doi:10.1093/nar/gkl254
TheMIGenASintegratedbioinformaticstoolkitforweb-basedsequenceanalysis
MarkusRampp*,ThomasSoddemannandHermannLederer
¨rPlasmaphysik,RechenzentrumGarchingderMax-Planck-Gesellschaft(RZG),amMax-Planck-Institutfu
Boltzmannstrasse2,85748Garching,Germany
ReceivedFebruary14,2006;RevisedMarch13,2006;AcceptedMarch30,2006
ABSTRACT
WedescribeaversatileandextensibleintegratedbioinformaticstoolkitfortheanalysisofbiologicalsequencesovertheInternet.Thewebportaloffersconvenientinteractiveaccesstoagrowingpoolofchainablebioinformaticssoftwaretoolsanddata-basesthatarecentrallyinstalledandmaintainedbytheRZG.Currently,supportedtaskscomprisesequencesimilaritysearchesinpublicoruser-supplieddatabases,computationandvalidationofmultiplesequencealignments,phylogeneticanalysisandprotein–structureprediction.Individualtoolscanbeseamlesslychainedintopipelinesallowingtheusertoconvenientlyprocesscomplexworkflowswithoutthenecessitytotakecareofanyformatcon-versionsortediousparsingofintermediateresults.ThetoolkitispartoftheMax-PlanckIntegratedGeneAnalysisSystem(MIGenAS)oftheMaxPlanckSocietyavailableatwww.migenas.org(click‘StartToolkit’).
cumbersome,ifnotimpracticable.Apartfromasmallsubsetofwell-definedapplicationswhicharewellsupportedbyexist-ingspecialpurposesoftware[e.g.theARBpackageforsequence-basedphylogeneticanalysis(1)]surprisinglyfewintegratedsoftwareenvironmentsformanagingsuchworkflowsofbasicanalysisstepsinaversatileanduser-friendlywayarepubliclyavailable.Existingclient–serverapplicationsmaybesubdividedintoclassicwebportals(2,3)and—emergingmorerecently—solutionsbasedonso-calledrichclientsforharvest-ingservicesanddatawhicharedispersedacrosstheInternet[cf.Ref.(4)foranexampleandrecentoverview;seealsowww.kepler-project.org].
Owingtoitsservice-orientedsoftwarearchitectureoursys-temcanservebothpurposes:whileinthisarticleweshallmainlyfocusonfunctionalitiesofferedbyapowerfulwebinterface,theMIGenASinfrastructurealsoprovidesSOAP-basedwebservicesthatcanbebeutilizedbythird-party(remote)clientapplications.
FEATURESANDFUNCTIONALITIES
TheMIGenASbioinformaticstoolkitisanewwebapplicationforprocessingbasicbioinformaticstasksaswellasorchestrat-ingthemintocomplexworkflowswithinasingle,coherentwebinterface.Targetusersareonlyassumedtobefamiliarwiththebasicfunctionalityofferedbythepopularsequenceanalysistools.Neitheradditionalcomputationalprerequisites(Amod-ernversionofoneofthepopularwebbrowsers,Mozilla/Firefox,OperaorInternetExplorerisrequiredwithJavaScriptenabled.)norin-depthbioinformaticsexperienceisconsideredtobenecessaryforworkingwiththetoolkit.ThesystemhasbeendevelopedwithsupportoftheMIGenASconsortiumoftheMax-Planck-Society.FoundingmembersaretheMax-Planck-Institute(MPI)ofBiochemistry(DepartmentofOesterhelt),MPIforComputerScience(DepartmentofLengauer),MPIforDevelopmentalBiology(DepartmentofLupas,andGroupS.C.Schuster:presentlyatPennsylvaniaStateUniver-sity,USA),MPIforMarineMicrobiology(DepartmentofAmann)andtheRZG.Servicesareprovidedandhostedby
INTRODUCTION
Alargepoolofindividualwebsitesofferingconvenientaccesstobasicbioinformaticssoftwareanddatahavecertainlygreatlyhelpedtoestablishmanycomputationalmethodsasstandardtoolsinlifesciences.Meanwhile,almostanynewlypublishedbioinformaticssoftwarepackagewhichisdistrib-utedforinstallationonPCsissupplementedbyawebserver(hostedbythesoftwaredevelopersand/orprovidedfordown-loadandlocalinstallation)inordertoenhanceusability,attractandguideusers,andtopromotevisibilityofthesoftwareinthescientificcommunity.NCBI’sBLASTservicesaretheproto-typicalexample.
