1Office hoursWednesday 3-4pm304A Stanley HallReview session5pm Thursday, Dec. 11GPB100Association vs. linkageStrong, easy to detect,but rare in population;may not be reflective ofcommon disease.Also, hard to collect familydata.Common but weakeffects; need 1000’sof samples to detect.If no common cause,can fail.UnrelatedindividualsRelatedindividualsAssociation vs. linkagesmallnumber ofgenerations;individualsshare bigchunks ofgenome;can get co-inheritancebetweendistantmarkersmanyrecombinationshave happenedsince commonancestor;shared regionis small; no co-inheritancebetweendistant markersSo you need very high densityof markers to get signal in anassociation study, but you getvery high spatial resolution.Association and admixtureCasesControls==At any one of these loci, Caucasian-like allele will beenriched in control samples.Genotyping by arrayFig. 11.8Expression microarraysFig. 1.132Labeled DNA sample vs.labeled mRNA (cDNA) sampleLabeled DNA sample vs.labeled mRNA (cDNA) sample(reverse transcription in the test tube)Coding sequence arrayFig. 1.13Coding sequence arrayFig. 1.13say, with chemotherapeuticCoding sequence arrayFig. 1.13Coding sequence arrayFig. 1.133Coding sequence arrayFig. 1.13expressionexpressionexpressednot expressedWhy measure expression ofall genes at once?“Regulon” expression responsehttp://www.mcb.mcgill.ca/~hallett/GEP/Lecture2/Image17.gif(e.g. cortisol)“Regulon” expression responsehttp://www2.kenyon. edu/Depts/BioEllipse/courses/biol114/Chap06/week08a_files/regulon.gifFinding regulatoryresponse on a large scaleOligo expression arraytranscripts(soybean component)4Expression profilesTime since drug administeredExpression profilesTime since drug administeredTime since drug administeredExpression profilesTime since drug administeredTime since drug administeredEach color is aregulon, or“cluster,” of co-regulated genesExpression effects of cancerCancer classificationCancer classification5Histological classification isfinicky: can we do better?Diagnosis via transcriptional profileDiagnosis via transcriptional profiletranscriptspatient samplesDiagnosis via transcriptional profileDiagnosis via transcriptional profileNatural variation among “normals”Two human chromosomes differ at ~1/1000 bases.96% of these differences are not in protein-coding sequence.Why?6Two human chromosomes differ at ~1/1000 bases.96% of these differences are not in protein-coding sequence.Most protein coding mutations are deleterious;appear but are culled by natural selection.Natural variation among “normals”Natural variation among “normals”Natural variation among “normals”Genetic variation in mRNA levelsORFTFTFGkinaseTFGenetic variation in mRNA levelsORFTFTFGkinaseTFLikely to be a complex trait.Many mRNA differences at once7Linkage mapping of mRNA levels“Black 6” mouse x “DBA” mouseLinkage mapping of mRNA levels“Black 6” mouse x “DBA” mouse~10% mRNA levels significantly differentLinkage mapping of mRNA levels“Black 6” mouse x “DBA” mouse111 F2 progeny…Linkage mapping of mRNA levels“Black 6” mouse x “DBA” mouse111 F2 progenyMicroarrayeach F2 liver…Linkage mapping of mRNA levels“Black 6” mouse x “DBA” mouse111 F2 progenyMicroarrayeach F2 liver…Genotypeeach F2Linkage mapping of mRNA levels“Black 6” mouse x “DBA” mouse111 F2 progenyMicroarrayeach F2 liver…Genotypeeach F2Looking for linkage (coinheritance) betweenmarker and mRNA level.8Marker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFMarker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFGGMarker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFGGMarker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFGGOne allele = high mRNA,the other = low mRNAMarker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFMarker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTF9Marker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFmRNA level shows linkage to locus of polymorphicregulator(s).Marker is linked to polymorphism inexpression regulation cascadeORFTFTFGkinaseTFmRNA level shows linkage to locus of polymorphicregulator(s).Locally acting polymorphismsLocally acting polymorphismsLocally acting polymorphismsPolymorphism responsible for mRNA difference is at the locusof the gene itselfLocally acting polymorphismsORFTFTFGkinaseTF10Locally acting polymorphismsORFTFTFGkinaseTFLocally acting polymorphismsPolymorphism responsible for mRNA difference is at the locusof the gene itselfLocally acting polymorphismsPolymorphism responsible for mRNA difference is at the locusof the gene itself~25% of varying mRNAsare caused by locallyacting polymorphismNonlocal polymorphismsNonlocal polymorphismsOne polymorphism in a key regulator can affect aregulon: 100’s of related mRNAs.Clinical applications“Black 6” mouse x “DBA” mouse111 F2 progenyMicroarrayeach F2 liver…Genotypeeach F2Measure fatpad each F211Clinical applicationsClinical applicationsColored curves= fat mass atdifferent bodylocationsClinical applicationsFinding polymorphism responsible fordifference in macroscopic phenotype is hardClinical applicationsFinding polymorphism responsible fordifference in macroscopic phenotype is hardIf mRNAs change too, can learn mechanismfrom known function of encoded proteinsClinical applicationsClinical applications12Clinical applicationsClinical applicationsCounts allele 1/allele 2, casesCounts allele 1/allele 2, controls= 1.5Clinical applicationsClinical applicationsMarker predicts quantitative expression level = associationCan we map expression traitsfirst, disease afterward?Linkage of human transcripts13Linkage of human transcriptsAssociation of human transcriptsAssociation of human transcriptslinkage(families)assoc(unrelated)Association in multiple populationsHan Chinese and JapaneseEuropean-Americans in UtahAssociation in multiple
View Full Document