MicroarrayandBioinformatics

基因芯片的生物信息學(xué)

AimsfortheMicroarrayBioinformaticsUnderstandbasicmicroarraytechnologyanditsuseingeneexpressionanalysis.基因芯片技術(shù)與表達(dá)譜分析中的應(yīng)用Learnbasicdataanalysismethodsandhowtoapplythemintheanalysisofgeneexpressiondata基因芯片的數(shù)據(jù)分析Dataacquisition數(shù)據(jù)獲得Datanormalization數(shù)據(jù)歸一化Dataanalysis數(shù)據(jù)分析DataClustering數(shù)據(jù)聚類Vocabulary-Review回顧Gene基因:

hereditaryDNAsequenceataspecificlocationonchromosome.Genetics遺傳學(xué):

studyofheredity&variationinorganisms.Genome基因組:anorgan'stotalcontent(fullDNAsequence)Genomics基因組學(xué):studyoforganismsintermsoftheirgenome.2002年2月12日,歷時10載耗資20億美元的人類基因組計劃最終完成,并報道了99%的人類基因組序列.Vocabulary-Review回顧Protein蛋白質(zhì)

:sequenceofaminoacidsthat“doessomething”Proteomics蛋白質(zhì)組學(xué)

:

studyofalloftheproteinsthatcancomefromanorganismsgenomeBioinformatics生物信息學(xué)

:thecollection,organization&analysisoflarge-scale,complexbiologicaldata.FunctionalGenomics功能基因組學(xué):studyofobtaininganoverallpictureofgenomefunctions,includingtheexpressionprofilesatthemRNAlevelandtheproteinlevelMicroarrayTechnology基因芯片技術(shù)BasicIdeaofMicroarrays

芯片基本原理TypesofMicroarraytechnologiesandhowtheywork芯片技術(shù)種類與原理OutputsofMicroarrays

芯片的制備ImageAnalysisrequiredtotransformoutputtogeneexpressionmatrices芯片數(shù)據(jù)采集與分析Microarray

基因芯片–AhighthroughputtechnologythatallowsdetectionofthousandsofgenesSimultaneously是指通過微陣列技術(shù)將高密度DNA片段陣列通過高速機(jī)器人或原位合成方式以一定的順序或排列方式使其附著在如玻璃片等固相表面，以熒光標(biāo)記的DNA探針，借助堿基互補(bǔ)雜交原理，進(jìn)行大量的基因表達(dá)及監(jiān)測等方面研究的最新革命性技術(shù)。–genechip,biochip,array–Muchrelyoncomputeraids–CentralplatformforfunctionalgenomicsHistory歷史HGP(humangenomeproject):suggestedbyDelbeccoonMar.7,1986，startedinOct.1990,rapidandsensitivetechniquesforhumangenomeinformationanalysis80S:

suggestionbasedoncomputerchip,WBrainstrieditfirstly.90S:

StephenFodor(PresentofAffymetrixnow)madeitsuccessfully.1995:QuantitativemonitoringofgeneexpressionpatternswithacomplementaryDNAmicroarray

Endof1996:

thefirstDNAchip1998年底美國科學(xué)促進(jìn)會將基因芯片技術(shù)列為1998年度自然科學(xué)領(lǐng)域十大進(jìn)展之一

1997年《財富》雜志：在20世紀(jì)科技史上有兩件事影響深遠(yuǎn)：一是微電子芯片，它是計算機(jī)和許多家電的‘心臟’，改變了我們的經(jīng)濟(jì)和文化生活，并已進(jìn)入每一個家庭；另一就是生物芯片，它將改變生命科學(xué)的研究方式，革新醫(yī)學(xué)診斷和治療，極大地提高人口素質(zhì)和健康水平?！盚istory歷史MicroarraysarePopular芯片技術(shù)的普及AtNYUMedCenternowcollectingabout3GBofmicroarraydataperweek(60chips,6-10differentexperiments)PubMedsearch“microarray”=24,431papers（2008）38,467(2010)MicroarraysarePopular芯片技術(shù)的普及NSFCHistory芯片中國——博奧生物博奧生物有限公司暨生物芯片北京國家工程研究中心成立于2000年9月30日，現(xiàn)已迅速發(fā)展成為國際領(lǐng)先的生物高技術(shù)公司2002年，博奧生物作為亞洲唯一入選的公司，被美國《財富》雜志評為2002年度“全球最有發(fā)展前景的生物技術(shù)公司”；

