windows form application in c# with database pdf: Add hyperlink to pdf online SDK application service wpf azure web page dnn 61.full0-part1940

Combining literature text mining with

microarray data: advances for system

biology modeling

AlbertoFaro,DanielaGiordanoandConcettoSpampinato

Abstract

Ahugeamountofimportantbiomedicalinformationishiddeninthebulkofresearcharticlesinbiomedicalfields.

Atthesametime,thepublicationofdatabasesofbiologicalinformationandofexperimentaldatasetsgeneratedby

high-throughputmethods is in greatexpansion, and a wealth h of f annotatedgene databases, chemical, genomic

(includingmicroarraydatasets),clinicalandothertypesofdatarepositoriesarenowavailableontheWeb.Thus

acurrentchallengeofbioinformaticsistodeveloptargetedmethodsandtoolsthatintegratescientificliterature,

biologicaldatabasesandexperimentaldataforreducingthetimeofdatabasecurationandforaccessingevidence,

eitherintheliteratureorinthedatasets,usefulfortheanalysisathand.Underthisscenario,thisarticlereviews

theknowledgediscoverysystemsthatfuseinformationfromtheliterature,gatheredbytextmining,withmicro-

arraydataforenrichingthelistsofdownandupregulatedgeneswithelementsforbiologicalunderstandingandfor

generatingandvalidatingnewbiologicalhypothesis.Finally,aneasytouseandfreelyaccessibletool,GeneWizard,

thatexploitstextminingandmicroarraydatafusionforsupportingresearchersindiscoveringgene^diseaserela-

tionshipsisdescribed.

Keywords:literaturetextmining;microarraydata;biologicaldatabases;knowledgediscovery

INTRODUCTION

Ahugeamountofbiomedicalinformationishidden

in millionsofresearcharticlespublishedinthelast

20 years and this quantity is bound to increase

exponentially[1].Similarly,thepublicationbiologi-

cal databasesisingreatexpansion, anda wealth of

annotatedgenedatabases,chemical,genomic,clinic-

al and other types of data repositories, including

drugs and microarrayexperiments are available on

theWeb.Thusatopicalchallengeofbioinformatics

is to leverage on the combination of multi-type

informationsources,foramoreeffectivesystembiol-

ogymodelingandknowledgediscovery[2,3].Afirst

importantsteptowardstheorganizationandintegra-

tion of multi-type biomedical information is the

National Center for Biotechnology Information’s

(NCBI)EntrezCross-Database[4]thatinterconnects

PubMedabstractswith NCBI’sdatabasesonDNA

sequenceandchemicalstructure, thusspeedingup

the research of data related to a given disease.

However, this system does not include disease–

gene or disease–protein compendia and it is not

Corresponding author. Concetto o Spampinato, Department of Informatics and Telecommunication Engineering – University of

Catania, Viale Andrea Doria, 6 – 95127 – Catania, Italy. Tel: : þ39 (0) 95 7382372; ; Fax: þ39 (0) 95 7382397; E-mail:

cspampin@diit.unict.it

AlbertoFaroisFullProfessorofArtificialIntelligenceattheEngineeringFacultyoftheUniversityofCatania,Italy,whereheisalso

theDeanoftheComputerEngineeringdegree.Hiscurrentresearchinterestsinclude:dynamicsystemstheoryofcognition,intelligent

learningenvironments,computervisionandmobilecomputing.

DanielaGiordanoholdstheLaureadegreeinElectronicEngineering,grade110/110 cumlaude,fromtheUniversityofCatania,

Italy(1990),andaPhDinEducationalTechnologyfromConcordiaUniversity,Montreal(1998).Since2001sheisAssociateProfessor

ofInformationSystemsoftheDIEEIDepartment,EngineeringFacultyoftheUniversityofCatania,wheresheteachesthegraduate

levelcourse‘CognitiveSystemsandHuman-ComputerInteraction’.Herresearchactivityhasdevelopedalongthefollowingtracks:

(i) Knowledge Management; (ii) Data a Mining, , information n retrieval l and d visualization; (iii) ) Image and signal l processing g with

soft-computingtechniques;and(iv)Advancedlearningtechnologies.

ConcettoSpampinatoreceivedtheLaureainComputerEngineeringin2004,grade110/110cumlaude,andthePhDin2008from

theUniversity ofCatania,whereheiscurrently ResearchAssistant.Hisresearchinterests includeimageandsignal processing for

environmentalapplications,biomedicalimageprocessing,multimediaretrievalandbioinformatics.

BRIEFINGS IN N BIOINFORMATICS. . VOL13.NO O 1. 61^82

doi:10.1093/bib/bbr018

Advance Access s published on 15 June 2011

TheAuthor2011.PublishedbyOxfordUniversityPress.ForPermissions,pleaseemail:journals.permissions@oup.com

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

Add hyperlink to pdf online - insert, remove PDF links in C#.net, ASP.NET, MVC, Ajax, WinForms, WPF

Free C# example code is offered for users to edit PDF document hyperlink (url), like inserting and deleting

adding a link to a pdf; add hyperlink to pdf in

Add hyperlink to pdf online - VB.NET PDF url edit library: insert, remove PDF links in vb.net, ASP.NET, MVC, Ajax, WinForms, WPF

Help to Insert a Hyperlink to Specified PDF Document Page

adding links to pdf in preview; add hyperlink to pdf in preview

capabletolinkitsresultstoexternaldatabases, e.g.

