DNA diseases and databases disastrously deficient - PDF document

DNA,diseasesanddatabases:disastrouslydeÞcientGeorgeP.PatrinosandAnthonyJ.BrookesErasmusUniversityMedicalCenter,FacultyofMedicineandHealthSciences,MGC-DepartmentofCellBiologyandGenetics,POBox1738,3000DR,Rotterdam,TheNetherlandsDepartmentofGenetics,UniversityofLeicester,UniversityRoad,Leicester,UKLE17RHRecentprogressindiseasegeneticsandgenome-relatedmedicinehasbeensubstantial,withvastamountsofdatabeinggenerated.However,thisprogresshasnotbeenmatchedbyadequatedatabaseprojectsthatgatherandorganizethesedatatoenabletheiruseful Correspondingauthor:Patrinos,G.P.(g.patrinos@erasmusmc.nl).Availableonline16April2005 TRENDSinGeneticsVol.21No.6June20050168-9525/$-seefrontmatter2005ElsevierLtd.Allrightsreserved.doi:10.1016/j.tig.2005.04.004 mutationsandpolymorphisms,butalmostnoÞnedetailtoaidproperunderstanding.ThebestcurrentexampleofacoredatabaseistheHumanGeneMutationDatabasehttp://www.hgmd.org.hgmd.org,whichbyMarch2005con-45000differentlesionsinalmost1800differentnucleargenes,withnewentriesaccumulatingatanaveragerateof2300perannum.Bycontrast,LSDBscontaininformationaboutoneorafewspeciÞcgenesgenes,usuallyrelatedtoasingledisease.Theyarehighlycuratedrepositoriesofpublishedandunpublishedmutationswithinthosegenesand,assuch,theycomplementthecoredatabases.Dataqualityandcompletenessaretypicallygood,withupto50%ofstoredrecordspertainingtounpublishedmutations.Thedataarealsorichandinformative.Forexample,LSDBswilltypicallypresenteachofthemultiplediscoveriesofrecurrentmutationalevents,enablingmutationhotspotstobeidentiÞed;whenthesemutationsoccurondifferentchromosomalbackgrounds(linkedtoothermutations)suchthattheyresultinseveral,ordifferent,diseasefeatures,thesecorrelationsarealsorecorded.AgoodexampleofanLSDBisHbVar((Ðarelationaldatabaseofhemoglobinvariantsandthalassemiamutations,providinginfor-mationonpathology,hematology,clinicalpresentationandlaboratoryÞndingsfornumerousDNAalterations.Geneandproteinvariantsareannotatedwithrespecttobiochemicaldata,analyticaltechniques,structure,stab-ility,function,literaturereferences,andqualitativeandquantitativedistributioninethnicgroupsandgeographicgeographic.AsiscommoninLSDBs,entriescanbeaccessedthroughsummarylistingsoruser-generatedqueriesthatcanbehighlyspeciÞc.Forinformationon350currentlyavailableLSDBsseehttp://www.hgvs.orghttp://www.hgmd.organdRef.Ref..InadditiontocoredatabasesandLSDBs,DNAvariationisalsorecordedinvariouspolymorphismdatabases,suchasthesinglenucleotidepolymorphismdatabase(dbSNP;http://www.ncbi.nlm.nih.gov/projects/ts/,theHAPMAPDataCoordinationCenterhttp://www.hapmap.org/.hapmap.org/andtheHumanGenomeVariationDatabase(HGVbase;[11Ð12].TheseresourcesmakenoexplicitattempttoconnectDNAinformationtophenotypes,andtheyarenotyetperfectindesignorcontentcontent,buttheydomakeavailableanextensivelistofÔnormalÕvariationthatoccursinthehumangenome.Thesedatabasesareimportantbecausetheyhelptocompletethepictureforanygeneorregionofinterest,bysummarizingalloftheneutralvariantsthataretypicallynotincludedincoredatabasesorLSDBs.CoredatabasesandLSDBsthussharethesameprimarypurposeofrepresentingDNAvariationsthathaveadeÞnitiveoraprobablephenotypiceffect.Thecurrentdatabasesare,however,toolimitedinnumberandintheirdegreeofinter-connectiontocapturealloftheinformationaboutpathogenicDNAmutations.Theisbecausethemodernresearchethosfailstoprovideadequateincentives(i.e.publicationoptions,peerrecog-nitionandfunding)toencourageresearcherstobuildnewcoredatabasesorLSDBs.Initiativesdesignedtomakeittechnicallysimpletosetupandusesuchdatabasesarewelcomed,suchasspecializedsoftwaresoftwareorinter-activeuserinterfaces(e.g.Genewindow;dow.nci.nih.gov),asarethosethatdirectlytransferdatafromclinicaldiagnosticslaboratoriesintothesedeposi-deposi-(http://dmudb.org/);buttheseinitiativeswillnotchangethefundamentalproblem.Instead,thebiomedicalcommunitymustÞrstappreciatetheover-whelmingneedforimprovedmutationdatabasesystemstobegintosolvethisproblem.Nationalmutationdatabases:anewtrendThespectrumofmutationsobservedforanygeneordiseasewilloftendifferbetweenpopulationgroups,andalsobetweendistinctethnicgroupswithinageographicalregion.Thisisanimportantextradimensiontoconsiderwhenbuildingmutationdepositories,anditisreßectedintheemergenceofseveralnewnationalmutationdata-bases(NMDBs)(NMDBs).NotonlydoNMDBshelptoelaboratethedemographichistoryofhumanpopulationgroups, Box1.Examplesofusefulmutationandrelateddatabases OnlineMendelianInheritanceinMan(OMIMAcomprehensive,authoritativeandtimelyknowledgebaseofhumangenesandgeneticdisorderscompiledtosupportresearchandeducationinhumangenomicsandthepracticeofclinicalgenetics.EachOMIMentryhasafull-textsummaryofageneticallydeterminedphenotypeand/orgeneandhasnumerouslinkstoothergeneticdatabasessuchasDNAandproteinsequence,references,generalandlocus-speciÞcmutationdatabases(nlm.nih.gov/entrez/query.fcgi?dbOMIManditoolHumanGeneMutationDatabase(HGMD)Adatabaserecordingvarioustypesofmutationwithinthecodingregionsofhumannucleargenescausinginheriteddisease.HGMDdoesnotusuallyincludemutationslackingobviousphenotypicconsequences.Dataarecollectedweeklybyacombinationofmanualandcomputerizedsearchprocedures(http://www.hgmd.orgSNPdatabasesRepositoriesofsinglenucleotidepolymorphisms(SNPs)andotherlocalizedvariationsacrossthegenome.ThemostcomprehensiveSNPdatabaseisdbSNPatNCBI(http://www.ncbi.nlm.nih.gov/);additionalinformationisavailableatothersitessuchastheHumanGenomeVariationDatabase()andtheInternationalHapMapProjectwebsite().SomeprojectshaveaspeciÞcresearchfocus,suchastheenvironmentalgenomeSNPdatabase(egSNP:niehs.nih.gov/egsnp/home.htm),whichlistscommonSNPsinselectedenvironmental-responsegenes.AnextensivelistoftheSNPdatabasesisprovidedathttp://www.genomic.unimelb.edu.au/mdi/dblist/dblist.htmlApubliclyavailableknowledgebasethatconsistsofacentralrepositoryforclinicalandgeneticinformationthataidsresearchersinunderstandinghowgeneticvariationamongindividualscontrib-utestodifferencesinreactionstodrugs(http://www.pharmgkb.org/Amulti-speciesgenotypeÐphenotypedatabase,mergingpublicgenotypeandphenotypedatafromawiderangeofhumanandothermodelorganisms.Itsuserinterfaceenablesscientiststocompareandbrowseknownphenotypessimultaneouslyforagivengeneorasetofgenesfromdifferentorganisms(http://www. TRENDSinGeneticsVol.21No.6June2005 theyarealsoaprerequisitetotheoptimizationofnationalDNAdiagnosticservices.Thatis,theywillprovideessentialreferenceinformationforuseinthedesignoftargetedmutation-detectioneffortsforclinicaluse,andtheymightalsoservetoenhanceawarenessamonghealthcareprofessionals,bio-scientists,patientsandthepublicabouttherangeofcommongeneticdisorders(andtheirenvironmentalcorrelates)sufferedbyparticularpopulationgroups.TwooftheÞrstonlineNMDBsareaFinnishdatabasehttp://www.