LEVEL 10
TO THE NEGATIVE 6, 10 TO THE
NEGATIVE 7.
INSTEAD OF SEQUENCING EVERYBODY
YOU'RE GOING TO SEQUENCE THE
ENDS OF THE DISTRIBUTION.
HERE IS HEMOGLOBIN CHANGE FOR
ALL THE SUBJECTS IN THE STUDY
THAT WE DID, THE GWAS STUDY.
LET'S SAY YOU TAKE THE UPPER AND
LOWER 5 PERCENTILE.
THOSE WITH THE MOST CHANGE AND
THOSE WITH THE LEAST CHANGE.
SEQUENCE THESE, SEQUENCE THESE.
THE COUGHS OF DOING THAT -- COST
OF DOING THAT IS JUST OVER
$100,000.
SO THAT IS NOT AN EXPENSIVE
EXPERIMENT.
NOW, THE LABOR TO GENERATE THE
VARIETIANTS, YOU HAVE TO HAVE
PEOPLE THAT CAN DO SOMETHING
WITH THE VARIANTS.
BUT THE DIRECT REAGENT COST.
WHAT HAPPENS WHEN YOU DO THAT?
THE MOST IMPORTANT OF THOSE
FUNCTIONAL VARIANTS THAT I
DESCRIBED HAVE AP VALUE OF 10 TO
THE NEGATIVE 13.
IN A SIMPLE TASTE CONTROL
COMPARISON BETWEEN THESE AND
THESE.
VERY EASILY FINDABLE BY
SEQUENCING, NO QUESTION ABOUT
IT.
WHOLE EXOME SEQUENCING, HERE IS
AN INTERESTING COMMENT.
HAD WE NOT DONE A GENOME WIDE
ASSOCIATION STUDY AND JUMPED TO
HIT IT WITH WITHQUENCING, THIS
WOULD HAVE BEEN A COMPLEX
STRAIGHT THAT YOU CAN EASILY
FIND THE GENETIC CAUSE FOR BY
SEQUENCING.
THAT REALLY IS AN ENCOURAGING
EXAMPLE.
MAKES YOU WISH IT WASN'T
DISCOVERRABLE BY GWAS.
BUT IT IS CERTAINLY POSSIBLE
THAT SOME TRAITS WILL BEAMINABLE
TO DISCOVERY.
AS FAR AS I KNOW THERE ISN'T AN
EXAMPLE OUT THERE.
OKAY.
SO I'M GOING TO TURN TO THE
BASIC APPROACH WE'RE USING FOR
COMPLEX TRAIT DISCOVERY.
I WANT TO GIVE A LITTLE BIT OF
BACKGROUND TO OURING AND LAUNCH
INTO THE STUDIES WE'RE DOING.
THE FIRST THING TO REALLY
EMPHASIZE, WHICH IS A BIT
DIFFERENT FROM GENOME WIDE
ASSOCIATION, IS JUST HOW
IMPORTANT IT IS TO BE CAREFUL
WITH WHO YOU DECIDE TO SEQUENCE.
IT REMAINS EXPENSIVE PER
INDIVIDUAL COMPARED TO GENOME
WIDE ASSOCIATION.
YOU'RE GOING TO BE MAKING FINE
DISTINCTIONS, YOU WANT TO MAKE
REALLY SURE OF PHENOTYPES IN THE
INDIVIDUALS YOU SEQUENCE.
HERE ARE THE KINDS OF DESIGNS
WE'RE THINKING ABOUT.
ONE OF THE DESIGNS IS TO GET
FAMILIES, SEEK OUT A NUMBER FROM
THE FAMILIES, AND DEPENDING ON
THE INFORMATION YOU HAVE IN THE
FAMILY A GOOD DESIGN MIGHT BE
FOR EXAMPLE TAKINGINGTANTLY
RELATED COINFECTED, OR MAYBE
JUST ONE INDIVIDUAL.
BUT SEQUENCING ALLOWS YOU TO
LEVERAGE COSEGREGATION DATA IN
THOSE FAMILIES IN THE ANALYSIS.
THE OTHER DESIGN THAT WE ARE
THINKING ABOUT IS TAKING
EXTREMES OF SOME KIND FROM THE
POPULATION SO THAT THAT ENRICHES
FOR CAUSAL VARIANTS, AS YOU CAN
SEE HERE.
AND THEN SEQUENCE THOSE
EXTREMES, LOOK FOR THAT
ENRICHMENT OF CAUSAL VARIANTS,
BUT TAKE THOSE VARIANTS BACK OUT
INTO THE LARGER POPULATION FOR
CONFIRMATION OF THEIR
ASSOCIATION WITH THE TRAIT.
THOSE ARE THE 2 DESIGNS WE'RE
TRYING TO MAKE THE MOST USE OF.
BEFORE I MOVE INTO THAT, LET ME
MAKE SOME GENERAL COMMENTS ABOUT
WHAT I SEE AS THE RELATIONSHIP
BETWEEN GENOME WIDE ASSOCIATION
GOING FORWARD AND SEQUENCING.
SO THE FIRST COMMENT IS IF YOU
HAVE A NEW TRAIT, I DO TAKE THE
VIEW THAT IT'S VERY SENSIBLE TO
CARRY OUT A GENOME WIDE
ASSOCIATION STUDY FIRST.
FOR SOME TRAITS IT CAN BE A
QUICKER ROOT TO UNCOVERING THE
GENETIC CONTROL.
AND I SHOULD JUST CLARIFY FOR
THE RECORD, FOR ANY TRAIT THAT
YOU'RE INTERESTED IN, YOU WOULD
LIKE THERE TO BE GWAS DATA
ALREADY.
IT WAS VERY VALUABLE TO GO
THROUGH THE GENERATE THE DATA.