2005年，博奧生物成為美國RedHerring雜志評選的“亞洲非上市技術(shù)100強(qiáng)公司”中的八家生物技術(shù)公司之一

2007年，博奧生物申報的“系統(tǒng)化生物芯片和相關(guān)儀器設(shè)備的研制及應(yīng)用項目”榮獲國家技術(shù)發(fā)明獎二等獎。

2008年，全國生物芯片標(biāo)準(zhǔn)化技術(shù)委員會經(jīng)國務(wù)院標(biāo)準(zhǔn)化主管部門批準(zhǔn)成立，生物芯片北京國家工程研究中心作為秘書處單位負(fù)責(zé)日常管理，中心主任程京博士任標(biāo)委會主任委員。

2009年，程京博士當(dāng)選工程院院士2009年12月28日，博奧生物研發(fā)生產(chǎn)和服務(wù)基地正式簽約入駐位于四川溫江的成都國際醫(yī)學(xué)城，同一天，成都博奧獨立醫(yī)學(xué)實驗室也正式開始運行

芯片中國——中國生物芯片中心世界上第一張遺傳性耳聾基因檢測芯片；世界上第一張17種分枝桿菌菌種鑒定芯片；世界上第一張結(jié)核耐藥檢測芯片；世界上第一張乙肝耐藥基因檢測芯片；世界上第一張SARS病毒檢測芯片；世界上第一張HLA分型檢測芯片；世界上第一張轉(zhuǎn)錄因子活性譜檢測芯片；世界上第一張細(xì)胞電旋轉(zhuǎn)檢測芯片；世界上第一張家蠶全基因組表達(dá)譜芯片；2007年開始局部贏利，2009年銷售過億并全面贏利，銷售收入平均每年增長75%。預(yù)計公司2015年銷售收入將達(dá)到30億元以上。863計劃與973計劃863計劃標(biāo)志863計劃即國家高技術(shù)研究發(fā)展計劃，是中華人民共和國的一項高技術(shù)發(fā)展計劃。這個計劃是以政府為主導(dǎo)，以一些有限的領(lǐng)域為研究目標(biāo)的一個基礎(chǔ)研究的國家性計劃。中國根據(jù)本身的經(jīng)濟(jì)實力，以“有限目標(biāo)，突出重點”為方針，主要的科學(xué)研究集中在生物技術(shù)、航天技術(shù)、信息技術(shù)、激光技術(shù)、自動化技術(shù)、能源技術(shù)、新材料領(lǐng)域以及1996年新加的海洋高技術(shù)。1997年由國家科技領(lǐng)導(dǎo)小組第三次會議決定實施的一項計劃。實施“973計劃”的戰(zhàn)略目標(biāo)是加強(qiáng)原始性創(chuàng)新，在更深的層面和更廣泛的領(lǐng)域解決國家經(jīng)濟(jì)與社會發(fā)展中的重大科學(xué)問題，以提高我國自主創(chuàng)新能力和解決重大問題的能力，為國家未來發(fā)展提供科學(xué)支撐。戰(zhàn)略目標(biāo):加強(qiáng)原始性創(chuàng)新，在更深的層面和更廣泛的領(lǐng)域解決國家經(jīng)濟(jì)與社會發(fā)展中的重大科學(xué)問題，以提高我國自主創(chuàng)新能力和解決重大問題的能力，為國家未來發(fā)展提供科學(xué)支撐。Features特點–Parallelism高平行

?

Thousandsofgenessimultaneously–Miniaturization小型化

?

Smallchipsize–Multiplexing高通量

?

Multiplesamplesatthesametime–Automation自動化

?

Chipmanufacturing

?

ReagentsWhatproblemscanitsolve?