Þndis.org[20]andanArabiandatabasehttp://www.agddb.org.agddb.org.Althoughrichininformation,theseparticularresourcesunfortunatelyprovidelimitedquerycapacity,particularlyforallelicfrequencies.Theywill,nevertheless,continuetodevelopuser-friendlydesigns,offeringgoodqueryingcapacityandextensiveexpertdatacuration.TheHellenicandCypriotNMDBs(availableathttp://www.goldenhelix.org/hellenichttp://www.goldenhelix.org/cypriot,respectively)areaim-inghigher,byintroducingaspecializeddatabasemanage-mentsoftware(ETHNOS)thatenablesbothcompoundqueryformulationandrestricted-accessdataentrysothatallrecordsaremanuallycuratedtoensuregoodandconsistentdataqualityquality.TomaximizetheutilityofNMDBs,thewayinwhichtheircontentisprovidedneedstoensureaseamlessintegrationwithrelatedcontentinLSDBsandcoredatabases.ThisisconceptuallyillustratedinFigure1Furthermore,extensivelinkstootherexternalinfor-mation(e.g.toOMIMandtovarioustypesofgenomesequenceannotation)wouldideallybeprovidedtoconnectNMDBstothegrowingnetworkofgenomicdatabases.GenotypeÐphenotypedatabases:alookatthefutureExcellentprogresshasbeenmadeinconstructingdata-basesofprimaryDNAinformation(i.e.genomesequenceandpolymorphism)and,asdescribedpreviously,therehasbeensomeprogressincreatingmutationdatabasestocatalogDNAalterationsthathaveaphenotypiceffect.However,thephenotypedataintheseresourcesarepresentedinabasicway,forexample,infreetextentriesand/orwithlittledetail.Thissituationneedstobeimproved,andthecomprehensiveanalysisofphenotypesisrequiredÐagoaltermedÔphenomicsÕÔphenomicsÕ.Infor-maticssolutionsthatsupportphenomicsmustbeBioinformaticsforthephenomicserawillhavetosolvenewproblems.AlthoughitisrelativelyeasytocreatedatabasesforÔuni-dimensionalÕDNAsequenceinformation(comprisingmerelyafour-lettercode),devisinggenericdataanddatabasemodelsfortheÔmulti-dimensionalÕboundlessuniverseofdiversephenotypesremainsachallenge.However,nomajorpublicdatabaseexiststhatcurrentlypresentsextensiveandsophisticatedgenotypeÐphenotypeconnections.OMIMremainsthebestcontenderinthiscategory,althoughthePhenomicDBisarecentandinnovativeproject,which,bymeansoforthologousgenerelationships,alignsOMIMdatawithmodelorganismdatainasingledatabase(http://www.phenomicDB.de[25].AlsoworthyofnoteisthePharmGKBproject,whichfocusesonpharmacogenetics(http://www.pharmgkb.org/[26Ð27]SowhatwillgenotypeÐphenotypedatabasesofthefutureneedtoachieve?Theymustaspiretobemuchmorethanmulti-diseaseLSDBsorcoremutationdatabases.Becausetheirphenotypedatacontentwillbesodiverse,computationalexploitationofthedatawillnotbepossiblewithanythingassimpleassequencesimilaritysearches.Instead,sophisticatedandintricatephenotypedatamodelswillberequiredtoempowercomputationalanalyses,andthesesolutionswillhavetomakerigoroususeofextensivephenotypeontologies.Ultimatelygeno-typeÐphenotypedatabaseswillneedtoprovidethefullrangeofÔomicsÕdata(e.g.transcriptomics,proteomicsandmetabolomics)thatmechanisticallyconnectgenotypedifferencestophenotypeconsequences.Initiallyhowever,suchanall-encompassingÔsystemsbiologyÕapproachmightbetoodaunting,andprojectswillprobablyÞrstconcentrateontyingDNAchangesdirectlytophenotypes.Achievingeventhis,giventheenormousamountofdatanowbeinggenerated,willdependonthecreationofsystematicandstandardizedwaystomanagephenotypedata,andthisbyitselfwillrequiregoodinternational Population IPopulation II BCDEFGHIMNO Central database Figure1.Relationshipsbetweenvarioustypesofmutationdatabases.Thisdepictionisborrowedfromelectroniccommerce,anditusestheconceptofanancienttempletoillustratethefruitfulsynthesisofcore,LSDBsandNMDBs.Coredatabasesrepresentmutationsfrommanygenesbutwithonlylimiteddetail(frequentlyreferredtoasÔmilewideandinchdeepÕdeepÕ)andsotheyaresymbolizedasthebroadfoundationsofthetemple.LSDBsprovideextensivedetailbutonlyforafewgenes(oftenreferredtoasÔinchwideandmiledeepÕdeepÕ),andhencetheyaresymbolizedasthetallandnarrowcolumnsofthetemple(AÐO).NMDBsprovidealayerontopofLSDBs(i.e.theroofofthetemple),inthattheyspecifypopulation-speciÞcdetailsforalterationsinmanydifferentgenes;theheightofeachroofisindicativeofthedepthofrecordedgeneticdiversityforagivenpopulationorethnicgroup. TRENDSinGeneticsVol.21No.6June2005335 cooperationandopendatasharingaswassofruitfulinthepublicefforttosequencethehumangenome.Principaldatacomponentswillbephenotypedescrip-tionsandinformationongenomeÐphenomerelationships.Therewillbemanysourcesofthesedata,includingresultsfromresearchintomendeliandiseases,datafromanimalmodelstudies,observationsfromgeneticassociationstudies,mutationÞndingsfrommoleculardiagnosticlaboratoriesandpatientdatafromclinicalinvestigations.Thewealthofdatapublishedinjournalswillneedtobeincorporatedintothesedatabasesorriskbeinglosttothemedical-informaticsfuture(althoughitiscurrentlyunclearhowthismightbeachieved).Certainly,publisherscoulddomoretoencourageauthorstosubmittheirÞndingstosuitableelectronicdatabases.Textminingsoftwareisanactiveareaofresearchresearch,butintheforeseeablefuturesuchtoolswillprobablyonlyprovidesufÞcientfunctionalitytohelpcuratorsmanuallyextractliteraturedata(e.g.byÞndingandÞlteringpublications),ratherthanaccomplishthistaskwithouthumaninter-vention.Forthesereasons,thephenotypechallengewefaceissigniÞcant,andperhapsaninternationalÔHumanPhenomeProjectÕProjectÕshouldnowbeorganizedasafollow-ontothesequencingphaseoftheHumanGenomeProject.Patientdatabases:takingphenotypedatabasesonestepTheconstructionofdepositorieswithphenotypeinfor-mationkeyedtomany(orevenall)individualsinapopulationcouldbeconsideredtheultimatephenotypedatabase.Certainly,whenwhole-genomesequencingbecomesroutineandpersonalizedmedicineiscommon,thenÔpatientdatabasesÕmightbesomethingwetakeforgranted.Butthatissomewayinthefuture.Inthemeantime,theÞrsteffortsinthatdirectionhavebeenlaunched,drivenbypopulation-wideepidemiologicalprojectsinitiatedinvariouscountries,suchasIceland,EstoniaandtheUK((andreferencestherein).Theseearlyendeavorsfocusmainlyoncommondisordersratherthanmendeliandisease,andprimarilyintendtocapturefunctionalrelationshipsbetweendiseasephenotypesandtheunderlyingpolymorphisms,genotypesandepigen-epigen-.Intime,theymustbemaderelevanttomonogenicdisorders,becausevariablepenetranceofevenmendeliandiseasecanbeascribedtomodiÞergenesandtogenomicvariationvariationthatpre-existedwithinthefound-ingpopulationofcontemporaryhumanshumans.Forexample,thesicklecellCd6(AT)andthethalassemiamajorCd39(CT)mutationsarefoundtoexistinÞveandninedifferenthaplotypebackgrounds,respectively,accountingformuchofthephenotypevariabilityobservedinaffectedindividualsindividuals.ThesamephenomenonisseenfortheIVSI-110(GA)andIVSI-6(T-thalassemiamutations((;G.P.Patrinos,unpublished).Thus,acompleteunderstandingofhumandiseasegeneticswillrequireco-analysisofmendelianandcommondiseasecausation.Patientdatabasesraiseparticularlycomplexethicalchallengesthatdemandcarefulattention.Primarily,theinclusionofclinicalandmoleculardataconnectedtospeciÞcindividualsmustbeperformedinawaythatensuresanonymity.Howbesttoachievethishasnotyetbeenestablished,butitiswidelyagreedthatstrictgovernanceframeworksmustbeestablishedtoaddressallconÞdentialityconcerns.Otherissuesthatneedtobeconsideredincludecopyrightandintellectual-propertyprotection,thenatureofinformedconsent,data-accessrights,inferentialrelationshipsandsoon.Therearenouniversallyagreedsolutionstotheseproblems,buttheseissuesmustberesolvedifpatientdatabasesandperson-alizedmedicinearetoadvancesubstantially.FutureconsiderationsAssummarizedabove,databaseactivitiesrelatingtoDNAanddiseasedataaredisastrouslydeÞcient.Theyarenotdoingservicetothewealthofdatabeinggeneratedortheinsightsintodiseasemechanismsthatcouldbegleanedifthisinformationwereproperlymanaged.ButasthesayinggoesÔnecessityisthemotherofinventionÕinventionÕ,andsothissituationwillhavetochange,andarguablyhasalreadybeguntodoso.Forexample,theHumanGenomeVariationSociety(http://www.hgvs.org/)hasbeenestab-lishedtosupportLSDBsandtherelatedcommunity,andaconsortiumofmodel-organism-databasegroups(www.gmod.org/)hascometogethertoworkonthephenotypedatabasechallengechallenge.TheHumanGenomeVariationDatabase(HGVbase:[11Ð12]isbeingevolvedintoagenotypeÐphenotypedatabase,andtothisendhasformedtheÔPhenofocusÕnetwork(http://www.phenofocus.net/)toprovideadiscus-sionforumandcontactlistofinterestedparties.Theseandotherprojectshaveeachprogressedtovaryingdegreestowardscreatingphenotypedatamodels.Forexample,theHGVbaseteamalongwiththeJapaneseBiologicalInformaticsConsortiumhavetogetherledamulti-groupinternationalefforttoproposeaglobalstandard-datamodelforgenomesequencepolymorphism[PolymorphismMarkupLanguage(PML);],andtheirpreliminarymodelforphenotypedataandforgeneticassociationÞndingswillsubsequentlybeofferedforincorporationintoPML.Onapurelytechnicallevel,oneofthemosturgentgoalsmustbethecreationofapowerfulmeansforcomputa-tionallyrepresentingphenotypedata.Anoptimaldesignforthiswouldbe:(i)Standardizedandwidelyaccepted:justastheFASTAformatmadeitpossibleforallDNAsequencedepositoriesandanalysistoolstointeracteasily,somustwidelyacceptedstandarddatamodelsandexchangeformatsbedevelopedforphenotypedata.