IT'S A SEPARATE QUESTION OF HOW
MUCH GWAS DATA YOU NEED.
BUT WHETHER YOU DO A CREDIBLE OR
GIVEN TRAIT, I THINK THAT'S NOT
SOMETHING THAT REASONABLE PEOPLE
CAN DISAGREE ABOUT.
BUT FOR MOST CONDITIONS WHERE
THERE ALREADY BEEN A GWAS DONE
OF REASONABLE SIZE, I DON'T
THINK THAT GWAS PRACTICED ALONE
IS AN IMPORTANT DRIVER OF
DISCOVERY ANYMORE.
THE REASON IS THAT IF YOU HAVE A
CONDITION WHERE YOU HAVE LOTS OF
SIGNALS AND YOU'RE HAVING
TROUBLE TRACKING DOWN THE
CAUSES, ADDING MORE SIGNALS, I
THINK, IS NOT REALLY WHERE WE
WANT TO CONCENTRATION OUR
ATTENTION.
WE WANT TO CONCENTRATION MORE ON
FINDING THE CALLS OF THE
SIGNALS.
AND SO THAT MEANS THAT I THINK
WE'LL BE INTERPRETING THE GWAS
SIGNALS ALONG WITH THE SEQUENCE
DATA TO TRY TO FIND THE CAUSES
OF THE GWAS SIGNALS.
I'D ALSO LIKE TO EMPHASIZE THAT
THAT COMBINED USE OF GWAS AND
SEQUENCE DATA IS PARTICULARLY
IMPORTANT, GIVEN THAT GWAS IS A
VERY ACCURATE AND WELL
UNDERSTOOD EXPERIMENT AND
SEQUENCING IS NOT.
SOMEONE HAS TO BE ABSOLUTELY
CLEAR ABOUT THAT.
WE'RE REALLY GOOD AT DOING
GENOME WIDE ASSOCIATION STUDIES,
REALLY, REALLY GOOD AT IT.
SO WE HAVE THE GENOME TYPING
CRASSRY RIGHT THERE WHEN IT'S
DONE CAREFULLY.
WE'VE GOT IT RIGHT WHERE WE WANT
IT TO BE.
WE KNOW WHAT THE ARTIFACTS ARE.
WE KNOW HOW TO CONTROL FOR
POPULATION STRATIFICATION.
WE KNOW HOW TO CORRECT FOR THE
NUMBER OF HEALTH DISPARITIES
WE'RE ALLEGATIONS OF ABUSE --
HIGH BOTH SEES WE'RE TESTING.
THAT'S WHY WE WANT TO THINK
ABOUT DOING THEM IN TAN --
TANDEM.
AND WITHOUT THINKING, LOADING UP
ON THE COMMON ONES WITHOUT
TAKING A VIEW ON THE PROPORTION,
I THINK IT'S CLEAR THAT SOME
PROPORTION OF THE SIGNAL MUST BE
DUE TO THIS EFFECT AND IF SO,
THEY PROVIDE A GOOD POINTER,
LOOKING FOR SOME OF THE IMPACT
RARER VARIANTS.
WINE I SEE THE VOLOF USING THESE
2 TYPES OF APPROACHES TOGETHER.
I'D ALSO LIKE TO EMPHASIZE A
LITTLE BIT FROM OUR OWN
EXPERIENCE, JUST HOW -- I GUESS
DANGEROUS IS THE RIGHT WORD,
SEQUENCING IS FROM THE
PERSPECTIVE OF FALSE POSITIVES.
AND I REALLY WANT TO MAKE THAT
POINT IN THAT I THINK ONE OF THE
UNAPPRECIATED CONTRIBUTIONS OF
GENOME WIDE ASSOCIATION WAS TO
REALLY CLEAN UP THE LITERATURE
AND REDUCE ATTENTION TO A WHOLE
BUNCH OF FALSE POSITIVES THAT
REALLY DID PLAYING A LOT OF
FIELDS.
THERE WAS SOME DISCIPLINES WHERE
THE VAST MAJORITY OF GENETIC
DISCOVERIES ARE SIMPLY UNTRUE.
WHEN YOU LOOKED AT THE GENOME
WIDE ASSOCIATION PROJECT
CAREFULLY, THAT BECAME CLEAR.
THAT REPRESENTS A TREMENDOUS
WAIST OF EFFORT.
AND.
>> WAS REALLY DID, I THINK, HELP
TO CLARIFY WHAT WAS REAL
ASSOCIATION AND WHAT WASN'T.
IT REPRESENTS A BETTER
ASSOCIATION STUDY THAN CANDIDATE
STUDIES IN EVERY SENSE.
AND SO I THINK IT'S REALLY
CRITICAL WE DON'T LET SEQUENCING
GO DOWN THAT OLD OLD BEING
FLOODED WITH FALSE POSITIVES.
BUT IT WOULD BE VERY EASY TO DO.
HERE IS AN EXAMPLE WHERE WE'RE
SEQUENCING SCHIZOPHRENIA GENOMES
AND WE PUMP INTO A MUTATION THAT
INTRODUCES A PREMATURE STOP IN A
DOPAMINE RECEPTOR.
WE HAVE A FUNCTIONAL VARIANT IN
A GENE THAT IS OF CLEAR
RELEVANCE TO SCHIZOPHRENIA BY
REALLY JUST ABOUT ANYBODY'S
DEFINITION.
AND COULD WE HAVE MANAGED TO
PUBLISH THAT IF WE WANTED TO?
YOU BET YOU COULD PUBLISH THAT.
BUT YOU ALSO KNOW THESE THINGS
HAPPEN.
SO WE DECIDE TO SEQUENCE IT, AND
THEN WE SHOW WE CAN ACTUALLY SEE
THE VARIANT IN SOME MORE CASE
BUT WE ALSO SAY IT CROPPING UP
IN CONTROLS.