drugsdatabases. Thislinkageisessentialtosupport

the cross-linking of textual information with the

relevant biological databases and to reinforce the

connectionbetweenannotationsinbiologicaldata-

bases. Moreover, given the fast development of

high-throughput methods with h consequent release

of experimental l data, the need of f developing tar-

getedbioinformaticstoolsandmethodsthatcombine

literature,biologicaldatabasesandexperimentaldata

for reducingthe time ofdatabase curation and for

knowledge discovery in literature is heavily

demanded [5]. . Under this scenario, , a a first aim of

thisarticleisreviewingtheknowledgediscoverysys-

temsthat integrate literature information, gathered

bytextmining, with microarraydata. Usually,this

isperformed either withthe goal of enrichingthe

listsofdownandup-regulatedgeneswithelements

for biological understandingor for generating and

validating new biological hypothesis. To illustrate

how text mining g and microarray data integration

maybeachieved,wefirstprovide,inthenextsec-

tion,anoverviewofthemethodsandtoolsthatare

used to perform text mining, i.e. information

retrieval(IR),namedentityrecognition(NER),in-

formation extraction(IE)andknowledge discovery

(KD), using as a starting point previous reviews

[6–9],andfocusingontheaspectsoftheintegration

betweenbiologicaldataandtextdatathathavebeen

recently investigated. . This overview allows us, , in

‘Combining text and microarray data’ ’ section, to

pointoutthedifferencesamongthecurrentattempts

atintegratingexperimentaldataintheminingloop.

Finally, in ‘GeneWizard’ section we illustrate a

new tool,GeneWizardthatusesmicroarraydatato

evaluate and validate biological hypothesis mined

fromtext.GeneWizard,basedonthemethodspro-

posedbyFaroetal.[10,11],proposesnovelrelation-

ships between genes and diseases by integrating

literature discoveries and gene sets gathered from

microarray data analysis. . In detail, starting from m a

gene–disease relationship, itextracts a set of genes

(related to the gene of the derived association)

involved in the disease. Biological functions,

namely, biological processes, cellular components

andmolecularfunctions, arethenassociatedtothe

validated set of genes by using Gene Ontology

(http://www.geneontology.org/) (GO). Finally, in

the conclusions s we outline the key challenges s to

advancingthistypeofintegratedknowledgediscov-

erysystems.

TOOLSANDMETHODSFOR

LITERATURETEXTMINING

Theprimarygoalofliteraturetextmining[12]isto

distillknowledgethatishiddenintextofpublished

papersandtopresentittotheusersinacoherentand

concise form. More formally, , the e ultimate goal of

textminingconcerns thediscoveryofnew, previ-

ouslyunknown information, byautomatictextre-

sourcesprocessing. Generally, systemsfor literature

textminingincludefourmainmodules[13]:(i)IR

togatherrelevanttextbyqueryingdatabasesofbio-

medicalpapers;(ii)NERtofindthebiologicalenti-

ties (e.g. genes, , proteins) within text; (iii) IE to

identify predefined relationships among biological

entities from explicit statements in text; and

(iv)KDtoelicitrelationshipshiddenintheinforma-

tion derived by the previous module. Recent

text-miningsystemshavestartedtakingintoconsid-

eration the integration betweenliterature and bio-

logical,chemical,medicalanddrugsdatabases.Inthe

nextsectionseachmoduleofatext-miningsystemis

reviewed,focusingonhowtheintegrationaspectis

takenintoaccount.

Informationretrievalisthefirststepofanyliterature

text-miningsystem and aimsatfindingdocuments

related to the user’s query [9] or at identifying

thetextsegments(articles,abstracts,etc.)pertaining

to a specific topic. The most t famous IR R tool for

biomedical papersisPubMed, that ismainlybased

ontwosearchmodels:(i)amodelthatusesBoolean

operators to retrieve documents by performing

queriesintheform of<DiseaseX> and<GeneY>

and(ii)avectorspacemodel[14]thatrepresentseach

documentbyavectorofindexterms,inwhicheach

term ischaracterizedbyavalueaccordingtoafre-

quency-based weighting g system. The e space vector

model is used to train machine learning methods

for discriminating relevant papers and irrelevant

papers with respect to the queries issued by the

user.However,inordertoexploitthefullpotential

ofIRsystemsformakingscientificknowledgemore

accessibleandenablingautomaticknowledgediscov-

ery,somesystemshaveextendedtext-basedsearch-

ingtooperate on other sourcesofdata (biological,

chemical, medical, drugs annotated databases).

Examples ofthese tools are: (i) Query Chem [15]

thatcombinestext-basedIR on biochemical data-

basesand WebAPItoretrievetheinformationand

relationships between compound structures and

Faroetal.

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

VB.NET Create PDF from Word Library to convert docx, doc to PDF in

Change Word hyperlink to PDF hyperlink and bookmark. VB.NET Demo Code for Converting Word to PDF. Add necessary references: RasterEdge.Imaging.Basic.dll.

add links to pdf online; adding hyperlinks to pdf documents

VB.NET Create PDF from Excel Library to convert xlsx, xls to PDF

Change Excel hyperlink to PDF hyperlink and bookmark. VB.NET Demo Code for Converting Excel to PDF. Add necessary references: RasterEdge.Imaging.Basic.dll.

add links in pdf; add url to pdf

(ii) EBIMed [16] that retrieves sentencesbased on

co-occurrencesbetweenbiologicalentitiesandiden-

tifiesrelationshipsbetweenprotein/genenamesand

drugs.