(ii)Ontologybased:useofstandardtermsandmean-ings,asmadeavailablebytheopenbiologicalontologywebsite(),willbeimportantforeffectivedataintegrationandanalysis.(iii)Flexible:systemswillneedtomanagediversephenotypesthatmightconcernmolecular,organelle,cellular,organ-systemorwhole-organismfeatures,withanylevelofdetail.(iv)Scalable:useswillrangefromsmall-scalestudiestomajordatabases,butcoredatamodelsmustbeequallyuseableforthesedifferentapplications. TRENDSinGeneticsVol.21No.6June2005 (v)Adaptable:withtimeandimprovedknowledge,phenotypesevolve(e.g.therangeoftestsperformedforadiseasecanchange),implyinganeedforversioning.(vi)Notdependentonthegenome:thephenotypeshouldbeastand-alonecomponent,becauseourknowl-edgeofitsDNAbasiswillprobablyimprovewithtime,andthegenomesequenceisnotyetdeÞnitive.Beyondthetechnicalchallenges,itwillbeevenmoredifÞculttoovercometheproblemsassociatedwiththewaydatabaseresearchisorganized,motivatedandrewarded.Forexample,formingconsensusopinionsandtrulycommittedconsortiatocreatestandardsisnoteasyinthehighlycompetitiveworldofscience.Thismightexplaininpartwhyleadingbioinformaticsactivitiestodayareoftenconductedinlargespecializedcenters(e.g.theEuropeanBioinformaticsInstituteandtheUnitedStatesNationalCenterforBiotechnologyInfor-mation)wherethepoliticalinßuenceandcriticalmassissuchthatwhattheyproduceautomaticallybecomesthedefactostandard.Thesegroups,however,cannotbuildallofthenecessaryLSDBs,coredatabases,NMDBsandgenotypeÐphenotypedatabasesthatareneeded,buttheycouldhelpothers(biologicaldomainexperts)tobuildthemandthenintegratetheireffortsts.Thiskindofdistributedandcoordinatedeffortwouldalso,ideally,bemanagedinclosepartnershipwithspecializedjournalsjournalstoensurethatcontributorsalsohaveameanstopublishtheirefforts.ConcludingremarksThemostfundamentalhurdleofallthathindersprogressinthemutationdatabaseÞeldislimitedfunding.Becauseofthis,almostallmutationdatabasesthatcurrentlyexisthavebeenbuiltbyresearchersÔonthesideÕfortheirownuse,withasmalldegreeofcorporatesponsorshipatbest.ToadvancebeyondthisÔcottageindustryÕstateofaffairs,projectsneedtobeincreasedinscale,qualityanddurability,andthiscanonlyhappenifstrategicallymindedfundingagenciesmakeavailablesubstantialtargetedfunds.ThenewdatabasesthatemergewillthenneedtoÞndsupportforgeneralmaintenanceandongoingdevelopment.Toprovidethis,theprojectscouldperhapsberunasself-sustainingÔbusinessesÕthatchargethedatasuppliersorthedatausers(equivalenttohowscientiÞcjournalsworktoday).ItmightalsobepossibletodevelopnovelformsofjointacademicÐcorporatefunding.Thefundingchallengeisthusfarfromsimple:butitmighthelpthedebatetonotethatfundingagenciesinvestvastsumsofmoneytocreateprimarymutationdata,buttheythenfailtodirectsufÞcientfundstoensurethatthesedataßoweffectivelytocliniciansandscientistsinvolvedindiseaseresearchandpatientcare.Thissituationdeservestoberemedied.AcknowledgementsWethankHeikkiLehvaslaihoandRaymondDalgleishforcriticalreadingofthismanuscript.OurresearchissupportedbytheEuropeanCommission(FP6,ISTthematicarea)throughtheINFOBIOMEDNoE(IST-507585).1McKusick,V.A.