Advancedanalysis,however,mostoftenrequiresthecon-certedinteroperationofdifferenttoolsandheterogeneousdata.ProcessingthecorrespondingworkflowsbyconsecutivelyvisitingwebsitesdispersedovertheInternetisapparentlyvery
*Towhomcorrespondenceshouldbeaddressed.Tel:+4908932992176;Fax:+4908032991301;Email:markus.rampp@rzg.mpg.deÓTheAuthor2006.PublishedbyOxfordUniversityPress.Allrightsreserved.
Theonlineversionofthisarticlehasbeenpublishedunderanopenaccessmodel.Usersareentitledtouse,reproduce,disseminate,ordisplaytheopenaccessversionofthisarticlefornon-commercialpurposesprovidedthat:theoriginalauthorshipisproperlyandfullyattributed;theJournalandOxfordUniversityPressareattributedastheoriginalplaceofpublicationwiththecorrectcitationdetailsgiven;ifanarticleissubsequentlyreproducedordisseminatednotinitsentiretybutonlyinpartorasaderivativeworkthismustbeclearlyindicated.Forcommercialre-use,pleasecontactjournals.permissions@oxfordjournals.org
W16NucleicAcidsResearch,2006,Vol.34,WebServerissue
theGarchingComputingCentreoftheMax-Planck-Society(RZG),whichmaintainsallsoftware,hardwareanddatarelatedtotheMIGenAStoolkit.Technology
Emphasishasbeenplacedondesigningascalableandextens-ible,object-orientedsoftwarearchitecture(basedontheJava2EnterpriseEditionplatform).Detailsaboutarchitecture,designandimplementationaredescribedinRef.(5).Withawebapplicationandwebservicesasthemainclientinter-facesabroadspectrumofusecasescanbecoveredrangingfrominteractive,web-basedworkflowprocessingtotheinteg-rationof(web)servicesintosophisticatedremoteapplications.Inordertoensureprivacyandsecurityforusersallcom-municationsarehandledviathehttpsprotocol.UponstartofanewsessionwiththeMIGenAStoolkit(viaanonymouslogin‘Guest’)theusergetsredirectedtothesecure(SSL/TLSencryption)httpscommunicationport.Thewebportal’siden-tityisauthenticatedbyacertificateissuedbytheMax-PlanckCertificateAuthority(http://ca.mpg.de/).Tools
Thewebapplicationsupportsthemaincategoriesofclassicbioinformaticstasks(seeTable1).Wehaveoptedforaman-ageableselectionofpackagesforeachfunctionalcategoryratherthanprovidingananonymouscollectionofalargenumberoftools.Packagesarecarefullyselectedaccordingtotheirperformance,circulationandcomputationalefficiency.Newtoolsarescheduledforintegrationonrequest.Databases
ForefficientaccessbytheMIGenASserverthefollowingFASTAnucleicandaminoacidsequencedatabasesaremirroredlocallyatRZGwithatleastaweeklyupdateinterval(linkstooriginalresourcesarestatedwithinparentheses):nr,env_nr,nt,sts,ESTs(www.ncbi.nih.gov),Swiss-Prot,TrEMBL(www.uniprot.org),PIR-NREF(pir.georgetown.edu),PDB(www.rcsb.org)andKEGGGENES(www.genome.ad.jp).Acompleteandup-to-datecollectionoforganism-specificFASTAdatabasesofthecompletedmicrobialgenomesfrom
Table1.OverviewoffunctioncategorieswithalltoolscurrentlysupportedbytheMIGenAStoolkitSequenceMultiplePhylogeny/Structuresimilaritysequenceclassificationpredictionsearch
alignmentNCBI-BLAST(6)ClustalW(7)PHYLIP(8)Arby(9)HHSearch(10)DIALIGN2(11)seqbootJNet(12)HMMer(13)MUSCLE(14)protdist,PsiPred(15)neighborPSI-BLAST(6)PCMA(16)consenseSignalP(17)HMMAccel
POA(18)drawgramTMHMM(19)T-Coffee(20)CLANS(21)
MODELLER(22)
Blammer,CluCheck
Anup-to-datelistoftools(anddatabases)withlinkstodetaileddocumentationismaintainedontheMIGenASwebportal.Thetoolsnamed‘HMMAccel’(for
performingacceleratedHMMersearches;Frickey&So
¨ding),‘Blammer’(foraligningBLASThitsequences;Frickey&Lupas)and‘CluCheck’(forauto-maticassessmentofalignmentquality;Frickey&Lupas)arenotyetpublished.
