doi:10.1093/bioinformatics/btl571
SAGA:asubgraphmatchingtool
matching.Gra l(Yanetal.,2005),PIS(Yanetal.,2006)andClosure-Tree(HeandSingh,2006)introducesomeapproximationforsubgraphisomorphism.However,theseapproximatemodelsareverylimited.Noneofthesetoolsallownodegapsintheirmodels.PathAligner(ChenandHofestaedt,2004)isatoolforaligningpathways.However,itassumesthatallpathwaysarelinearpaths.ThetoolsmostcloselyrelatedtoourworkarePathBlast(Kelleyetal.,2004)andthesuccessiveNetworkBlast(Sharanetal.,2005),whicharedesignedforaligningproteininteractionnetworks.Theirgraphsimilaritymodelallowsnodemismatchesandnodegaps,butgraphstructuraldifferencesarelargelycon nedtoshortpaths.AsshowninSection3.5,ourgraphsimilaritymodeltoleratesmoregeneralstructuraldifferences,andcan ndbiologi-callyrelevantmatches,whenbothPathBlastandNetworkBlastfail.Anotherrelatedmethod(Koyuturketal.,2005)hasbeenproposedforaligningproteininteractionnetworks.However,thematchtech-niqueusedinthismethodlargelyfocusesoncapturingthepenaltyassociatedwithgeneduplication.Finally,PathBlast,NetworkBlastandthemethodproposedinKoyuturketal.(2005)canonlyperformonegraphcomparisonatatime.Tomatchaqueryagainstadatabaseofgraphs,thematchingalgorithmsmustberunforeachgraphinthedatabase.Asaresult,thesemethodsarenotcomputationallyef cientwhenqueryinglargegraphdatabases.
Inthispaper,wepresentanovelapproximatesubgraphmatchingtechniquecalledSAGA.AttheheartofSAGAisa exiblemodelforcomputinggraphsimilarity,whichpermitsnodegaps,nodemismatchesandgraphstructuraldifferences.Tospeeduptheexecutionofquerieswiththispowerfulmatchingmodel,weemployanindexingmethodforef cientqueryevaluation.Throughexperi-mentalevaluation,wedemonstratethatSAGAismore exibleandpowerfulthanexistingmodels.SAGAallowsadditionalinforma-tionderivedfromtherelationshipsbetweenentitiesinpathwaystobeincorporatedintocomparativeanalysis.OurexperimentalresultsshowthatSAGA ndsexpectedassociations,likeInsulinsignalinginType2DiabetesMellitus(T2DM).SAGAalso ndslesswellstudiedassociations,liketheToll-likereceptor,T-cellreceptorandApoptosispathwaysinHelicobacterpylori(H.pylori)infection,aswellasCalcium,WntandHedgehogsignalinginBipolarDis-order.Inaddition,SAGAprovidesapowerfultoolforbiomedicaltextcomparison.
Fig.1.(a)Anexamplegraph.(b)Anexamplesubgraphmatch.
2.2Distancemeasureforsubgraphmatching
Ourmodelmeasuressimilaritybyadistancevalue,sographsthataremoresimilarhaveasmallerdistance.Formally,thesubgraphmatchingisde nedasfollows:LetG1¼(V1,E1,f1)andG2¼(V2,E2,f2)betwographs.AnapproximatematchingfromG1(thequery)toG2(thetarget)isabijection
^1$V^2,whereV^1jV1andV^2jV2.mappingfunctionl:V
AnexamplematchisshowninFigure1.Thedashedlinesindicatethematchednodesinthetwographs.Notethatnodescanbemappedeveniftheyhavedifferentlabels.Also,notethatnotallnodesarerequiredtobemapped,e.g.v5inG1hasnomappinginG2,andisagapnode.
Thesubgraphdistance(SGD),withrespecttol,isde nedas:
SGDlðG1 G2Þ¼we·StructDistl
þwn·NodeMismatcheslþwg·NodeGapsl
where
StructDistl¼
X
1 u<vu v2V
ð1Þ
jdG1ðu vÞÀdG2ðlu lvÞjX
1u2V
ð2Þð3Þð4Þ
NodeMismatchesl¼mismatchðf1ðuÞ f2ðluÞÞX
gapG1ðuÞ
NodeGapsl¼
1u2V1ÀV
22.1
SYSTEMANDMETHODSGraphmodel
Thedistancemodelcontainsthreecomponents.TheStructDistcompo-nentmeasuresthestructuraldifferencesofthematch,theNodeMismatches
componentisthepenaltyassociatedwithmatchingtwonodeswithdifferentlabels,andtheNodeGapscomponentisusedtomeasurethepenaltyforthegapnodes.(Gapnodesarenodesinthequerythatcannotbemappedtoanynodesinthetargetgraph.)Eachofthesecomponentsisdescribedinmoredetailinsubsections2.2.1through2.2.3.