BUT MORE THAN THAT, WE ALSO WHEN
WE LOOK AT THE NEXT GENERATION
SEQUENCE DATA AND THE SANGER
SEQUENCE DATA, AT THE END OF THE
DAY WE CAN'T BE ABSOLUTELY 100%
POSITIVE THAT THERE IS A
VARIANTS THERE.
SOMETIMES IT CAN BE DIFFICULT TO
DETERMINE.
SO AT EVERY STAGE YOU'VE GOT
PROBLEM.
GENOTYPING ACCURACY AND STORY
TELLING.
SO I THINK IT'S ALL IMPORTANT WE
ESTABLISH A REALLY POWERFUL
FRAMEWORK TO ESTABLISH WHAT IS
CONVINCING ASSOCIATION AND ALSO
TO PREVENT TOO MUCH OF THE FALSE
POSITIVES COMING OUT AGAIN IN
THE LITERATURE.
AND SO I GUESS WHAT I'M REALLY
SAYING, I ACTUALLY PERCEIVE THE
POSSIBILITY, MAYBE THE
LIKELIHOOD THAT THERE IS GOING
TO BE A LITTLE BIT OF A WILD
WEST MENTALITY OF PEOPLE REALLY
FORCING ANSWERS OUT.  SEQUENCE
DATA ANY OLD WAY THEY CAN,
MAKING UP A LOT OF STORIES AND
LIKE HAPPENED FOR J WAS -- GWAS,
I WOULD LIKE TO SEE THE
SEQUENCING COMMUNITY COME
TOGETHER ON HOW TO DO THIS
RIGHT.  THIS IS OUR THINKING HOW
TO DO IT RIGHT.
YOU YOU FIRST COMMENT, YOU DO
NOT DECLARE A DISCOVERY BEFORE
YOU SEPARATELY GENOME TYPE THE
VARIANTS, SHOW IT'S ACCURATE,
AND CONVINCING WITH A SEPARATE
TECHNOLOGY, SO THE SEQUENCING IS
DISCOVERY.
YOU TAKE THE VARIANTS AND PROVE
THAT YOU'VE ACCURATELY GENOTYPED
THEM.
MAYBE IN A FUTURE DATE, WE CAN
SKIP THAT.
I THINK THAT'S A WAYS DOWN THE
ROAD.
HERE ARE THE OPTIONS FOR FOLLOW
UP RIGHT NOW.
ACTUALLY PRETTY GOOD.
IF YOU WANT TO DO SOME FEW
VARIANTS THAT YOU'RE INTERESTED
IN, I'LL TELL YOU ABOUT SOME WE
HAVE DONE.
YOU CAN [INDISCERNIBLE], THAT'S
A SEPARATE TECHNOLOGY THAT IS
OFTEN COMPLETELY CONVINCING,
SOMETIMES NOT.
>> IF YOU WANT TO DO SOME NUMBER
OF HUNDREDS OF VARIANTS, THEN
TECHNOLOGY IS BEAK EXPRESSED.
IF YOU WANT TO DO THOUSANDS,
APPROPRIATE TECHNOLOGY IS
[INDISCERNIBLE], I SHOULD ALSO
SAY THAT I DO NOT OWN SHARES.
I JUST HAPPEN TO THINK IS THIS
IS A GOOD APPROACH FOR THE
FOLLOW UP.
BUT THERE ARE PROBABLY OTHER
COMPANIES OUT THERE THAT COULD
PROVIDE RE REFOR SOME OF THESE.
IF YOU'RE FOCUS IS ON A GENE,
THERE ARE MULTIPLE VARIANTS THAT
APPEAR ASSOCIATED AS OPPOSED TO
A SINGLE VARIANT, THEN THERE IS
A NEW TECHNOLOGY COMING OUT THAT
IS REALLY EXACTLY PERFECT FOR
THAT, WHERE YOU CAN TAKE ON THE
EXXONIC REGIONS OF TO 30 OR 40
GENES, SO 400 EXXONS, SOME
NUMBER OF HUNDREDS OR THOUSANDS
OF SAMPLES.
AND THAT COST MIGHT BE 50 OR
$100 A SAMPLE.
THIS IS THE FOLLOW-UP FRAMEWORK.
I'M GOING TO TELL YOU ABOUT A
FEW STUDIES WE HAVE DONE TO CARE
OUT THESE STUDIES.
THE FIRST ONE IS TAKING A
PARTICULAR CLASS OF GENETIC
VARIATION.
AND SYSTEMICALLY TAKING ALL. 
VARIANTS THAT ARE IDENTIFIED FOR
FOLLOW UP FROM THAT CLASS OF
VARIATION.
SO WE FOCUSED FIRST ON VARIANTS
THAT INTRODUCE PREMATURE PROTEIN
TRUNCATION.
AND -- TRUNCATION.
THE REASON WE DID THAT WAS
TWO-FOLD.
THESE ARE FUNCTIONAL VARIANTS
AND THE OTHER REASON IS THAT WE
KNOW FOR PSYCHIATRIC DISEASE,
THAT COPY NUMBER VARIANTS, WHICH
DO THE SAME THING, BUT TO
MULTIPLE GENES IN HETEROZYGOTE
FORM REDUCED THE EXPRESSION ARE
ASSOCIATED WITH MANY PSYCHIATRIC
DISEASES, THAT MAYBE SUGGESTIONS
THAT MUTATIONS THAT ARE POINT
MUTATIONSTATIONS CAN KNOCK DOWN
THE EXPRESSION OF A GENE MIGHT
BE RISK FACTORS.
SO WHAT WE DID WAS COMBINE
SAMPLES THAT WE WERE SEQUENCING
IN SKITS -- SCHIZOPHRENIA AND
EPILEPSY.
