4NeoplasticProcessincludesMeSHtermsreferringtocancers SemanticType IntermediateBMeSHTerms G E A Genesjun 1 Genesfos 2 GenesAPC 3 GenesReporter 4 GenesDominant 5 Genesras 6 Genesrel 7 Genesbc ID: 143521
Download Pdf The PPT/PDF document "Figure3:TemporalDistributionofCurcuminDo..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Figure3:TemporalDistributionofCurcuminDocumentstivelyrecent,beingpublishedin1990orlater.Thisin-dicatesasurgeinscienticinterestinthehealtheffectsofthisspice,whichhaslongbeenvaluedinAsiaforitsmedicinalproperties.WelimitedST-BtothethreesemantictypesGeneorGenome;Enzyme;andAminoAcid,PeptideorProtein.WerestrictedST-CtoDiseaseorSyndromeandNeo-plasticProcess4andsetM(theparameterspecifyingthenumberofBtermstoselect)to10.Thesesemantictypesareappropriatesincewearelookingforbiochemicalandgeneticconnectionsbetweenturmericandnoveldiseases.Table1showsthetop10selectedMeSHtermsfromeachST-Btype(step2).Wecanobservefromthetablethatsomeofthetermsappearinmorethanonesemantictype.(Thisispossiblesinceatermmaybeassignedtomorethan1semantictypeintheUMLS).However,weremoveduplicatesinstep2.Also,sometermsareveryspecicsuchasProteinKinaseCwhileothersarebroadrepresentingfamiliessuchasDNA-BindingProteinsandIsoenzymes.AtpresentwedonotdistinguishbetweenBtermsusingspecicity.Ourplanistoexaminethisaspectinfutureresearch.TheBtermslistedinTable1arethetoptentermsthatwereretrievedfromasearchoftheliteratureforthese-mantictypesGenesorGenomes,Enzymes,andAminoAcid,PeptideorProteins.Thebiochemicaleffectsofcur-cuminbecomeapparentuponconductingasearchoftheliteratureforcurcuminandanyoftheseterms.Curcumin,forexample,hasastrongdown-regulatoryeffectonc-JunNH2-terminalkinase(JNK)(14627502,12859962,11370761,12097302)resultinginthearrestofcellpro-liferation(14627502)inprostatetumorcells(12853969)andinductionofapoptosis(12859962).Curcuminin-hibitsNF-kappaB(12714587)leadingtothesuppres-sionofcellproliferationandtheinductionofapoptosis 4NeoplasticProcessincludesMeSHtermsreferringtocan-cers. SemanticType IntermediateBMeSHTerms G E A Genes,jun 1 Genes,fos 2 Genes,APC 3 Genes,Reporter 4 Genes,Dominant 5 Genes,ras 6 Genes,rel 7 Genes,bcl-2 8 NucleolusOrganizerRegion 9 Genes,myc 10 MAPK 1 3 GlutathioneTransferase 2 5 ProteinKinaseC 3 9 Prostaglandin-Endoperoxide Synthase 4 10 Isoenzymes 5 Protein-TyrosineKinase 6 Caspases 7 Nitric-OxideSynthase 8 OrnithineDecarboxylase 9 MAPKinaseSignalingSystem 10 NF-kappaB 1 TranscriptionFactorAP-1 2 Proto-OncogeneProteinsc-jun 4 TumorNecrosisFactor 6 Glutathione 7 DNA-BindingProteins 8 Table1:IntermediateBTerms.G:GeneorGenome,E:Enzyme,AAPP:AminoAcid,PeptideorProtein.Num-bersindicateranks. TerminalCMeSHTerms Rank Retina 1 Spinalcord 2 Testes 3 ThyroidNeoplasms 4 IschemicAttackTransient 5 Table2:TopFiveNovelCTerms.inmultiplemyeloma(12393461)andovarycancercells(12520734).TGF-beta1inducedIL-6whichhasbeenim-plicatedinthemalignantprogressionofprostatecancerswasseverelyimpededbycurcuminthroughinhibitionofc-Jun(matcheswithGenes,juninthetable)JNK(anin-stanceofMAPKinthetable)orAP-1(12853969).Thecurcuminopendiscoveryprocessterminatedwitharankedlistofdiseases.Table2showsthetop5entries5.Oneobservationmadeatthispointwasthatthetypeofautomatedsearchconductedinstep5ofthealgorithmtocheckfornoveltyisinsufcient.Atpresent,thesearchinvolvesonlytheparticularMeSHtermintersectedwiththeAtopic.Wedonotyetautomaticallyconsidersyn-onymsoftheMeSHterm.Forexampleforthelastentryinthetable,althoughIschemicAttackTransientAND(turmericORcurcuminORcurcuma)retrieved0docu-ments,thesearchIschemiaAND(turmericORcurcuminORcurcuma)retrieves17documents.Hencethisentryisunlikelytobeimmediatelyinterestingtotheuser.How-ever,thetoptwoentriesdidnotretrieveanydocumentevenaftersearchingwithdifferentsynonyms.Testesisalsounlikelytobeinterestingsinceacurcuminsearchin-tersectedwithspermretrievedmanydocuments.Consid-eringretrievalsetsizealoneisinsufcient.Forinstance,curcuminintersectedwiththyroidretrieved5documents.However,theseappeartobeperipheraltocurcumin'sef-fectonthyroidneoplasmsfocusingmoreonaspectssuchashypothyroidismandtoxicity.Automatingqueryex-pansionusingsynonymswillbethesubjectoffurtherre-search.Atthispointtheusermayselectentriesandperusetheappropriateliteraturefurtherto(a)determinethenatureoftherelationshipbetweencurcuminandthediseases(asthesubstanceunderstudycouldbebenecialorharmful)and(b)assessthequalityofthebackgroundknowledgethatmaybeusedtoguidefurtherstudyofcurcuminandthedisease.ThismanualphasemaybeguidedbythespecicBterm-basedpathwaysconnectingtheselected 5AlthoughthemainsemantictypeforatermsuchasSpinalCordisBodyPart,Organ,orOrganComponent,intheUMLSSpinalCordislistedforatleastonevocabularyasasynonymforSpinalCordDiseases.ItisthusalsoassignedthesemantictypeofDiseaseorSyndrome.SimilarobservationsholdfortermsRetinaandTestes MeSHTerm Genes,fos Genes,Reporter Genes,Dominant MAPK GlutathioneTransferase ProteinKinaseC Isoenzymes Protein-TyrosineKinase Caspases Nitric-OxideSynthase OrnithineDecarboxylase Proto-OncogeneProteinsc-jun TumorNecrosisFactor Glutathione DNA-BindingProteins Table3:BTermsConnectingTurmericandRetina.diseasewithcurcumin.Table3liststheBtermsthatwereautomaticallyidentiedasconnectingcurcuminand`Retina'.Inthenextsectionwepresentsuchananalysisfor`Retina'.Thatis,we(thesecondauthor)examinetheliteraturetodetermineifretinaldiseasesmaybeagoodcontextinwhichabioscientistmaystudycurcumin.Ouranalysisindicatesthatindeedthereisgoodevidencesup-portingthehypothesisofabenecialroleforturmericinthecontextofdiabeticretinopathies,ocularinammationandglaucoma.Analysisoftheotherhighly-rankeddis-easesisleftforfuturework.4Turmeric-RetinalDiseasesConnectionTheprocedurefolloweduptothispointis`term-centric'.Thatis,weautomaticallyidentifystatisticallyinterestingBtermsandthengeneratearankedlistofCterms.Wenowpresentfurtheranalysisontheconnectionbetweenretinaldiseasesandcurcumin.Insomecasesreadingthetitleandabstractsofselectrecordsprovidedsufcientin-formation.Inadditionthefulltextofthedocumentwasavailable.Ourstrategywastoexaminepublicationsforbiochemicalormolecularbiologymechanisms.Inpar-ticular,wewereinterestedinascertainingwhetheranyofthegenesnotedearlierwerealsoinvolvedinthepatho-physiologyoftheseretinaldisorders.Wefocusedonthegenesasthecriticallinksthatconnecttheagentcurcumintothedisorders.Analysis:Theuser'sgoalistoidentifybiochemicalpathwayspotentiallyconnectingretinaldiseasesandcur-cumin.Retinaldiseasescouldresultfromcomplicationsduetodiabetes,orofinfectionandinammationoftheretina. withtheideathatsimilarevidencemaybefoundforreti-naldiseases.Insummaryitseemslikelythatcurcumin,takeninthedietorappliedtopically,couldprovebene-cialincasesofdiabeticretinopathies,retinalinjury,ocu-larinammationandglaucoma.5RelatedResearchTextmining,i.e.,uncoveringinformationthatmayleadtohypotheses,hasattractedtheattentionofmanyre-searchers(eg.Andrade&Valencia,1998;Gordon&Lindsay,1996;Masysetal.,2001;Smalheiser&Swanson,1996a;Smalheiser&Swanson,1996b;Srini-vasan&Wedemeyer2003;Srinivasan,2004;Swanson,1986;Swanson,1988;Swansonetal.,2001;Weeber,2000).Examplesofrecenttextminingapplicationsin-cludeautomaticallyidentifyingvirusesthatmaybeusedasbioweapons(Swansonetal.,2001),proposingther-apeuticusesforthalidomide(Weeber,2003)andnd-ingfunctionalconnectionsbetweengenes(Chaussabel&Sher,2002;Shatkayetal.,2000).Amajoremphasisintextminingresearchhasbeentodirectlyexploitco-occurrencerelationshipsinMED-LINE.Forexample,Jenssenetal.,(2001)generateaco-occurrencebasedgenenetworkcalledPubGenefromMEDLINEfor13;712namedhumangenes.EachofPubGene's139,756linksisweightedbythenumberoftimesthegenesco-occur.WilkinsonandHuber-man6identifycommunitiesofgenes.Startingwithaco-occurrencebasedgenenetworkforaparticulardis-easedomain,communitiesareidentiedbyrepeatedlyre-movingedgesofhighestbetweeness(numberofshortestpathstraversingtheedge).Applyingthistothedomainofcolorectalcancer,theyareabletoidentifyinterestinghypotheseslinkinggenesthatwereforexample,inthesamecommunitybuthadnoedgebetweenthem.OurresearchisbasedontheopendiscoveryframeworkproposedbySwanson.Asindicatedbefore,SwansonandSmalheisermadeseveraldiscoveriesusingtheiropenandcloseddiscoverymethods(Swanson,1986;Swan-son,1988;Swansonetal.,2001;Smalheiser&Swan-son,1996a;Smalheiser&Swanson,1996b),thatwerelatervalidatedbybioscientists.Thesediscoveriesto-getherofferatestbedofexamplesthatarebeingusedbyotherresearcherstodeveloptheirowndiscoveryal-gorithms(Gordon&Lindsay,1996;Lindsay&Gordon,1999;Srinivasan,2004;Weeberetal.,2001).Onecharacteristicthatmaybeusefulindistinguish-ingbetweentextminingeffortsistheextenttowhichtheyareproblemorsubdomainspecic.Forexample,PubGeneisdirectlytargetedtowardsbioinformaticsre- 6Wilkinson,D.,&Huberman,B.A.Amethodforndingcommunitiesofrelatedgenes.http://citeseer.nj.nec.com/546592.html.searchers.Incontrast,implementationssuchasoursthatderivefromtheopendiscoveryframeworkarenotprob-lemspecic.Thesemaybeusedforavarietyofgoals,asforexamplebygeneticistsinvolvedinunderstandingtheresultsofmicroarrayexperimentsandbyepidemiologistssearchingforlinksbetweenvirusesandspecicpopula-tions.Webelievethatthenextgenerationoftextminingsystemswillbejudgednotonlybytheireffectivenessbutalsobytheirexibilityinapplication.6ConclusionsWeappliedourimplementationofSwanson'sopendis-coveryalgorithmtotheproblemofidentifyingnoveldis-easeorproblemcontextsinwhichsubstancesmighthaveatherapeuticrole.WeusedourmethodstoinvestigatethepotentialofturmericorCurcuminLonga.Ouranalysisidentiesarankedlistofproblemsforwhichtreatmentwithcurcuminmaybebenecialwiththetoprankeden-trypointingtoretinaldiseases.Guidedbyouralgorithm,furtheranalysisoftheliteraturebyourexpertuser(age-neticist)yieldedgoodevidenceinsupportofthehypoth-esisthatcurcumin,takeninthedietorappliedtopically,couldprovebenecialincasesofdiabeticretinopathies,ocularinammationandglaucoma.Infutureworkwewillanalyzetheothersuggestionsmadebyouropendiscoverymethods.Forexample,thesecondsuggestionisproblemsrelatedtothespinalcord.Theanalysiswillagainfocusongeneticmechanismsthatcouldpotentiallyconnectcurcuminwiththeproblems.Wewillalsoexploremethodstoautomatequeryexpan-sionforthesearchinstep5ofthealgorithm.Onelim-itationofthediscoveryprocessconcernstheevidencegatheringphasewhenanalyzingindividualCterms.Thisprocessismanualandinvolvessignicantinvestmentoftimeandintellecttowardsiftingthroughtheliteratureandcollectingevidencerelevanttothehypothesizedconnec-tions.Inthenextphaseofourworkweplantostudymethodstoassistinthisphase.AcknowledgmentsThisresearchwaspartlyaccomplishedwhiletherstau-thorwasavisitingfacultyscholarattheNationalLibraryofMedicine,Bethesda,Maryland.ShethankstheUniver-sityofIowafortheFacultyScholarAwardandNLMfortheirhospitality.ThisworkwasalsofundedinpartbyNSFgrantno.IIS-0312356awardedtotherstauthor.Theauthorsthanktheanonymousreviewersfortheirin-sightfulcomments.ReferencesAndradeA,&ValenciaA.1998.Automaticextractionofkeywordsfromscientictext:applicationtothe