Ofcourse,IRshouldnotrelyonlyonmethodsfor

querytermsmatching,becausetermambiguitymay

causelow precisionandlowrecall. To addressthis

issue,anumberofIRtoolsthatexploitsestablished

domain ontologies, , to support semantic search in

biomedicalrepositoriesandtoguaranteemorepre-

cision with respect to the Boolean search h systems,

have been proposed. For instance, GoPubMed

(http://www.gopubmed.org/)[17]classifiesabstracts

using GOterms, GoWeb (http://www.gopubmed

.org/goweb) [18] combines keyword-based Web

searchwith textminingandontologiestoorganize

and navigate the results and facilitate question an-

swering, whereas Textpresso (http://www.text

presso.org/)[19]usesacustomontologytoquerya

collectionofdocumentsforinformationonspecific

classes of biological concepts s (gene, cell, etc.) and

theirrelations.

IRsystemsarecurrentlyfocusingonhowtopre-

sent and distill the search results and how to

cross-linktheseresultswiththebiologicaldatabases

usedintheretrievalprocess[17,20];infact,longlists

ofretrievedpapersprovideascarceoverviewofthe

problem andmaycreateconfusionin the userson

whichsourcesofdatahavebeenused.Forexample,

iHOP[21]convertstheinformationinPubMedinto

anavigablemulti-sourcesnetworkofgenesandpro-

teins(thatalsoincludesphenotypes,pathologiesand

genefunction), thusprovidingan intuitivealterna-

tivewayofaccessingthetenmillionofabstractsin

PubMed.

NER

Biological entities are the backbone of any

text-mining system, , but often the naming of f the

entitiesisinconsistentandimprecise[22]sincethey

arecitedwithavarietyofterms.Therefore,themain

goalofaNERsystemistofindthebiologicalentities

(mainly genes and proteins) that are mentioned

within a text and d to o associate them with known

namesor identifiers(IDs). Usually,thistaskisper-

formed in n two steps: first, the recognition of the

words that refer to entities s and then, the unique

identificationofsuchentities.

The earliest NER systems relied on rule-based

approaches (e.g. . in n [23]), , i.e. . they y were based on

manually crafted rules that described common

namingstructuresfor certain termclasses,basedon

morphological,orthographicandsyntacticcharacter-

istics[24].Asannotatedcorpora(inwhichgeneand

protein names are categorized) have become

available, the newer systems have relied on

machine-learningalgorithms [25,26], to recognize

the names on the e basis of f their peculiar features.

Differently, methods relying on dictionaries

[27, 28] depend on lists of synonyms of entities

names that are matched in n documents s usingalgo-

rithms that recognize variations in how w thenames

appear (e.g. gene ‘BRCA1’ may be written as

‘Brca1’, ‘BRCA 1’, ‘brca1’, etc.). However, the

most effective and recent t NER systems are based

onthecuration ofentitiesnameliststoreducethe

aliases[6,29],eventhoughtheirmaindifficultyisthe

lackofstandardizationofnames(e.g.eachgenehas

manynamesandabbreviations).Underthisscenario,

ontologies, taxonomies and controlledvocabularies

areofstrategicimportance for NER systems since

theyprovidesemanticinterpretation ofbio-entities

[30–32].

Arelevantexampleofcontrolledvocabularythat

canbeusedforNERisMedicalSubjectHeadings

(MeSH)containingabout30000termsandmainly

usedforindexingarticlesinMEDLINE(MEDLINE

isoneofthecomponentofPubmedthatindexesthe

records using MESH controlled vocabularies.) (i.e.

each article is summarized by a set of f controlled

terms). MeSH H covers s protein n functions in cellular

systems,butitisnotexhaustive.

Currently, the tendency y for NER systems is to

integrate different vocabularies or ontologies in

ordertoprovideastructured,accurateandcomplete

listofthebiologicalentitiesthatcanovercomethe

aforementioned drawback. In this direction, one

valuable approach, based on integration n of f several

controlled vocabularies, , is SemCat [33] consisting

ofalargenumberofsemanticallycategorizedterms

coming from different biomedical knowledge re-

sources (e.g. Unified Medical Language System

(UMLS) [34], Gene Ontology (GO) [35], Entrez

Gene [36], ProtScan [37] ] and ChemID D [38]) and

open-domain corpora [39]. An example of NER

systembasedonSemCatwasdevelopedbyTanabe

etal.[40].Thisapproachbuildsaprioritymodelfor

entityrecognitionbasedonthepositionofthewords

inasentence,i.e.awordontherightsideofasen-

tenceismorelikelyanentitywithrespecttoaword

on the leftside. Similarly, , knowledge-based d NER

approaches using platforms that t integrate different

Combiningliteraturetextminingwithmicroarraydata

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

.NET PDF Document Viewing, Annotation, Conversion & Processing

Extract bookmark & outlines. Extract hyperlink inside PDF. PDF Write. Insert images into PDF. Edit, remove images from PDF. Add, edit, delete links. Form Process

add a link to a pdf in acrobat; add page number to pdf hyperlink

VB.NET PDF: Basic SDK Concept of XDoc.PDF

You may add PDF document protection functionality into your VB.NET program. Hyperlink Edit. XDoc.PDF for .NET allows VB.NET developers to edit hyperlink of PDF

pdf link to specific page; accessible links in pdf

typesofontologiesfromgenetermstogeneticpath-

ways(Ageneticpathwayisalinearsequenceofgene

activities resulting from m the functional l interactions

between different genes.), to proteins, to clinical

trials],such asBioPortal (http://www.bioontology.

org/ncbo/faces/index.xhtml) and Open Biological

Ontologies (http://www.obofoundry.org/), have

beeninvestigated[41].