AND WE LOOKED AT THE SEQUENCE
DATA THAT WE GENERATED, LOOKED
AT ALL THE STOP VARIANTS THERE,
FOCUSING ON 114 GENOMES WE HAD
SEQUENCED WITH SCHIZOPHRENIA, 9
6 WITH EPILEPSY, COMPARED TO 150
CONTROLS.
ALL OF THE VARIANTS THAT
INTRODUCED PREMATURE STOPS THAT
WERE PRESENT IN HOMO ZYGOTE FORM
WE TOOK FOR FOLLOW UP.
AND THEN WE ALSO LOOKED AT A
SPECIAL REGION OF THE GENOME.
FOR EXAMPLE, REGIONS WHERE THERE
WERE COPY NUMBER VARIANTS
IMPLICATED ALREADY.
WE TOOK THEM AND PUT THEM INTO A
BEAD EXPRESSION, GENOTYPING
ASSAY IN A LARGER NUMBER OF
SAMPLES.
SO THAT'S THE KIND OF
INTERMEDIATE SCREEN TO SEE WHICH
OF THE VARIANTS LOOK
INTERESTING.
THERE WERE 384 STOP GAINS THAT
WE WERE FOLLOWING UP.
THESE ARE GENOTYPED IN MORE
CONTROLS IN MORE CASES.
AND THEN WE FOUND IN THAT FOLLOW
UP THAT 16 WERE PRESENT IN MORE
THAN 2 CASES AND NO CONTROLS.
AND THE REST OF THEM WENT AWAY
BECAUSE WE ACTUALLY FOUND THAT
THEY APPEARED IN CONTROLS.
AND SO NOW WE'RE DOWN FROM THE
384 TO 16.
AND WE FOLLOW UP THOSE 16 NOW
WITH THE PACKMAN APPROACH AND
REMAIN WITH 6 STILL INVARIANT.
I DON'T KNOW THAT WE'RE ZEROING
IN HERE.
WE DON'T HAVE YET A P VALUE THAT
IS CONVINCING.
YOU HAVE TO SAY WHAT WAS YOUR
ORIGINAL OPPORTUNITY SPACE.
AND THE ORIGINAL OPPORTUNITY
NASE DEPENDING HOW YOU DEFINE IT
COULD BE ONLY THE STOP VARIANTS
THAT ARE SEEN IN THE GENOMES OR
IT COULD BE ALL THE, SAY,UNE
EQUIVICAL FUNCTIONAL CATEGORIES,
LIKE ESSENTIAL SPLICE.
YOUR P VALUE MIGHT BE AROUND 10
TO THE NEGATIVE 5 OR 6.
AND 24 IS WHAT WE'RE GETTING SO
FAR.
BUT NOW, OF COURSE, WE WILL
CARRY OUT INTO FURTHER GENOME
TYPING AND THE HOPE IS THAT THIS
CHAIN WILL EVENTUALLY IDENTIFY
SUFFICIENT EVIDENCE FOR
ASSOCIATION.
I THINK THAT THAT REALLY IS
GOING TO BE ONE OF THE PRIMARY
DESIGNS FOR DISCOVERY UNTIL WE
CAN GET TO THE POINT OF REALLY
DOING TENS OF THOUSANDS OF
GENOMES.
I THINK THIS BE A PRIMARY ENGINE
OF DISCOVERY.
THE OTHER COMMENT I'D LIKE TO
MAKE IS THAT ONE OF THE
BEAUTIFUL THINGS I THINK ABOUT
HUMAN GENETICS, IT REALLY IS
SUPPOSED TO TELL YOU ABOUT
THINGS THAT YOU DIDN'T ALREADY
KNOW ABOUT.
AND FOR THAT REASON, WE HAVE
TAKEN A VERY STRONG VIEW THAT
YOU REALLY DON'T PICK GENES.
YOU LET THE GENETICS SPEAKING
FOR ITSELF.
YOU TAKE WHAT GENES COME OUT AND
WHATEVER THEY ARE.
SO THAT IS VERY MUCH OUR
PHILOSOPHY.
BUT EVEN SO, I HAVE TO SAY THAT
EVEN FROM THAT COMMITTED HUMAN
GENETICS PERSPECTIVE, SOME OF
THE GENES THAT YOU END UP WITH
ARE REALLY PRETTY CRUMBY FROM
THE PERSPECTIVE OF NEURO
PSYCHIATRIC DISEASE.
THIS ONE IN PARTICULAR, I DON'T
KNOW IF ANYONE CAN READ THAT,
TURNS OUT THAT THAT GENE IS ONE
OF THE PRINCIPAL COMPONENTS OF
HAIR.
AND SO IT DOESN'T MAKE IT HIGH
ON ANYBODY'S LIST FOR EPILEPSY
OR SCHIZOPHRENIA RISK FACTORS.
HOWEVER, IF YOU LOOK AT THE
MOUSE BRAIN ATLAS IT TURNS OUT
THAT THIS IS, INDEED, EXPRESSED
IN THE MOUSE BRAIN.
I HAVEN'T A CLUE WHAT THAT MEANS
BUT I ACTUALLY WOULD NOT BEING
FACETIOUS, I WILL TAKE THE VIEW
YOU CAN'T PICK AND CHOOSE.
YOU REALLY DO THE GENETICS AND
SEE WHERE IT TAKES YOU.
SOMETIMES IT SHOWS YOU SOMETHING
YOU WERE NOT EXPECTING.
SO THAT'S AN EXAMPLE OF
FOLLOWING UP CLASSES OF GENETIC
VARIANTS.
WE HAVE BEEN ALSO BEEN THINKING
ABOUT THE IDEA OF ZEROING ON ON
PARTICULAR EXTREMES.