基因芯片的應(yīng)用–Differingexpressionofgenesovertime,betweentissues,anddiseasestates

基因表達(dá)差異–Identificationofcomplexgeneticdiseases

復(fù)雜性基因疾病的診斷–Drugdiscoveryandtoxicologystudies

藥理與毒理學(xué)研究–Mutation/polymorphismdetection(SNP’s)

SNP檢測–Pathogenanalysis

診斷病原DifferentialGeneExpression

基因表達(dá)差異AFewExamples:Celltypespecific

-e.g.skincellvs.braincell

Developmentalstage

-e.g.embryonicskincellvs.adultskincellDiseasestate

-e.g.normalskincellvs.skintumorcellEnvironment-specific -e.g.skincelluntreatedvs.treated drugs,toxinsWhatisitspitfall缺陷與不足?–DetecttranscriptionmRNAlevel,nottranslationproteinlevel–Manyfactors(variations)canaffecttheresult：影響因素眾多

?Chipandprobedesign

?Experimentdesign

?Samplepreparation

?Imageacquisition

?Datanormalization

?Dataanalysis

?

….–Successcrucial成功關(guān)鍵:

?Youknowboththebiologyproblemandthecomputeraids(software,statistics).RequrimentsArrayspotter點樣儀Arrayscanner掃描儀Chemistrysystems雜交體系Softwares

軟件Marketpredict市場預(yù)期

At1999：1billionUSDLessthan5yrs:20billions2005：5billions(USA)2010：40billions(USA)Don’tincludediseasediagnosticThelargestindustryinsteadofmicroelectricsPrinciple原理?

SimilartoNorthern–Base-Pairing,hybridizationbetweennucleiccids

?

MajordifferencesfromNorthern–Detectsthousandsofgenessimultaneously/individual–Probesfixationonglassslide/nylonmembrane–Targetsampleslabelingwithfluorescent/radioactivedNTPTypesofMicroarrays

基因芯片的種類

1.cDNAchip(DNAmicroarray,two-channelarray)cDNA芯片

:–ProbecDNA(500~5,000baseslong)isimmobilizedtoasolidsurfacesuchasglass–Usingrobotspotting–TraditionallycalledDNAmicroarray–FirstlydevelopedatStanfordUniversity2.Genechip(DNAchip,Affymetrixchip)基因芯片:–Oligonucleotide(20~80-meroligos)issynthesizedeitherinsitu(on-chip)orbyconventionalsynthesisfollowedbyon-chipimmobilization–HistoricallycalledDNAchips–DevelopedatAffymetrix,Inc.,undertheGeneChip?trademark–ManycompaniesaremanufacturingoligonucleotidebasedchipsusingalternativetechnologiesSpottingProcess點樣過程

Spotrobot點樣儀Cheungetal.1999點樣針Affymetrix

基因芯片表達(dá)差異檢測ComparisonofProbeTypes

兩種探針比較

AdvantagesNoneedtoisolateandpurifycDNAsbecauseoligonucleotidescanbesynthesized.Shortoligonucleotidesarelesslikelytohavecross-reactivitywithothersequencesinthetargetDNA.DensityofchipsishigherthanwithcDNAs.LimitationsThesequencehastobeknown.Synthesiscanbeexpensiveandtime-consuming.TheshortsequencesarenotasspecificfortargetDNA,soappropriatecontrolsmustbeadded.In-situSynthesis/OligosPCRProducts/cDNAProbesAdvantagesFlexibilitytostudycDNAsfromanysource.cDNAsdonotrequireanyaprioriinformationaboutthecorrespondinggenes.Longersequencesincreasehybridizationspecificity,whichreducesfalsepositives.LimitationsIsolationofindividualcDNAstoimmobilizeoneachspotcanbecumbersome.Densityislowerthansynthesizingoligonucleotidesonthesurfaceofthechip.cDNAsarelongersequencesandaremorelikelytorandomlycontainsequencesfoundintargetDNA,whichresultsincross-reactivity.Manyothervariationsofthetechnologyexist,suchastheuseoflongeroligos,theuseoffibreoptics,etc.HomemadeTailoredCheaper?????Maximum24,000featuresperarrayPronetovariabilityCommerciallyavailable“Offtherack”Moreexpensive?????Maximum500,000featuresperarrayLessvariabilitySpottedArraysAffymetrixArraysProcessofmanufactureamicroarray