A recent approach based on web-services is

Whatitiz[42]thatimplementsaNERbasedonmor-

phologicalvariabilityofterms.Inparticular,itispro-

vided with numerous modules for annotating

different entities: chemical entities (whatizit

Chemical), diseases (whatizitDiseaseUMLS for ac-

cessing the UMLS Metathesaurus using the tool

MetaMap (http://metamap.nlm.nih.gov/), drugs

(whatizitDrugs maps s drugs in the e text with terms

of a controlled vocabulary built using Drugbank

(http://redpoll.pharmacy.ualberta.ca/drugbank/),

andgenes (whatizitGOsearches for geneontology

terms).

Mostrecently,algorithmsabletoidentifyanddis-

ambiguateacronymsautomatically,eveniftheseare

not mapped in any standard nomenclature, have

beeninvestigatedtoimproveNERsystemsperform-

ance[43,44].

IEfromliteratureaimsatextractingpredefinedtypes

of facts in the form m of f relationships between bio-

logical entities from the retrieved documents. . The

inputstothissteparesentences,whereastheoutputs

are relationships among biological entities.

Generally,twomainapproachesexist:

 Co-occurrences processing: these approaches

identify entities that co-occur within the text,

i.e. terms that appear r in the same e texts tend to

be related. Often these methods are able to

detect co-occurrences to o extract single relation-

shipsofa certain type: gene–gene, gene–disease,

protein–protein, etc. Several works have used

co-occurrence frequencies for r extracting known

single

relationships

[45].

For

example,

Al-Mubaid and Singh in [46] proposed a text

mining approach based on co-occurrence and

term frequency analysis, by which they found

andvalidatedsixsignificantgenesforAlzheimer’s

disease.

Co-occurrences approaches have been investi-

gated also for extracting facts that involve

multi-typedata,inlinewiththecurrentresearch’s

trendoftextandbiological dataintegration.For

instance, Mukhopadhyay et al. [47] identifies

multi-wayrelationshipsinvolvingmorethantwo

biological entities, i.e. genes, proteins, diseases,

drugsandchemicals,etc.Anexampleofidentified

relationshipis‘geneAactivatesproteinBindis-

easeCfororganDunderinfluenceofchemicalE’.

Co-occurrenceapproachestendtoprovidebetter

recallthan precision anderrors arise incomplex

sentences containing multiple relationships.

Theseapproachesareunabletoextractdirectional

relationships (i.e. A involves B B but B does not

involve A) and to distinguish different t types of

relationships, e.g. they cannot identify relation-

ships in the form ‘A is not connected to B’.

Precision can be improved by integrating

co-occurrencemethodswithruleorpattern-based

approaches.However,theseapproachestendtobe

datasetdependent,i.e.theruleorpatternsetsare

derivedfromtrainingdataoftennotapplicableto

other data different t from m the ones s used during

thetraining[48].

 Patterns parsingbyNaturalLanguageProcessing

(NLP): all the above mentioned issues are ad-

dressed byNLP approaches, which combine the

analysisofsyntaxandsemanticsinatextforob-

tainingrelationshipsbetweenfacts.Theworkflow

oftheseapproachesis:first,thetextistokenizedto

identify the boundaries of the words and sen-

tences, then a part-of-speech tagging (e.g.

[49,50])systemassignslabelssuchasnoun,verb,

adjectivetoeachword.Afterwards,asyntaxtreeis

computed for each sentence to detect noun

phrases and represent their relationships [6].

NERisthenusedto tagthe relevantbiological

entitiesintheserelationships.Finally,inorderto

identifythe evidences for entitiesrelationships,a

rule set based on the syntax tree and on the

semantic labels [51, 52] is used. For example

Fundel’setal.[53]developedRelExtoobtainde-

pendencytreesfromMEDLINEabstractsbyusing

Stanford Lexicalized Parser (http://nlp.stanford.

edu/downloads/lex-parser.shtml). Thesetreesare

then enriched with genes andproteinsbyusing

ProMiner[54],adictionary-basedNER.Finally,

a set of f three simple rules (e.g. ‘A A activates B’,

‘ActivationofAbyB’ and ‘Interaction between

AandB’)isappliedtoobtaincandidaterelation-

shipsthatarethensubmittedtoafilteringmodule

thatusesnegation check,enumerationresolution

Faroetal.

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

VB.NET Create PDF from PowerPoint Library to convert pptx, ppt to

Export PowerPoint hyperlink to PDF. VB.NET Demo Code for Converting PowerPoint to PDF. Add necessary references: RasterEdge.Imaging.Basic.dll.

pdf link to attached file; pdf hyperlinks

How to C#: Basic SDK Concept of XDoc.PDF for .NET

You may add PDF document protection functionality into your C# program. Hyperlink Edit. XDoc.PDF for .NET allows C# developers to edit hyperlink of PDF document

adding a link to a pdf in preview; active links in pdf

and restriction to the domain of interest for

screeningthecandidaterelationships.Oftenpartial

language-parsingapproachesareunableto detect

relationshipsthatspanmultiplesentences[6],and

fullparsing,providingmoreelaboratesyntacticin-

formation,isadoptedtoachievepotentiallybetter

results [55]. A typical full grammar parsing g ex-

ample is the Pro3Gres dependency parser [56]

thatintegrateshand-writtengrammarwithastat-

isticallanguagemodelforparsingunrestrictedtext

byusingdeepknowledgeoftheEnglishlanguage.