I'LL TELL YOU ABOUT A TO.  WE'RE
PARTICULARLY INTERESTED IN THE
IDEA THAT YOU CAN USE DRUG
RESPONSE TO ZERO IN ON A MORE
HOMO GENIUS SUBJECT SET OF
PATIENTS.
SO IN BOTH EPILEPSY AND IN
SCHIZOPHRENIA WE'RE LOOKING AT
PARTICULAR SUBGROUPS BASED UPON
RESPONSIVENESS.
IN IN EPILEPSY, THERE IS A
DISTINCTION OF PATIENTS WHEREBY
THE MAJORITY OF THEM RESPOND
VERY, VERY WELL TO
[INDISCERNIBLE] ACID.
A MINORITY DO NOT HAVE THEIR
SEIZURES CONTROLLED.
SO WE HAVE TAPPED INTO OUR
CLINICAL COHORTS IN EPILEPSY TO
TRY TO BUILD A COHOLY SPIRIT OF
REFRACTORY -- COHORT OF
REFRACTORY EPILEPSY.
SO FIRST ON THE JME SIDE.
AS I SAID, A FRACTION OF
INDIVIDUALS DO NOT HAVE THEIR
SEIZURES WELL CONTROLLED.
WE COMBINED SELECTION OF THOSE
INDIVIDUALS FROM SPORADIC
FAMILIES WITH AGAIN A FAMILY
STUDY BEING DONE THROUGH NINDS
FUNDING.
AND COMBINING BOTH WHAT WE SEE
IN FAMILIES WITH WHAT WE SEE FOR
SPORADIC CASES.
WE HAVE BEEN SEQUENCING JME
PATIENTS.
AND THEN, AGAIN, ALL THE
VARIANTS THAT SHOW EITHER
PRESENCE OR ENRICHMENT IN THE
REFRACRY PATIENTS, WE'RE
CARRYING INTO FOLLOW UP
GENOTYPING AND LARGER COHORTS.
HERE IS AN EXAMPLE OF SOME OF
THE VARIANTS THAT HAVE BEEN
TAKEN INTO FOLLOW UP GENOTYPING.
THIS IS BASED UPON ENRICHMENT IN
THE CASES, EITHER ON A RECESSIVE
MODEL, OR IN REGIONS THAT SHOW
LINKAGE FOR JME, OR IN FAMILIES
AND THAT FOLLOW-UP GENOME TYPING
FOR 9 VARIANT OF INTEREST SHOWED
THAT 8 OF THEM WERE NOT INVOLVED
AT ALL.
SO WE COULD RULE THEM OUT.
AND ONE OF THEM THEY WANTONED
QUITE -- STRENGTHENED QUITE
SHORTSTOPLY.
IT'S POSSIBLE THAT THAT JEAN MAY
BE A RISK FACTOR FOR JME.
I DO NOT KNOW YET WHETHER THAT'S
TRUE.
I'LL NEED YET MORE FOLLOW-UP
GENOTYPING BUT WE'RE PROBABLY ON
A PATHWAY THAT WILL EVENTUALLY
WORK.
AND I GUESS THE ONE THING THAT I
REALLY, REALLY LIKE ABOUT THIS
CANDIDATE GENE, IT HAS NOTHING
TO DO WITH HAIR.
THIS IS A TRANSPORTING WHICH
TRANSPORTS GLUTAMATE, PROBABLY.
SO IT'S AN OKAY CANDIDATE GENE.
BUT YOU FOLLOW WHERE THE GENETIC
ASSOCIATION TAKES YOU.
WHERE ARE WE IN TERMS OF P.
VALUES?
YOU THROW ALL THE CASES
TOGETHER, YOU'RE ACTUALLY -- P.
VALUE IS FINE, 10 TO THE
NEGATIVE 7.
THAT IS NOT OKAY TO DO.
WE KNOW VERY WELL THAT
POPULATION STRUCTURE CAN INFLATE
THESE ASSOCIATION STATISTICS.
SO YOU DON'T GET TO IGNORE THAT
JUST BECAUSE YOU'RE LOOKING AT
SEQUENCE DATA.
IF YOU, THEN, TAKE THESE
VARIANTS AND DO A PROPER CONTROL
FOR STRATIFICATION IN THE SAME
KIND OF WAY WE USED TO DO FOR
GWAS, SO IN OUR CASE WE DID IT
FOR EUROPEAN ANCESTRY, P. VALUE
IS LOWER.
AND WE'RE NOT WHERE WE NEED TO
BE.
I THINK ACTUALLY FOR THAT ONE WE
MAY BE MOVING IN THAT DIRECTION.
AND WE'RE DOING A VERY SIMILAR
STUDY FOR TREATMENT RESISTANT
SCHIZOPHRENIA.
THIS IS TAPPING INTO A VERY LONG
TERM SET OF COHORTS THAT HERB
MELTER HAS BEEN DEVELOPING WITH
VERY WELL CHARACTERIZED
TREATMENT SCHIZOPHRENIA AND
TAPPING INTO COHORTS BEING
DEVELOPED WITH RO1 FUNDING AND
ALSO A COLLABORATOR OF HOURS AT
McLANE HOSPITAL, DEBORAH, SO
THESE SOURCES ARE DESCRIBED
HERE.
IS THE DESIGN IS WE'LL BE
SEQUENCING AS MANY OF THE
TREATMENT RESISTANT PATIENTS AS
WE CAN.
IT'S A COMBINATION OF WHOLE
EXOME AND WHOLE GENOME, THEN
TAKING THE VARIANTS OF INTEREST
IN THE SAME DESIGN AS BEFORE.
SO FAR WE HAVE ABOUT 100 OF
THOSE COMPLETE.