NCBIisavailabletogetherwithanumberofeukaryoticgenomes.ClusteredESTsequencesareprovidedasFASTAdatabasesforHomosapiens,MouseandDrosophila(http://genenest.molgen.mpg.de/).Inaddition,HMMlibrariesbasedonPfam-A(http://pfam.wustl.edu/)canbesearched.Upload-ingofuser-suppliedsequencedatabasesissupportedbythemajorityoftools.Such(private)dataarenotvisibleoutsideoftheuser’ssession.Basicuserinterface
Theessentialuserinteractionoccursinthelarge,centralpartofthewebportalwhichdisplaystheformspromptingtheuserforinputdataandparametersandrenderstheoutputofcom-pletedcomputations(Figures1and2).Thesetofsupportedtoolsisarrangedinahierarchicaltabbedstructure.Theusernavigatesbetweentoolsbyfirstselectingthetabwiththecorrespondingtoolcategoryandthenclickingtheparticulartool.Basiccontrolsforworkingwithatoolarelocatedinthenarrowhorizontalbarshownatthetopofthepage.Thiscontrolbarhostsanumberofpull-downmenuswhichallowtoswitchbetweendifferentrunswiththesametool(‘Runs’),tonavigatebetweeninputform,documentationandoutputdisplay(‘View’),toredirectresultstoothertools(‘Forward’)andtodownload(‘Export’)results.The‘submit’buttonneedstobeclickedforstartingcomputations(seeFigures1and2).Theuserprovidesprimaryinputdata(e.g.proteinsequencesandmultiplesequencealignments)tobeanalyzedbyeitherpastingoruploadingthedatainoneofthepopularformatsorbydirectlyselectingoutputfromaprecedingcomputationperformedwithinthetoolkit(seebelow).Tool-specificparameters,suchasE-valuecut-offs,databasestobesearchedandsoon,aredefinedbymakingselectionsinthecorrespondingformfieldswhicharelocatedbelowtheaforementionedinput-datafields(Figure1).Smallpop-up‘tooltips’withabriefexplanationofaspecificpara-meteraredisplayedwhentheuserhoversoverthecorrespond-inghyperlinkwiththemousepointer.Clickingthehyperlinkredirectstoamoredetaileddocumentationofthetoolanditsparameters.
Theparameterspaceofinterestcanbesystematicallyexploredbycreatinganew‘run’foreachrelevantcombinationofinputparametersforaparticulartool.Obtainedresultsmaybeforwardedtoanothertoolordownloadedindifferentformatstotheuser’sPCbymakingthecorrespondingselec-tionfromthepull-downmenunamed‘Forward’or‘Export’,respectively(seeFigure2).
Thenarrowverticalareaontheright-handsideoftheportalshowsastatusoverviewofcomputingtasksandfacilitatesquicknavigationtoallrunsperformedwithinasession.Theupperpartofthisareaisreservedforcreatingandman-agingpersistentprojects.Thisfeature,whichiscurrentlyavailableonlytoacoreusercommunityequippedwithper-sonalizedaccounts,willsoonbereleasedforpublicuse.Pipelining
Thenotionofa‘run’withatoolisthecentralconceptunder-lyingthepipeliningcapabilitiesofthisapplication:ifoutputdataoftoolAcan(inprinciple)beusedasinputforanothertoolB,allrunstheuserhasalreadyperformedwithtoolAareofferedasselectableinputfortoolB.Forexample,thetarget
NucleicAcidsResearch,2006,Vol.34,WebServerissueW17
Figure1.SelectionofinputdataandparametersformultiplesequencealignmentcomputationwiththeClustalWtool.InthisexamplethreeindependentsetsoftargetsequencesidentifiedbythreedifferentprecedingBLASTsearcheswillbesubjectedtomultiplesequencealignment.
sequencesfoundinarunwithasearchtoolsuchasBLASTcanbeimmediatelyusedasinputforanalignmenttoolsuchasClustalW(seeFigure1).Theabovementioned‘Forward’pull-downmenuwhichisdisplayedwheninspectingtoolresultsfacilitatestheforwardingofresultstoanothertoolforfurtherprocessing(Figure2).