芯片制備流程Startwithindividualgenes,e.g.the~6,200genesofthegenomeorY.pestisAmplifyallofthemusingpolymerasechainreaction(PCR)“Spot”themonamedium,e.g.anordinaryglassmicroscopeslideEachspotisabout100μmindiameterSpottingisdonebyarobotComplexandpotentiallyexpensivetaskB21B22B23B24B25B26B27B28B29B30B31B32B17B18B19B20B5B6B7B8B9B10B11B12B13B14B15B16B1B2B3B44×8矩陣17×17點陣一共8448個點；4005條鼠疫菌基因+若干對照DNA；每樣品相鄰重復(fù)兩個點?；蜻x擇4015條芯片點樣基因的PCR擴(kuò)增產(chǎn)物純化和濃縮4005條基因全基因組芯片研制引物設(shè)計MicroarraySteps基因芯片分析過程?

ExperimentandDataAcquisition實驗過程與數(shù)據(jù)獲得–Chipmanufacturing芯片制備–Samplingandlabeling點樣–Hybridization雜交–Imagescaling圖像掃描–Dataacquisition數(shù)據(jù)獲得

?

Datanormalization數(shù)據(jù)歸一化

?

Dataanalysis數(shù)據(jù)分析

?

Biologicalinterpretation生物學(xué)解釋Readinganarray(cont.)BlockColumnRowGeneNameRedGreenRed:GreenRatio111tub12,3452,4670.95112tub23,5892,1581.66113sec14,1091,4692.80114sec21,5003,5890.42115sec31,2461,2580.99116act11,9372,1040.92117act22,5611,5621.64118fus12,9623,0120.98119idp23,5851,2092.971110idp12,7961,0052.781111idh12,1704,2450.511112idh21,8962,9960.631113erd11,0233,3540.311114erd21,6982,8960.59ColorCoding

掃描結(jié)果TablesaredifficulttoreadDataispresentedwithacolorscaleCodingscheme:Green=repressed(lessmRNA)geneinexperimentRed=induced(moremRNA)geneinexperimentBlack=nochange(1:1ratio)OrGreen=controlcondition(e.g.aerobic)Red=experimentalcondition(e.g.anaerobic)WeonlyuseratioMicroarrayDataAnalysis

芯片數(shù)據(jù)分析

?Imageacquisition圖像獲得

?Datanormalization數(shù)據(jù)歸一化

?Dataanalysis數(shù)據(jù)分析-Clusteranalysis

聚類分析Noise干擾Noisesources干擾來源:Samplepreparation,labeling,amplificationReactionvariationsEnvironmentTargetvolumeHybridizationparameters(temperature,time,...)AspecifichybridizationDustScannersettingsQuantizationOtherImageProcessingProblems

SpotQualityProblemsUnevengridpositionsCurveswithinagridVariableSpotsizeorshapeVariableDistancebetweenspotsTypicalProblemsofRawOutputTwoslidesP04vs.P01(pg2)A1vs.P01(pg2)Noisefiltering干擾過濾Gridding:identifyspotlocationsSegmentation:distinguishforegroundfrombackgroundFixedCircle:putacirclearoundtheforegroundareaSeededregiongrowing:identifyinitialspot“seeds”andgrowhighintensityregionsEdgedetectionalgorithmsBackgroundcancellationIntensity=FGintensity-BGintensityNoisefiltering干擾過濾Normalization

歸一化Thewordnormalizationdescribestechniquesusedtosuitablytransformthedatabeforetheyareanalysed.Goalistocorrectforsystematicdifferencesbetweensamplesonthesameslide,orbetweenslides, whichdonotrepresenttruebiologicalvariationbetweensamples.Normalization歸一化NoralizedatatocorrectforartificialvariancesRed=FGred-BGredGreen=FGgreen–BGgreenPixelValue=log2(Red/Green)-log2(Redavg/Greenavg)Pixelcolor:Green ifpixelvalue<0Yellow ifpixelvalue=0Red

ifpixelvalue>0Normalization歸一化Calibrated,redandgreenequallydetectedUncalibrated,redlightunderdetectedTheoriginofsystematicdifferences