THE LAST 2 RELATIVELY EXTREME
TRAITS OR VARIANCE TRAITS ARE
DRUG INDUCED AND RESISTANCE TO
INFECTION BY HIV.
THE DRUG INDUCED EXAMPLE IS
INTERESTING IN THAT THIS
REPRESENTS A VERY DRAMATIC
EXTREME FOR THE POPULATION.
FOR SOME DRUGS IT MAY BE ONE OF
10,000, 50,000 OR FEWER
EXPOSURES THAT CAN LEAD TO
SERIOUS LIVER INJURY.
SO WE'RE TALKING ABOUT A REAL
MINORITY OF THE POPULATION.
AND IT'S ENTIRELY POSSIBLE THAT
THERE ARE IN THOSE CASES
EXTREMELY STRONG GENETIC CONTROL
THAT WOULD BE WOULD BE PICK UP
IN SMALL SAMPLE SIZES.
IF YOU, THEN, FOUND THE GENETIC
CONTROL IT WOULD MAKE A REAL
IMMEDIATE DIFFERENCE.
YOU COULD, THEN, AVOID THE DRUG
IN THOSE INDIVIDUALS THAT ARE
SUSCEPTIBLE.
THERE ARE INDICATIONS OF EXACTLY
THAT KIND OF GENETIC CONTROL FOR
DRUG INDUCED LIVER INJURY, FOR
EXAMPLE, WITH [INDISCERNIBLE],
HAVING SUCH A STRONG DEPENDENCE
ON [INDISCERNIBLE] 5701.
THIS IS REALLY ENCOURAGED US TO
PURSUE SEQUENCING EFFORT AND
DRUG INDUCED LIVER INJURY, LED
BY TOM URBAN IN THE GROUP.
AND WE HAVE BEEN ABLE TO
ESTABLISH A COLLABORATION WITH A
REALLYCEPTIONAL RESOURCES --
EXCEPTIONAL RESOURCE PUT
TOGETHER BY THE DYLAN
CONSORTIUM, AND THIS IS
PARTICIPATING SITES.
THIS IS A CONSORTIUM HEADED BY
PAUL WATKINS.
AND WE ARE NOW PURSUING A
VARIETY OF SEQUENCING STUDIES IN
PARTNERSHIP WITH THEM.
THEY HAVE A TOTAL OF 1,000
PATIENTS WITH CAREFULLY
CONFIRMED DRUG INDUCED LIVER
INJURY, WHETHER THEY REALLY HAVE
A HIGH DEGREE OF CONFIDENCE THAT
THEY KNOW A DRUG THAT'S
RESPONSIBLE.  THERE ARE A MODEST
NUMBER OF DRUGS WHERE THERE ARE
LARGEST NUMBER OF PATIENTS
EFFECTED AND A BUNCH OF DRUGS
WITH ONLY A FEW INDIVIDUALS.  SO
THE DESIGN THAT WE HAVE BEEN
FOLLOWING IS TO SEQUENCE ALL THE
CASES WHERE THERE ARE MORE THAN
5 CASES FOR A DRUG.
AND THEN CARRY OUT THOSE
ANALYSES FIRST.
THEN WE MAY GO BACK AN EVEN
SEQUENCE WHERE THERE IS ONLY A
FEW -- FEWER THAN 5 FOR EACH
DRUG TO LOOK FOR COMMON RISK
FACTORS ACROSS THE DRUGS.
SO FAR WE HAVE DOPE SEQUENCING
FOR ACID AND [INDISCERNIBLE].
VERY SMALL NUMBERS.
AND SO WE DON'T HAVE ANYTHING
DEFINITIVE YET.
ALTHOUGH I SAW A VERY POSSIBLY
GOOD POSSIBILITY WITH
[INDISCERNIBLE] ACID.
FINALLY, THE AREA WHERE THERE IS
THE MOST PROMISE FOR NEAR TERM
DISCOVERY, AND MAYBE ALREADY A
DISCOVERY.
THAT IS LOOKING AT THE GENETIC
CONTROL OF WHO CAN AND WHO
CANNOT BE INFECTED WITH HIV. 
THIS WORK IS LED BY KIM AND IT
CAPITALIZED ON A COHORT OF
INDIVIDUALS THAT WERE EXPOSED
BETWEEN 79 AND 82, TO EITHER
KNOWN OR SUSPECTED CONTAMINATED
INFUSION FACTOR AS PART OF THEIR
TREATMENT FOR HEMOPHILIA.
THIS COHORT IS BEING ASSEMBLEED,
AN NIH FUNDED NETWORK, TO STUDY
THE IMMUNOLOGY.
THERE ARE 500 HA HAD BEEN
ASSEMBLED THAT HAD IF FUSIONS
AND DID NOT BECOME HIV POSITIVE,
ALTHOUGH THE MAJORITY DID.
WE ARE MAKING USE IN THIS EFFORT
OF ANOTHER COHORT THAT JIM
GETTER ASSEMBLED AT THIS NCI
MANY YEARS BEFORE.
AND THAT COHORT PROBABLY HAS
SLIGHTLY HIGHER OVERALL EXPOSURE
RATES, BECAUSE JIM PUT A GREAT
DEAL OF EFFORT INTO SELECTING
THE VERY MOST EXPOSED
INDIVIDUALS.
THOSE INDIVIDUALS ARE AVAILABLE
TO MIND THE GENOME FOR VARIANTS.
WHAT I'D LIKE TO EMPHASIZE HERE
IS THAT WE KNOW OF ONE GENETIC
FACTOR THAT CERTAINLY CONFERS
PROTECTION AGAINST HIV
INFECTION.
THAT'S HOMO DIZZY GOESTY FOR
DELETION.
SO IF YOU DON'T MAKE THE
APPROPRIATE FORM, THEN MOST
FORMS OF HIV CAN INFECT.