InEquation(1),we,wnandwgaretheweightsforeachcomponentinthismatchingmodel,andcanbeusedtochangetheemphasisonthedifferentpartsofthesimilaritymodel.WhileEquation(1)computesthesubgraphdistanceforaspeci cmatchingl,theactualsubgraphdistancefromaquerytoitstargetistheminimumdistanceoverallpossiblematchings,namely:
SGDðG1 G2Þ¼minSGDlðG1 G2Þ
l
Inourmodel,agraph,G,isa3-tupleG¼(V,E,f).VisthesetofnodesandE V·Visthesetof(directedorundirected)edges.Nodesinthegraphshavelabelsspeci edbythemappingf:V!L,whereListhesetofnodelabels.Thismodelcapturesthefeaturesthatarecommonlypresentinmostbiologicalgraphdatasets,inwhichnodesrepresentmolecules/complexes,labelsdenotemolecule/complexnamesandedgesindicaterelationshipsbetweennodes.WeassumethateachnodeinthegraphhasauniqueID.ThisIDisusedtoestablishatotalorderamongthenodes.
IntheexamplegraphinFigure1(a),viisusedtorepresenttheuniquenodeIDandLkisthenodelabel.Notethattwodifferentnodesinagraphcanhavethesamelabel.
Ourdistancemodelandmatchingalgorithm(discussedbelow)supportbothdirectedandundirectedgraphs.Wepresentourmethodusingundirectedgraphs;adaptationsofthedistancemeasureandthematchingalgorithmfordirectedgraphsarestraightforwardandomittedhere.
ð5Þ
TheStructDistcomponentmea-suresthestructuraldifferencesforthematchingnodepairsinthetwographs.InEquation2,thedGi(u,v)functionmeasuresthe‘distance’betweennodeuandnodevingraphGi,andisde nedasthelengthoftheshortestpathbetweenuandv.TheStructDistcomponentcomparesthedistancebetweeneachpairofmatchednodesinonegraphtothedistancebetweenthecorrespondingnodesintheothergraph,andaccumulatesthedifferences.TheNodeMismatchescom-ponentinEquation(3)isthesumofthepenalties(quanti edbythemismatchfunction)associatedwithmatchingnodeswithdifferentlabels.
Acommonandbiologicallyintuitivemismatchpenaltymodelistoimplicitlygroupnodelabelsbasedonsimilarity,allowingfora


node
2.2.1TheStructDistcomponent
2.2.2TheNodeMismatchescomponent
233
ren Brunank , Bioinformatics:The Machine Learning Approach, A Bradford Book, The MIT press ,Cambridge, Massachusetts,London,England ,1998 [6] Eric ...
49页 免费 Bioinformatics 06_swis... 30页 免费 bioinformatics-1 暂无评价 74页 2下载券 bioinformatics softwar... 13页 免费喜欢此文档的还喜欢...
Bioinformatics 65页 免费bioinformatics-3 暂无评价0人阅读0次下载举报文档§2.3 核酸...
Bioinformatics_lect_1_理学_高等教育_教育专区。生物信息学 pptBioinformatics 主讲人: 陈海峰 主讲人 Shanghai Jiaotong University 2011-11-13 1 Contact ...
bioinformatics-4_农学_高等教育_教育专区。生物信息学§2.4 核酸序列的多序列比对 目的: ? 发现多个序列的共性 ? 发现与结构和功能相关的保守序列片段设:有k...
免费 Bioinformatics 09_blast(... 暂无评价 50页 免费 Bioinformatics 03_genban... 暂无评价 46页 免费 Bioinformatics 04_DNA co... 暂无评价 49页 免费...
bioinformatics-4 隐藏>> §2.4 核酸序列的多序列比对 目的: ? 发现多个序列的共性 ? 发现与结构和功能相关的保守序列片段设:有k个序列s1, s2, ... ,...
bioinformatics.btu635_生物学_自然科学_专业资料。宏基因功能分析中特征选择和比较的二级统计程序Bioinformatics Advance Access published September 24, 2014 Genome ...
BIOINFORMATICS APPLICATIONS NOTE Genetics and population analysis Vol. 26 no. 23 2010, pages 2990–2992 doi:10.1093/bioinformatics/btq565 Advance Access ...
Bioinformatics Advance Access published March 21, 2006 BIOINFORMATICS BicAT: A Biclustering Analysis Toolbox Simon Barkow a? Stefan Bleuler a , Amela Preli?...
热门文档