Fullparsingapproacheshavebeenrecentlyinte-

grated with multi-type data for extracting facts

involvingmoreconcepts: InfoPubMed[57]rec-

ognizes different types of interaction between

gene and proteins on MEDLINE abstracts by

combining full parsing, , machine learning tech-

niques and ontologies for NER; Pharmspresso

[58], identifies important pharmacogenomics

factsinarticlesreferencedtohumangenes,poly-

morphisms,drugsanddiseasesbyfulltextparsing

andbyexploringbiological, chemicalanddrugs

databases.

Anapproachbasedonfull-textprocessingthatalso

usesthe‘not’ conceptinproteinsrelationships, i.e.

protein A bindsprotein Bbut notprotein C,was

developedbyKim[59].Theyfound41471protein–

proteincontrastsavailableattheweb-addresshttp://

biocontrasts.biopathways.org/.

The currenttrend is to favor fulltexts over ab-

stractssincebiologicalentitiesidentifiedfrommining

only abstracts can be strongly underestimated be-

cause of abstracts’ concise nature. Methods for

mining full l biomedical texts need to be improved

substantially, especially in converting PDF or

HTML documents to plain n text and in handling

grammatical errors [60]. Another shortcoming of

currentmethodsisthattheydonotconsiderinfor-

mation hidden in tables and figures. Recently,

approachesthatintegrate textdata, biologicaldata-

bases and non-textual data (e.g. . images, graphics,

etc.)havebeen proposed and a comparative listis

providedin[7].AnexampleisSLIF[61]thatcom-

binesfigure’scaptionmining,imageprocessingand

specificdomainontologiestoextractbiomedicalin-

formation from fluorescence microscopy images.

A biologicalentityrecognitionsystemfindsprotein

andcelltypenamesintheminedcaptionsandthese

entities are associated with the patterns extracted

from the related images. Finally, a web-interface

andaXML-basedweb-serviceallowuserstoinves-

tigateandquerythederivedinformation.

Trying to discover r hidden or implicit biomedical

links and to propose them as potential scientific

hypothesesisthemaingoalofknowledgediscovery

systems.Infact,thepreviouslydescribedIEsystems

extract onlypre identified or explicit relationships.

Swanson,pioneeroftheresearchinknowledgedis-

coveryfromtext,in[62]demonstrated,byusingthe

semi-automatedArrowsmithsystem[63],hownew

knowledge canbeinferredfromexistingliterature.

Inferringindirectrelationshipsimpliestousefactsin

the form A A leads s to B and B leads to C, then a

relationshipsmaybeinferredbetweenAandC.In

detail, theuser providesahypothesisbetweentwo

biologicalentities(AisrelatedtoC)thatisfurther

provedbysearchingforrelatedterms(B)supporting

thegiven hypothesis. Anexample ofinferred rela-

tionshipsistheone‘fishoil-Raynaud’sdisease’dis-

covered by Swanson [64] or the relationship

betweenmagnesiumdeficiencyandmigrainehead-

ache [65]. These two discoveries were confirmed

experimentally [66, 67]. Several methods relying

on natural language processing exist to discover

knowledgeaboutgeneregulation[68],proteinphos-

phorylation [69, 70], gene–disease or gene–gene

interaction[71–73].

OneofthemostcompletesystemsthatusesNLP

is GeneWays [74] that examines entire articles s to

extract thephysical interactions amongdiseaseand

genes hidden in n the literature. Differently, many

other systemsare basedon co-occurrence, i.e. the

idea that two concepts (biological entities) are

relatediftheyoccurinthesamecontextsinthelit-

erature.Theycanbebased eitheron (i) firstorder

co-occurrences,e.g.entityAco-occurswithentityB

[73], [10] or (ii) second-order co-occurrences

[75, 76], i.e. entity A co-occurs with entity B

whichco-occurswithentityC,thereforethereisa

relationship between entities A and C. These

approaches share the assumption that hidden and

validrelationshipsmaybefoundbysuitablyscreen-

ing the huge number of facts retrieved by the

co-occurrence approaches.Forinstance, Jelieretal.

[76]proposetheassociativeconceptspace(ACS)to

filtertheirrelevantrelationshipsandtermsobtained

byapplyingasecondorderco-occurrenceapproach.

Indetail,ACSreflectsnotonlytheco-occurrenceof

Combiningliteraturetextminingwithmicroarraydata

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

C# Create PDF from Word Library to convert docx, doc to PDF in C#.

Change Word hyperlink to PDF hyperlink and bookmark. C#.NET Sample Code: Convert Word to PDF in C#.NET Project. Add necessary references:

add a link to a pdf; add hyperlink pdf

C# Create PDF from PowerPoint Library to convert pptx, ppt to PDF

Export PowerPoint hyperlink to PDF in .NET console application. C#.NET Demo Code: Convert PowerPoint to PDF in C#.NET Application. Add necessary references:

pdf edit hyperlink; add hyperlink to pdf

two entities, but t also indirect, , multi-step p relation-

shipsbetweenentities.