BUT WE ALSO KNOW THAT THERE ARE
OTHER PEOPLE, PROBABLY A
MAJORITY OF PEOPLE THAT DON'T
HAVE THAT, THAT CAN'T BE
INFECTED.
THEY ALMOST CERTAINLY HAVE
SOMETHING YET TO BE FOUND
GENTLY.
THAT'S WHAT WE'RE TRYING TO DO.
THAT'S THE GOAL, TO FIND THESE
PROTECTIVE MECHANISMS.
ONE OF THE REASONS WE'RE
PARTICULARLY INTERESTED IN DOING
THAT, IS THE IDEA THAT IF THE
STRATEGY FOR CONTROLLING HIV
WERE TO TURN MORE THAN TOWARD
PROPHYLAXIS, NOT SAKE IT IT
WILL, BUT THEN IT'S MORE DRUGS
AND OPTIONS FOR
PROPHYLACTICTREME, AND ONE ROUTE
IS TO FIND DRUGS THAT WORK ON
THE HOST.
WE NEED TO FIND MORE NATURAL
VARIANTS.
GIVING CLUES TO DRUG
DEVELOPMENT.
SO FAR WE HAVE SEQUENCED 50
INDIVIDUALS THAT WERE EXPOSED
FROM 79-82, NOT HIV POSITION.
48 AT THE TIME.
NOW DONE 50.
WE'RE COMPARING THESE EITHER TO
WHOLE GENOME SEQUENCE THAT WE
GENERATED IN THE LAB THAT ARE
NOT EXPOSED OR WHOLE EXOME
SEQUENCE AND LOOKING FOR
VARIANTS THAT ARE ENRICHED.
WHEN WE DID THAT, WE ACTUALLY
FOUND SOME QUITE STRIKING
ASSOCIATION IMMEDIATELY JUST IN
THE SEQUENCING DATA.
AND SO HERE ARE A FEW OF THE
MOST INTERESTING ASSOCIATIONS
THAT EMERGED FROM THE SEQUENCE
DATA.
SO WE HAVE VARIANTS BASED ON THE
DATA ACHIEVE SIGNIFICANCE QUITE
READILY.
IN 2 CASES THESE VARIANT ARE IN
GENES THAT CAME IN IN AN RNAi
SCREEN FOR GENE PRODUCTS THAT
NOONENING REPLICATION OF NIH.
YOU DON'T KNOW ANYTHING UNTIL
YOU FOLLOW UP GENOTYPES.
I'M NOT CLAIMING IT'S REAL.
WE NEED TO FOLLOW IT BUT THIS
LOOKS VERY ENCOURAGING AT THIS
STAGE.
WE'RE NOW IN THE PROCESS OF
FOLLOWING UP THESE VARIANTS AND
OTHER VARIANTS.
ONE POINT I REALLY DO WANT TO
EMPHASIZE, IN DOING ANY OF THIS
KIND OF WORK IT'S ABSOLUTELY
CRITICAL TO HAVE REALLY GOOD
FOLLOW UP RESOURCES TO CONFIRM
THE EFFECT OF THE VARIANCE.
AND IN OUR CASE FOR HIV, WE'RE
IN A TERRIFIC POSITION BECAUSE
OF THE COHORT AND BUILDING GOING
ON.
THE WAY TO FOLLOW THIS UP IS TO
JOBO TYPE VARIANTS OF INTEREST,
AND FURTHER, TO -- WE HAVE
AROUND 500 OF THE EXPOSED
INDIVIDUALS, ALL INDIVIDUALS
EXPOSED THROUGH SEXUAL CONTACT.
WE CAN LOOK AT THE VARIANTS IN
THE GENERAL POPULATION.
AND THE EXPECTATION, OF COURSE,
IS THAT THERE WILL BE AN
EASTBOUND -- ENRICHMENT IN THE
EXPOSED BUT NOT INFECTED
INDIVIDUALS RELATIVE TO THE
GENERAL POPULATION.
FINALLY WE CAN LOOK AT VARIANTS
IN HIV POSITIVE COHORTS.
THERE WE EXPECT THE VARIANTS TO
BE DEPLETED.
AND SO THIS WAS OUR ALREADY, IN
OUR OWN GROUP, THE NUMBERS THAT
WE HAVE AVAILABLE FOR THIS
FOLLOW UP EFFORT.
THIS IS OVER 500.
SO WE REALLY ARE HERE IN A
POSITION TO BE ABLE TO CONFIRM
ANYTHING WE FIND.
THE SEQUENCING HAS TO GIVE US
THOSE OPTIONS.
ONE COMMENT HERE, I THINK THIS
REALLY ILLUSTRATES THE
IMPORTANCE IN DISCOVERY GENETICS
OF GROUPS REALLY WORKING
TOGETHER.
I THINK WHAT WE DON'T WANT TO
SEE IS ONE GROUP TRYING TO GET
EVERYTHING THAT HE POSSIBLY CAN
OUT OF SEQUENCING 150
INDIVIDUALS, BUT NOT HAVING MANY
FOR FOLLOW-UP, ANOTHER WITH 200,
SO ON.
AS FAR AS IS POSSIBLE AND
REASONABLE, THIS WORK DOES WORK
BEST WHEN IT'S PURSUED TOGETHER
AND YOU CAN ACTUALLY HAVE THIS
KIND OF A SIZE, FOLLOW-UP COHORT
FOR CONFIRMATION THAT WILL ALSO
REDUCE THE NUMBER OF FALSE
POSITIVES THAT ARE CONTRIBUTED
TO THE LITERATURE.
SO THAT'S REALLY WHAT I WANTED
TO TALK ABOUT.
I'D LIKE TO END WITH SOME
COMMENTS ABOUT WHERE I SEE
THINGS GOING.