系統(tǒng)誤差的產(chǎn)生原因Systematicdifferencesdueto…Dyebiases,whichvarywithspotintensity,Locationonthearray,Plateorigin,PrintingqualitywhichmayvarybetweenPinsTimeofprintingScanningparameters,…Normalization

methods&issues

歸一化方法MethodsGlobaladjustmentIntensity

dependent

normalizationWithin

print-tip

group

normalizationAnd

many

other…Selection

of

spots

for

normalizationDNAarrayDataAcquisition

DNA芯片數(shù)據(jù)的獲得?ImageAnalysissoftwarepackagesexistfortheanalysisoftheoutputofcustommadechips(e.g.GenePixPro,ArrayVision,TIGRSpotFinder,etc)?Needchipdescriptionfile(CDF)–ForprobelocationIntroductionofSoftware----SAM

SAM軟件介紹SignificanceAnalysisofMicroarraysTusher,TibshiraniandChu(2001):Significanceanalysisofmicroarraysappliedtotheionizingradiationresponse.PNAS200198:5116-5121,(Apr24).ExcelpluginFreeMostpublishedmethodofmicroarraydataanalysisHandlingMissingData

丟失數(shù)據(jù)的操作TherearecurrentlytwooptionsforimputingmissingvaluesinSAM.RowAverageEachvalueisimputedwiththeaverageofnon-missingvaluesforthatgene.K-NearestNeighborIntheother(default)option-missingvaluesareimputedusingak-nearestneighboraverageingenespace(defaultk=10):Clustering聚類軟件Hypothesis:

GeneswithsimilarfunctionhavesimilarexpressionprofilesFindgroupofgeneswithsimilarexpressionprofilesFindgroupdofindividualswithsimilarexpressionprofileswithinapopulationClustering

=GroupidentificationBasicissuesinclustering

聚類的基本原則Issues

todecidebefore

clusteringWhichgenes/arrays

touse?Which

similarity/dissimilarity

measure?Which

clustering

algorithm?It’sanexploratorytechniqueThere’snooptimalsolutionAny

method

will

yield

groupsWhich

method

gives

good

groups?ClusteringSteps

聚類分析步驟Chooseasimilaritymetrictocomparethetranscriptionalresponseortheexpressionprofiles:PearsonCorrelationSpearmanCorrelationEuclideanDistance…Chooseaclusteringalgorithm:HierarchicalK-means…Clusteralgorithm

聚類算法 -UnsupervisedAnalysis非監(jiān)督算法

-HierarchicalK-meanSelf-organizingmapsOthers-SupervisedAnalysis:classificationrules

監(jiān)督算法系統(tǒng)聚類法步驟

1、將n個樣品各作為一類;

2、計算n個樣品兩兩之間的距離，構(gòu)成距離矩陣;

3、合并距離最近的兩類為一新類;

4、計算新類與當(dāng)前各類的距離。再合并、計算，直至只有一類為止;

5、畫聚類樹形圖，確定距離切點、類組，解釋。

在SPSS軟件中的操作步驟：

Analyze-----Classify-----HierarchicalHierarchicalClustering

分層聚類法g1g2g3g4g5g10.230.000.95-0.63g20.910.560.56g30.320.77g4-0.36g5g1g4g1g2g3g4g5g10.230.000.95-0.63g20.910.560.56g30.320.77g4-0.36g5Findlargestvalueissimilaritymatrix.Joinclusterstogether.

Recomputematrixanditerate.HierarchicalClustering

分層聚類g1,g4g2g3g5g1,g40.370.16-0.52g20.910.56g30.77g5g1g4g2g3g1,g4g2g3g5g1,g40.370.16-0.52g20.910.56g30.77g5Findlargestvalueissimilaritymatrix.Joinclusterstogether.

Recomputematrixanditerate.HierarchicalClustering

分層聚類g1,g4g2,g3g5g1,g40.27-0.52g2,g30.68g5g1g4g2g3g5g1,g4g2,g3g5g1,g40.27-0.52g2,g30.68g5Findlargestvalueissimilaritymatrix.Joinclusterstogether.

Recomputesimilaritymatrixanditerate.InterpretingtheResults