Inadditiontosuchsemi-automaticworkflowmanagementwheretheuserinteractivelycoordinatesthesuccessionoftoolsitisalsopossibletopreconfigureacustom‘Meta’-tool(tab-group‘Pipelines’)asapipelineofindividualtoolsandintermediatefilters.Thesamepipelinecanthenbeemployedforconvenientlyprocessingdifferentsetsofinputdataandparameters.Forexample,suchatoolpipelinecouldstartbyasequencesimilaritysearchwiththetargetsequencesbeingfilteredaccordingtoachosenE-valuecut-off,subsequentlybeingsubjectedtomultiplealignment,automaticvalidationandfinallyphylogenetictree-building.Customizationofresults,dataintegration
Allrelevantresultsofcomputationsareinternallyinterpreted(‘parsed’)bytheserver.Thisisnotonlyafundamentalprerequisiteforthepipeliningcapabilitiesdescribedabovebutalsoallowsustoaddvaluetotherawresultsdeliveredbytheunderlyingsoftwarepackages.Figure2,forexample,showsacolor-codedversionofascoredmultiplesequencealignmentascomputedby‘ClustalW’togetherwitharulerforresidue-positionnumbers.Asanexampleforamoreadvancedfeaturewepointoutthecapabilityforcomprehens-iveandreliableannotationofsequencesbyspeciesandgenenames,proteinnamesaswellaspossiblesynonymsandacces-sioncodesinvarioussequencedatabases.ThisisbasedonthePIR-NREF(23)andUniProt(24)databases(sincerecently,PIR-NREFhasbeensupersededbyUniProt)andappliestoallsequenceswhichhavebeenextractedfromoneofthemajorproteinsequencedatabases.WealsoshowliteraturelinkstoPubMed(www.ncbi.nlm.nih.gov),whicharerelated(accord-ingtotheinformationprovidedbyPIR-NREF/UniProt)totheproteinunderconsideration.ThecompletetextofPubMedabstractsgetsasynchronouslyretrievedandisdisplayedinasmallframewhentheuserhoversthemousepointeroverthePubMedicon,whichisdisplayednextto,e.g.aBLASThit.Tasksfordisplayandpostprocessingofresults,whichrequireahigherdegreeofinteractivitythananHTML-based
W18NucleicAcidsResearch,2006,Vol.34,WebServerissue
Figure2.ResultofamultiplesequencealignmentcomputationwiththeClustalWtool.Thepulled-downmenunamed‘Forward’(topright)offersaselectionoftoolssuitableforsubsequentprocessingofthealignment.
webapplicationconceivablycanoffer,aredelegatedtoJavaApplets.Examplesaretheappletsnamed‘ATV’(25)fortreeviewing,‘JalView’(26)foreditingalignments,‘Jmol’(www.jmol.org)forrendering3Dproteinstructuresand‘CLANS’(21)forinteractivevisualizationofpairwisesequencesimilarities.Parallelprocessing
ThemajorityoftoolssupportedbytheMIGenAStoolkitallowparallelprocessingofmultiple,mutuallyindependentinputdata.Whenpastingoruploadingasetofproteinsequences,forexample,orselectingmultipleoutputfromaprecedingrunforfurtherprocessingwithanothertool,anewrunwiththistooliscreatedautomaticallyandexecutedinparallelforeachindividualinputwithonlyasinglestepofuserinteraction.SOAP-basedwebservices
Naturally,notallconceivablesortsofanalysisandpost-processingproceduresfortoolresultscanbeanticipatedandimplementedintoawebapplication.Inordertoallow
advanceduserstotakeadvantageofexistingMIGenASservices,yetexertmaximumcontrol(e.g.byembeddingthemintheirownscripts),programmaticaccesstoindividualtoolinterfacesisexportedintheformofSOAP-basedwebservices(cf.27).This,inparticular,allowsintegrationwithotherthird-partyremoteapplications[seeRef.(28)andref-erencescitedtherein].ExamplecodewritteninthePerlorJavaprogramminglanguageforanumberofwebserviceclientsoftheMIGenAStoolkitisdistributedonrequest.