ONE OF THE THINGS, I THINK,
THAT'S REALLY CRITICAL TO
APPRECIATE IN THE WAY GENETICS
IS MOVING IS THAT WHEN WE FIND
VARIANTS THAT HAVE RELATIVELY
HIGH IMPACT ON THE TRAIT, IT'S
REALLY IMPORTANT TO BE ABLE TO
LOOK AT THE SUBJECT THAT CARRIES
IT AND THE SUBJECT'S FAMILY, TO
WORK OUT WHAT THAT VARIANT DOES.
SO FOLLOW UP PHENOTYPING BECOMES
MORE IMPORTANT.  THIS IS VERY
COMMON IN MENDELIANIAN GENETICS
LESS SO IN THE AREA OF GENOME
WIDE ASSOCIATION STUDY.
NOW WE NEED TO GO BACK IN THAT
DIRECTION AND WE WANT TO BE
SEQUENCING FOR DISCOVERY FOR
COMMON DISEASES.
INDIVIDUALS THAT ARE STILL
AVAILABLE FOR FURTHER
EVALUATION, THAT'S REALLY
IMPORTANT.
WE HAVE RETURNED TO CLASSICAL
HUMAN JOKES.
ANOTHER CRITICAL QUESTION WE
HAVE TO FACE AS EARLY AS
POSSIBLE IS THE EXTENT TO WHICH
CAUSING MUTATIONS THAT WE
IDENTIFY CLUSTER INTO COMMON
PATH WEIGHS.
ONE OF THE CONCERNS ABOUT HOW TO
PERSONALIZE MEDICINE, IF A LOT
OF THE ACTION OR RARE VARIANTS,
THAT EVERYBODY HAS A DIFFERENT
CAUSE OF DISEASE.
TO A CERTAIN AGREE THAT WILL
HAPPEN, TO A CERTAIN DEGREE.
BUT IT'S ENTIRELY POSSIBLE THAT
THE DIFFERENT MUTATIONS THAT
INDIVIDUALS HAVE WILL BE
CLUSTERABLE INTO COMMON
PATHWAYS.
THAT WILL BE A REALLY IMPORTANT
AREA 067 FOCUS.
THE FINAL COMMENT, TO THE EXTENT
WE START IDENTIFYING RELATIVELY
HIGH IMPACT RARE VARIANTS, I
THINK IT WILL BE POSSIBLE TO
MAKE MORE MEANINGFUL
DISTINCTIONS AMONGST INDIVIDUAL
IN TERMS OF RISKS THEY FACE FOR
DISEASE.
THAT MIGHT START TO HAVE
IMPLICATIONS FOR PROPHYLAXIS.
YOU CAN IMAGINE IF YOU COULD
PREDICT WHAT INDIVIDUALS WILL
DEVELOP SCHIZOPHRENIA YOU MIGHT
START TREATMENT OR EVALUATE
TREATMENT OPTIONS.
SO I THINK WE'RE IN A VERY
EXCITING TIME.
THERE WILL BE A TREMENDOUS
AMOUNT OF SCORCHRY WITH
SEQUENCING, WE HAVE TO BE
EXTREMELY CAREFUL ABOUT HOW WE
DEPLOY IT TO MAKE SURE WE GET IT
AS RIGHT AS WE CAN MOVING
FORWARD.
THAT'S WHERE I WANTED TO END AND
THANK THE PEOPLE THAT HAVE BEEN
WORKING WITH ME ON THE VARIOUS
PROJECTS.
I THINK THAT I MENTIONED, I
HOPE, MOST OF THEM.
THE OTHERS ARE INDICATED ON THE
SLIDE.
THANK YOU VERY MUCH FOR YOUR
ATTENTION.
>> WE HAVE TIME FOR 5 MINUTES OF
QUESTIONS AND YOU CAN JOIN US
FOR A RECEPTION AT 4:00.
THERE IS CERTAINLY TIME FOR
QUESTIONS FROM THE AUDIENCE.
I'LL START WITH THE FIRST ONE.
SO ONE THING I WASN'T SURE I
FOLLOWED WHEN YOU WERE TALKING
ABOUT ONE OF THE STORIES ABOUT
SCHIZOPHRENIA, YOU SAID YOU HAD
DATA FROM PLATFORMS AND SANGER
BASED.
YOU SEEMED UNCLEAR ABOUT WHAT
THE TRUTH WAS.
>> FOR THIS ONE VARIANT IN DRD5
IT TURNS OUT THAT THE NEXT
GENERATION SEQUENCE DATA IS
MESSY, AND SANGER IS A BIT
MESSY.
>> IS IT JUST PART OF THE.
JUST PART -- IT'S HARD TO
SEQUENCE
SEQUENCE CLEANLY.
I THINK WE'LL EVENTUALLY GET IT
CLEAN ENOUGH SEQUENCE.
RIGHT NOW, WE DON'T.
AND IT REALLY JUST -- I MEAN IT
DOES EMPHASIZE THAT IN THE GWAS
WE WERE FOCUSING ON VARIANT THAT
WE KNOW HOW TO TYPE WELL.
NOW WE'RE DEALING WITH VARIANTS
THAT WE DON'T KNOW HOW TO TYPE
WELL.
>> DID THE GENOTYPING CHIP GIVE
YOU A ACCURATE CALL AT THAT
POSITION?
>> BECAUSE THAT VARIANT IS NOT
KNOWN TO ANYBODY IT'S NOT ANY
CHIP AT ALL.
>> GOT IT.
OKAY.
>> QUESTIONS?
NO QUESTIONS FROM THE AUDIENCE?
IF NOT WE CAN START THE
RECEPTION EARLY.
A RECEPTION JUST A FEW YARDS
DOWN IN THE HALL.
>> THANK YOU VERY MUCH.
[APPLAUSE]