Veryfew systemshave been designed to extract

complexandmultipletypesofrelationships(e.g.find

allthegenesinvolvedinadiseaseandalltherelated

proteins) that necessarily require different types of

data to be integrated. Anni 2.0 [77] uses an

ontology-basedinterface to MEDLINE to identify

different types of associations between biomedical

concepts, including g genes, , proteins and diseases. It

resortstotheideaofconceptprofiling,i.e.a listof

concepts is presented d where each h concept t is asso-

ciated to o the analyzed d text together with a a score

describingitsimportance.Anexampleofassociation

derivedwithAnni2.0.is:‘GeneKLK3isboundto

theprostatecancer,morespecificallywithmalignant

neoplasmofprostate’. Polysearch [78] is arecently

developed web-tool able to identify associations

from published abstracts and many y well-annotated

databases. Itenablesuserstoperformqueriesinthe

form: ‘Find all genes associated with a prostate

cancer’.Uptodateitsupportsmorethan50classes

ofqueries miningmore than a a dozen oftext, ab-

stractsandbioinformaticsdatabases.Akeyfunction-

alityofPolySearchisthatitextractsandanalyzestext

data not only from PubMed but also from other

databasessuchasDrugBank[79]andHumanGene

Mutation Database (HGMD) [80]. Gendoo [81]

identifiesdiseaserelevantgenesandaimsatunder-

standingtheirmechanismsbyinterpretingdatapro-

videdbygenomesequencesandtranscriptomics.In

detail, in Gendoo the On-line Mendelian

Inheritance in Man (OMIM) (http://www.ncbi

.nlm.nih.gov/omim/) knowledge-based system,

thatcontainsabout20000entriesforhumangenes

and for genetic diseases, , is re-organized by using

MeSH, thus improving OMIM’s exploitability by

computerautomation.

Insummary,severalKDmethodshavebeenpro-

posed in the last years s where integration between

biologicaldataandunstructured/structuredtexthas

beenachievedinatleastoneoftheIR,NERandIE

sub-systems.However,torealizethefullpotentialof

text mining, , new w methods s that integrate complex

texts,biologicalandalsorawexperimental dataare

needed,withafocusonenablingbiologiststoexploit

biologicalknowledgemoreeffectively.Thisisneces-

sarybecauseanyknowledgediscoverymethodgen-

erateshypothesesaboutrelationshipstobevalidated

empirically, andreusingavailableexperimentaldata

isaneffectivestrategytospeedupscientificprogress.

COMBININGTEXTAND

MICROARRAYDATA

Literature text-mining methods are useful to dis-

cover hidden or indirect relationships, however

their integration with high-throughput methods

(e.g. microarray) is heavily demanded. To fulfill

this need, in the last years bioinformatics efforts

have been n directed toward the implementation of

toolssupportingtheintegrationofbiologicalandex-

perimentaldatawithliteratureinformationinorder

to infer biological l hypothesis that can n assume the

form of pathways, , gene regulatory networks, or,

moreingeneral,biologicalnetworksinvolvingdif-

ferententitiessuchasgenes,proteins,diseases,drugs

from

experimental data gathered from

high-

throughputmethods.

DNA microarray technology, one of the most

commonhigh-throughputmethods,allowsresearch-

erstocomeacrossbiologicalfunctionsonagenomic

scale. However, the list of the produced down

and upregulated genes is very cryptic, thus

requiring a huge e effort in data a interpretation [13,

82]. Moreover, , the e selection ofsuch lists of genes

(i.e.theclusterstobeanalyzed)isdemandedtothe

researchersthat,giventheamountofdatainvolved,

mightnotpursuethebestselectionsinceitisdifficult

tocatchthecorrelationbetweentheclusterandthe

biologicalaspecttobeinvestigated.

Therefore,inthelast10years,theattentionofthe

bioinformaticscommunityhasbeendirectedmainly

toelicitingsuchcorrelationstounderstandthebio-

logicalmeaningofthe producedlistsofgenes, in-

stead of investigatingnovel clusteringand d statistics

methodologies.Understandingthebiologicalmean-

ingofasetofupanddownregulatedgenesderived

frommicroarrayexperimentsisoneaspectofcurrent

bioinformaticseffortsin data integration; theother

oneforeseestextminingcombinedwithmicroarray

datatargetedtothegenerationandtotheevaluation

of biological hypotheses, which can be obtained

either byminingmicroarraydatathatinvolvesdata

clusteringandmanualselectionoftheclusterstobe

analyzedorbyminingtheliteratureusingtheknow-

ledgediscoverysystemsbeforedescribedor byex-

ploiting the knowledge of the biomedical

researchers.

Understandingbiologicalmeaning

Themostnaturalwaytoassignabiologicalmeaning

to asetofgenesthathasbeenobtainedbymining

microarray data is to project it onto biological

Faroetal.

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

processes that t can n be represented in the form, in

orderofincreasingcomplexity,ofGOterms,path-

waysandgeneregulatorynetworks(Ageneregula-

torynetworkrepresentsacollectionofsegmentsof

DNA that maps gene regulations in living cells.)

previously identified or manually compiled by

researchers[82](seeFigure1).

The most common (and simple) approach for

genelistprojectionistouseGOforinterrelatinga

listofgeneswithabiologicalprocessand/oramo-

lecularfunctionand/oracellularcomponent.GOis

alsousedtoranksignificantgenes(producedinthe

microarray experiment) in n relationship to o the GO

categories. A list of about 70 methods and

tools that carry y out GO-based microarray analysis

is reviewed in [83]. An approach that goes

beyond simple GO classification is Onto-Express

[84], since it associates lists of up and down

regulated genes with functional profiles built by

correlatingGOterms(biologicalprocesses,chemical

components, molecular functions) with expression

profiles.

However,theGOclassificationdoesnotprovide

exhaustiveinformationaboutthebiologicalcontext

ofagivensetofupanddownregulatedgenes.This

canbeachievedbypathwayanalysisand/orbyregu-

latorynetworkanalysis.Pathwayanalysismainlyin-

vestigates the functional and physical interaction

among genes instead of using the gene-centered

view as GO-based approaches. These systems try

to mapgenesderivedfrommicroarrayexperiments

onto precompiled pathways derived by manually

analyzingthe literature. Mostnon-commercial sys-

temsforpathwayanalysisrelyontheKEGGdatabase

(http://www.genome.jp/kegg/) that contains a

collection of pathways representing the current

knowledge on gene and molecular interaction.