(1966)MendelianInheritanceinMan.ACatalogofHumanGenesandGeneticDisorders,(1stedn.),JohnsHopkinsUniversityPress2Hamosh,A.etal.(2002)OnlineMendelianInheritanceinMan(OMIM),aknowledgebaseofhumangenesandgeneticdisorders.NucleicAcidsRes.30,52Ð553Cotton,R.G.etal.(1998)TheHUGOmutationdatabaseinitiative.279,10Ð114Stenson,P.D.etal.(2003)Humangenemutationdatabase(HGMD):2003update.Hum.Mutat.21,577Ð5815Beroud,C.(2005)Theuseofmutationdatabasesinmoleculardiagnostics.InMolecularDiagnostics(Patrinos,G.P.andAnsorge,W.,eds),Elsevier(inpress)6Hardison,R.C.etal.(2002)HbVar:Arelationaldatabaseofhumanhemoglobinvariantsandthalassemiamutationsattheglobingeneserver.Hum.Mutat.19,225Ð2337Patrinos,G.P.etal.(2004)ImprovementsintheHbVardatabaseofhumanhemoglobinvariantsandthalassemiamutationsforpopu-lationandsequencevariationstudies.NucleicAcidsRes.32(Databaseissue),D537ÐD5418Claustres,M.etal.(2002)TimeforauniÞedsystemofmutationdescriptionandreporting:areviewoflocus-speciÞcmutationGenomeRes.12,680Ð6889Wheeler,D.L.etal.(2004)DatabaseresourcesoftheNationalCenterforBiotechnologyInformation:update.NucleicAcidsRes.(Databaseissue),D35ÐD4010TheInternationalHapMapConsortium.(2003)TheinternationalHapMapproject.426,789Ð79611Fredman,D.etal.(2004)HGVbase:acuratedresourcedescribinghumanDNAvariationandphenotyperelationships.NucleicAcids32(Databaseissue),D516ÐD51912Fredman,D.etal.(2002)HGVbase:ahumansequencevariationdatabaseemphasizingdataqualityandabroadspectrumofdataNucleicAcidsRes.30,387Ð39113Marsh,S.etal.(2002)SNPdatabasesandpharmacogenetics:greatstart,butalongwaytogo.Hum.Mutat.20,174Ð17914Aerts,J.etal.(2002)DataminingofpublicSNPdatabasesfortheselectionofintragenicSNPs.Hum.Mutat.20,162Ð17315Dvornyk,V.etal.(2004)CurrentlimitationsofSNPdatafromthepublicdomainforstudiesofcomplexdisorders:atestfortencandidategenesforobesityandosteoporosis.BMCGenet.5,416Brown,A.F.andMcKie,M.A.(2000)MuStaRandothersoftwareforlocus-speciÞcmutationdatabases.Hum.Mutat.15,76Ð8517Patrinos,G.P.etal.(2005)HellenicNationalMutationDatabase:aprototypedatabaseformutationsleadingtoinheriteddisordersintheHellenicpopulation.Hum.Mutat.25,327Ð33318Patrinos,G.P.andWajcman,H.(2004)RecordinghumanglobingeneHemoglobin28,v-vii19Horaitis,O.andCotton,R.G.(2004)Thechallengeofdocumentingmutationacrossthegenome:thehumangenomevariationsocietyHum.Mutat.23,447Ð45220Sipila,K.andAula,P.(2002)DatabaseforthemutationsoftheFinnishdiseaseheritage.Hum.Mutat.19,16Ð2221Teebi,A.S.etal.(2002)Arabgeneticdiseasedatabase(AGDDB):apopulation-speciÞcclinicalandmutationdatabase.Hum.Mutat.22Gerlai,R.(2002)Phenomics:Þctionorthefuture?TrendsNeurosci.23Scriver,C.R.(2004)Afterthegenome-thephenome?J.Inherit.Metab.Dis.27,305Ð31724Hall,J.G.(2003)AclinicianÕsplea.Nat.Genet.33,440Ð44225Kahraman,A.etal.(2005)PhenomicDB:amulti-speciesgenotype/phenotypedatabaseforcomparativephenomics.26Licinio,J.(2004)PharmGKB:thepharmacogeneticsandpharmaco-genomicsknowledgebase.PharmacogenomicsJ.4,127Hewett,M.etal.(2002)PharmGKB:thepharmacogeneticsknowledgeNucleicAcidsRes.30,163Ð16528Zhou,G.etal.(2004)Recognizingnamesinbiomedicaltexts:amachinelearningapproach.20,1178Ð119029Chiang,J.H.etal.