FUTUREDIRECTIONS
DevelopmentoftheMIGenAStoolkitwhichweintroducedinthisarticlehasbeenuser-drivenfromthebeginning.Thefunc-tionalitiesofthetoolkitarecontinuallybeingupdatedandextendedinresponsetorequestsandsuggestions,whichareemergingfromthecoreusercommunityoftheMIGenASconsortium.Accordingtotheconsortium’soriginalfocusonmicrobialgenomeresearchthemajorityofstudiesconduc-tedsofarhasbeendealingwithmicrobialgenes.Althoughthetoolkitinprincipleisnotlimitedtothesetypesofanalysis,the
NucleicAcidsResearch,2006,Vol.34,WebServerissue
W19
currentselectionoftools,databasesandespeciallysupporteduse-casesisprobablyslightlybiased.
Accordingly,weplantoextendandgeneralizescopeandfunctionalityoftheserver,andwouldliketoencouragepro-spectiveuserstoprovideuswithfeedback,inparticularonusabilityofthesystemanddesirablenewfeatures.
Inaddition,acomprehensivesetofSOAP-basedwebserviceswithcorrespondingclientcodesandworkflowtoolswillbemadeavailableontheMIGenASwebportalinthenearfuture.
ACKNOWLEDGEMENTS
WeareindebtedtothemembersoftheMIGenASconsortiumforsharingtheirsoftwareandexpertisewithus.Ananonymousrefereeisgratefullyacknowledgedforvaluablecriticismwhichhelpedustoimprovetheusabilityofthesystem.FundingtopaytheOpenAccesspublicationchargesforthisarticlewasprovidedbytheMax-Planck-Society.Conflictofintereststatement.Nonedeclared.REFERENCES
1.Ludwig,W.,Strunk,O.,Westram,R.,Richter,L.,Meier,H.,Yadhukumar,Buchner,A.,Lai,T.,Steppi,S.,Jobb,G.etal.(2004)ARB:asoftwareenvironmentforsequencedata.NucleicAcidsRes.,32,1363–1371.2.Crass,T.,Antes,I.,Basekow,R.,Bork,P.,Buning,C.,Christensen,M.,Claussen,H.,Ebeling,C.,Ernst,P.,Gailus-Durner,V.etal.(2004)TheHelmholtzNetworkforBioinformatics:anintegrativewebportalforbioinformaticsresources.Bioinformatics,20,268–270.
3.Gracy,J.andChiche,L.(2005)PAT:aproteinanalysistoolkitforintegratedbiocomputingontheweb.NucleicAcidsRes.,33,W65–W71.
4.Navas-Delgado,I.,Rojano-Mun
˜oz,M.,Ramı´rez,S.,Pe´rez,A.,Andre´sLeo
´n,E.,Aldana-Montes,J.andTrelles,O.(2006)Intelligentclientforintegratingbioinformaticsservices.Bioinformatics,22,106–111.
5.Rampp,M.andSoddemann,T.(2005)Aworkflowengineformicrobialgenomeresearch.InKremer,K.andMacho,V.(eds),ForschungundWissenschaftlichesRechnen2004,Volume68ofGWDG-Reports
(ISSN0176-2516).Ges.fu
¨rwissenschaftlicheDatenverarbeitung,Go
¨ttingen,Germany,pp.17–46.6.Altschul,S.,Madden,T.,Schaffer,A.,Zhang,J.,Zhang,Z.,Miller,W.andLipman,D.J.(1997)GappedBLASTandPSI-BLAST:anewgenerationofproteindatabasesearchprograms.NucleicAcidsRes.,25,3389–3402.7.Chenna,R.,Sugawara,H.,Koike,T.,Lopez,R.,Gibson,T.J.,
Higgins,D.G.andThompson,J.D.(2003)MultiplesequencealignmentwiththeClustalseriesofprograms.NucleicAcidsRes.,31,3497–3500.8.Felsenstein,J.(1989)Phylip—phylogenyinferencepackage.Cladistics,5,164–166.