A comprehensive list of tools for pathways

analysis can be found at the weblink http://

www.geneontology.org/GO.tools.microarray.shtml.

Pathwaymappingofmicroarraydata, usually,gen-

eratesmorethanonepathway,therefore,itisneces-

sarytorankthemaccordingtotheirrelevancetothe

dataset.PathwaysrankingisprovidedinGenMAPP

Figure 1: Understandingbiologicalmeaningofasetofregulatedgenes.Themostcommonwaysforunderstanding

thebiologicalmeaningofasetofgenesare:(i)toprojectitontobiologicalprocessesrepresentedintheformof

GO terms,pathwaysandgeneregulatorynetworks(bluerectangle)and/or(ii)toannotatethelistsofregulated

genesbasedonliteratureprofiling(redrectangle).Pathwaysandgeneregulatorynetworksareusuallyderivedby

manualliteratureanalysis.

Combiningliteraturetextminingwithmicroarraydata

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

2 [85],where the users can rank, , andat thesame

time, customize the pathways; an extension of

GenMAPP2 [85] proposes the Fisher’s exact test

for ranking the relationships between genes and

pathways. PathExpress [86, 87], instead, identifies

the most relevant metabolic pathways associated

with a subset of genes using P-values. The

KEGG-based web-tool KOBAS [88] proposes a

controlled vocabulary for r gene e pathways mapping

andtherelevanceofthediscoveredpathwaysisesti-

mated using g binomial, Chi-square and hypergeo-

metric distribution test. Although pathway-based

approachesprovidedeeperinformationonbiological

processes possibly y relevant to o a a set of genes, their

mainshortcomingisthatbiologicalprocessesusually

dependonmorethanonepathwayandtheconnec-

tions between suchpathwaysis relatedto thebio-

logicalcontext.Theinterconnectionofpathwaysis

definedas generegulatorynetwork. Thisnetwork

cannot be easilyderived by simplycombining pre

compiled pathway y because the e networks’ morph-

ology changes with the biological context. The

earliest attempts for building g gene regulatory net-

works have been successful only for lower

eukaryotes withsimplegenomes[89, 90]. Current

approaches (both stand alone and also combined

with GO classification and pathways), , instead, are

directedtowardmorecomplexmammaliansystems.

Forinstance,ARACNe[91] buildsregulatorynet-

worksinmammaliancellsbyidentifyingtranscrip-

tional interactions among genes from microarray

expression profiles. An interesting g effort is repre-

sented by MONET [92] a method based on

Bayesian networks for inferring gene regulatory

networks. It t mainly consists of two steps: : the e first

aimsatsplittingthewholegenesetintooverlapped

groupsthatcontaingeneswhoseGOannotationsor

microarrayexpressionpatternsarehighlycorrelated.

Finally, the second step infers Bayesian networks

over each group and integrates such groups into

globalregulatorynetworks.BioCAD[93]integrates

both the above inference tools (ARACNe and

MONET) for building g gene regulatory networks.

Thetoolalsosupportsvalidationoftheinferrednet-

works by y integrating gene and protein regulatory

networksderivedfromMEDLINEabstractsusinga

text-miningsystembasedonSTRING-IE[94].

The described approaches provide as outcomes

precomputedrelationships between genesandbio-

logicalprocesses.However,theliteraturemayenrich

the information about relationships regulated

genes-biological processes much more e than n struc-

tured ontologies or precompiled pathways can do.

Toextracttheadditionalinformationhiddeninthe

literature, severalmethodsthatannotatethelistsof

regulated genes based on literature profiling g have

been proposed [95, 96]. Most of f these approaches

arebasedonkeywordsover-representationofaset

ofgenes,similarlytoGO-basedmicroarrayanalysis,

butwherethekeywordstobeassociatedtothegene

setaregatheredbyminingdirectlyMEDLINEand

theyareusedtointerpretgenesindomainsscarcely

coveredbyGO.In detail, suchmethodsretrievea

subsetofMEDLINEabstractsassociatedwithoneor

moregenes,e.g.aclusterofgenesderivedbygene

setanalysis methods [97, 98].Then,theseabstracts

areusedtoidentifyrelevantkeywordsinthetextor

annotated MeSH terms (medical subject heading

terms), thus helpingthe gene sets characterization.

Forexample,GenClip[99],oneofthemostrecent

tools, builds s functional clusters of genes related to

disease pathogenesis starting from a list of genes

from microarray. The tool first identifies keywords

astermsthatco-occurinatleasttwooftheanalyzed

genesbyminingliteratureabstractsandthenclusters

thelistofgenesbasedonkeywordoccurrences,thus

obtainingfunctionalclusters.Differently,Chagoyen

etal.[100]proposesasystemforliteratureprofilingof

largesetsofgenesorproteinsthatcanbeusedtofind

similarities among g genes. The method starts from

creatinga pool of documents related d to a a specific

gene. Afterwards, the pool of f documents is con-

vertedintoavectorspacerepresentationandfinally,

thenon-negativematrixfactorization[101]isapplied

to thevectorspace,thusobtainingforeachgenea

literatureprofile(Aliteratureprofilecanbeseenasa

pictureofthefunctionalrelationships,derivedfrom

scientific papers, between set of genes). CoPub

[102],providesaninsightintothebiologicalmech-

anisms relatedto a a setofregulatedgenesfor liver

pathologiesbycalculatingstatisticsforgene-keyword

co-occurrencesusingtheentireMEDLINEabstracts,

insteadofonlyasubset,asthepreviousapproaches

do.Theinputsofthetoolareasubsetofgenesob-

tained by microarray data a processing and d a a set of

keywords, whereas a navigable network of

MEDLINEabstractswherethegenesandthekey-

words co-occur is provided as output. The text

mining method extracts networks of abstracts by

analyzingtheco-occurrencesofhuman,mouseand

ratgeneswithkeywordsdescribingliverpathologies,

pathways,GOterms,diseases,drugsandtissues.An

Faroetal.