(2004)GIS:abiomedicaltext-miningsystemforgeneinformationdiscovery.20,120Ð121 TRENDSinGeneticsVol.21No.6June2005337 30Hirschman,L.etal.(2002)Accomplishmentsandchallengesinliteraturedataminingforbiology.18,1553Ð156131Freimer,N.andSabatti,C.(2003)Thehumanphenomeproject.Genet.34,15Ð2132Kaiser,J.(2002)Biobanks.Populationdatabasesboom,fromIcelandtotheU.S.Science298,1158Ð116133Bjornsson,H.T.etal.(2004)Anintegratedepigeneticandgeneticapproachtocommonhumandisease.TrendsGenet.20,350Ð35834Reich,D.E.andLander,E.S.(2001)Ontheallelicspectrumofhumandisease.TrendsGenet.17,502Ð51035Collins,A.etal.(1999)Geneticepidemiologyofsingle-nucleotidepolymorphisms.Proc.Natl.Acad.Sci.U.S.A.96,15173Ð1517736Lander,E.S.(1996)Thenewgenomics:globalviewsofbiology.Science274,536Ð53937Labie,D.etal.(1985)CommonhaplotypedependencyofhighGgamma-globingeneexpressionandhighHbFlevelsinbeta-thalassemiaandsicklecellanemiapatients.Proc.Natl.Acad.Sci.U.S.A.82,2111Ð211438Pirastu,M.etal.(1987)Thesamebeta-globingenemutationispresentonninedifferentbeta-thalassemiachromosomesinaSardinianpopulation.Proc.Natl.Acad.Sci.U.S.A.84,2882Ð288539Patrinos,G.P.etal.(2001)Agamma-haplotypes:anewgroupofgeneticmarkersforthalassemicmutationsinsidethe5regulatoryregionofthehumanAgamma-globingene.Am.J.Hematol.40Plato(360BC)TheRepublic41Stein,L.D.etal.(2002)Thegenericgenomebrowser:abuildingblockforamodelorganismsystemdatabase.GenomeRes.1599Ð161042Stein,L.(2002)Creatingabioinformaticsnation.Nature119Ð12043Anonymous.(1999)NewspaperandtheInternet:caughtintheweb.352,17Ð1944Scriver,C.R.etal.(2000)PAHdb:alocus-speciÞcknowledgebase.Hum.Mutat.15,99Ð104 ArticlesofinterestinTrendsCurrentOpinionDendritic-cell-basedtherapeuticvaccinationagainstcancerFrankO.Nestle,ArpadFarkasandCurdinConradCurrentOpinioninImmunology,17,163Ð169LessonsfromDNAmicroarrayanalysis:thegeneexpressionproÞleofbioÞlmsBethA.LazazzeraCurrentOpinioninMicrobiology,8,222Ð227WhenX-raysmodifytheproteinstructure:radiationdamageatworkOlivieroCarugoandKristinaDjinovic«CarugoTrendsinBiochemicalSciences,30,213Ð219Pharmaceuticals:athreattodrinkingwater?OliverA.Jones,JohnN.LesterandNickVoulvoulisTrendsinBiotechnology,23,163Ð167Environmentalmicrobiology-on-a-chipanditsfutureimpactsWen-TsoLiuandLiangZhuTrendsinBiotechnology,23,174Ð179Immuno-PCR:highsensitivitydetectionofproteinsbynucleicacidampliÞcationChristofM.Niemeyer,MichaelAdlerandRonWackerTrendsinBiotechnology,23,208Ð216TheaviangenomeuncoveredHansEllegrenTrendsinEcologyandEvolution,20,180Ð186TacklingthepopulationgeneticsofclonalandpartiallyclonalorganismsFabienHalkett,Jean-ChristopheSimonandFrancüoisBallouxTrendsinEcologyandEvolution,20,194Ð201MonogeniclowreninhypertensionMariaI.New,DavidS.Geller,FrancescoFalloandRobertC.WilsonTrendsinEndocrinologyandMetabolism,16,92Ð97GeneticsofhumanhypertensionAnupamAgarwal,GordonH.WilliamsandNaomiD.L.FisherTrendsinEndocrinologyandMetabolism,16,127Ð133MHCpolymorphism:AIDSsusceptibilityinnon-humanprimatesRonaldE.BontropandDavidI.WatkinsTrendsinImmunology,26,227Ð233Mechanismsofcelldeathinrhodopsinretinitispigmentosa:implicationsfortherapyHugoF.Mendes,JacquelinevanderSpuy,J.PaulChappleandMichaelE.CheethamTrendsinMolecularMedicine,11,177Ð185 TRENDSinGeneticsVol.21No.6June2005