9.VonO
¨hsen,N.,Sommer,I.,Zimmer,R.andLengauer,T.(2004)Arby:automaticproteinstructurepredictionusingprofile–profilealignmentandconfidencemeasures.Bioinformatics,20,2228–2235.10.So
¨ding,J.(2004)ProteinhomologydetectionbyHMM–HMMcomparison.Bioinformatics,199,133–154.
11.Morgenstern,B.(1999)Dialign2:improvementofthesegment-to-segmentapproachtomultiplesequencealignment.Bioinformatics,15,211–218.
12.Cuff,J.andBarton,G.(1999)Applicationofenhancedmultiplesequence
alignmentprofilestoimproveproteinsecondarystructureprediction.Proteins,40,502–511.
13.Eddy,S.(1998)ProfilehiddenMarkovmodels.Bioinformatics,
14,755–763.
14.Edgar,R.C.(2004)MUSCLE:multiplesequencealignmentwithhigh
accuracyandhighthroughput.NucleicAcidsRes.,32,1792–1797.15.Jones,D.(1999)Proteinsecondarystructurepredictionbasedon
position-specificscoringmatrices.J.Mol.Biol.,292,195–202.16.Pei,J.,Sadreyev,R.andGrishin,N.(2003)PCMA:fastandaccurate
multiplesequencealignmentbasedonprofileconsistency.Bioinformatics,19,427–428.
17.Bendtsen,J.,Nielsen,H.,vonHeijne,G.andBrunak,S.(2004)Improved
predictionofsignalpeptides:Signalp3.J.Mol.Biol.,340,783–795.18.Lee,C.,Grasso,C.andSharlow,M.(2002)Multiplesequencealignment
usingpartialordergraphs.Bioinformatics,18,452–464.
19.Krogh,A.,Larsson,B.,Heijne,G.v.andSonnhammer,E.(2001)
PredictingtransmembraneproteintopologywithahiddenMarkovmodel:applicationtocompletegenomes.J.Mol.Biol.,305,567–580.20.Notredame,C.,Higgins,D.andHeringa,J.(2000)T-Coffee:anovel
methodformultiplesequencealignments.J.Mol.Biol.,302,205–217.21.Frickey,T.andLupas,A.(2004)CLANS:aJavaapplicationfor
visualizingproteinfamiliesbasedonpairwisesimilarity.Bioinformatics,20,3702–3704.
22.Fiser,A.andSali,A.(2003)Modeller:generationandrefinementof
homology-basedproteinstructuremodels.MethodsEnzymol.,374,461–491.
23.Wu,C.H.,Yeh,L.S.,Huang,H.,Arminski,L.,Castro-Alvear,J.,Chen,Y.,
Hu,Z.,Kourtesis,P.,Ledley,R.S.,Suzek,B.E.etal.(2003)TheProteinInformationResource.NucleicAcidsRes.,31,345–347.
24.Apweiler,R.,Bairoch,A.,Wu,C.H.,Barker,W.C.,Boeckmann,B.,
Ferro,S.,Gasteiger,E.,Huang,H.,Lopez,R.,Magrane,M.etal.(2005)UniProt:theuniversalproteinknowledgebase.NucleicAcidsRes.,32,D115–D119.
25.Zmasek,C.M.andEddy,S.R.(2001)ATV:displayandmanipulationof
annotatedphylogenetictrees.Bioinformatics,17,383–384.
26.Clamp,M.,Cuff,J.,Searle,S.M.andBarton,G.J.(2004)TheJalviewJava
alignmenteditor.Bioinformatics,12,426–427.
27.Pillai,S.,Silventoinen,V.,Kallio,K.,Senger,M.,Sobhany,S.,Tate,J.,
Velankar,S.,Golovin,A.,Henrick,K.,Rice,P.etal.(2005)SOAP-basedservicesprovidedbytheEuropeanBioinformaticsInstitute.NucleicAcidsRes.,33,W25–W28.
28.Oinn,T.,Addis,M.,Ferris,J.,Marvin,D.,Senger,M.,Greenwood,M.,
Carver,T.,Glover,K.,Pocock,M.R.,Wipat,A.etal.(2004)Taverna:atoolforthecompositionandenactmentofbioinformaticsworkflows.Bioinformatics,20,3045–3054.
因篇幅问题不能全部显示，请点此查看更多更全内容