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

approachthatexploitsliteratureforsupportinggene

listinterpretationistheoneproposedbyJelieretal.

[103],whichusesassociationsderivedfromliterature

(usingtheAnni2.0tool)toprovideaninterpretation

ofgeneexpressionchanges.Indetail,theypropose

the literature-weighted global test to o compute e the

correlation between associations (in the form

gene-biomedical concept) obtained bymining g the

literatureandlistofgenesextractedfrommicroarray

andtheyprovideasoutputthescoresreflectingthe

importanceofageneforaconceptofanassociation.

Table1showssomeoftheweb availabletoolsfor

understandingbiological meaningofasetofregu-

latedgenesderivedfromexperimentaldata.

Hypothesisgeneration

Hypothesisgeneration has the objective to suggest

undiscoveredassociations betweenbiomedicalcon-

cepts;thisisdifferentfromtheattemptsatproviding

a biological l meaningto a set offacts (e.g. lists of

genes)extractedfromexperimental data.The most

explored venue for hypothesis generation concerns

thediscoveryofgenes andotherbiological entities

(togetherwiththeirrole)involvedinaspecificdis-

ease(disease-centricanalysis).Infact,predictingbio-

logicalentities(andtheirrole)involvedinadisease

beforeexperimentalanalysismaysavetimeandeffort

byindicatingwhere the research should lookinto.

Textminingandmicroarraydata havebeen com-

bined in two main ways to achieve this

goal:(i)startingfromamicroarrayrelatedtoaspecific

disease, a list ofgenes(thehypothesis) is extracted

(e.g.oneormorecluster)andthentheroleofsuch

genes(i.e. theprioritization)inthegivendiseaseis

exploredusinginformationextractedfromliterature

and(ii) the hypothesis is generated from literature

mining in the e form m of associations between genes

anda given disease and then these associationsare

filteredandvalidatedbyresortingtomicroarraydata.

Thefirstapproachesfollowthisworkflow:givena

setofgenes(thehypothesis)eithergathereddirectly

frommicroarraydataanalysisorpreviouslystoredin

publicdatabases(suchasGeneExpressionOmnibus),

theliterature(mainlyMEDLINE)ismined,starting

fromthissetofgenes,inordertoelicitagenepri-

oritizationforthegivendiseaseortofindoutother

biological concepts involved in the same disease

(Diseasemodeling)(seeFigure2).Theworkflowis

similar to the biological understanding approach’s

onewiththedifferencethatinthiscasetheoutcome

isarefinementandacloseexaminationoftheinput

hypothesisregardingconceptsandtheir role in the

given disease, whereas inthe context ofbiological

understanding the output is the association of a

meaningtoalistofgenes(facts).

One of f the most complete tools for r hypothesis

generation from microarray is s G2D [104, , 105]. It

performsgenesprioritizationrelatedtoinheriteddis-

easesbycombiningMeshannotationsinMEDLINE

andasetofgeneswiththeGOannotationsofentries

Table1: Listofthewebavailabletoolsforunderstandingbiologicalmeaningofasetofregulatedgenesderived

fromexperimentaldata

Description

UsedResources

Availableat

GenMAPP2

Visualize gene expression n data

biologicalpathways

GO,KEGG

http://www.genmapp.org/

PathExpress

MappingofasetofGenesonto

Pathways.

KEGGSwiss-Protdatabase

Blastx

http://bioinfoserver.rsbs.anu.edu

.au/utils/PathExpress/

KOBAS

Identifystatisticallyenriched

pathwaysforasetofgenesor

proteins

PathwaysDatabase:KEGG,PID

Curated

BioCyc

andPanther

http://kobas.cbi.pku.edu.cn/home.do

ARACNe

Estimategeneregulatorynetworks

inmammaliancellsusing

microarrayexpressionprofiles

Expressionprofiledatasetofhuman

Blymphocytecellsbuiltbythe

authors

http://wiki.c2b2.columbia.edu/

califanolab/index.

php/Software/ARACNE

GenCLIP

Clusteringofgenelistsbyliterature

profiling

NCBIEUtilitiesforTextMining

GeneList:HUGO

EntrezGene,

orUnigene

http://www.genclip.com

CoPub

Findbiomedicalconceptsfrom

Medlinelinkedtoageneset

(Affymetrixidentifiers)

NCBIE-UtilitiesforTextMining

http://services.nbic.

nl/cgi-bin/copub/CoPub.pl

http://expasy.org/sprot/;

http://blast.ncbi.nlm.nih.gov/;

http://pid.nci.nih.gov/;

http://biocyc.org/;

http://www.pantherdb.org/pathway/;

http://www.genenames.org/;

http://www.ncbi.nlm.nih.gov/unigene.

Combiningliteraturetextminingwithmicroarraydata

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

inNCBI RefSeq[106](collectionofannotatedse-

quences, including genomic DNA, , transcripts and

proteins). More specifically, it receives as input a

genomic region n and an OMIM disease identifier

andprovidesasoutputthegenespotentiallyinvolved

inthegivendisease.Indetail,forthediseaseunder

analysis the MESH terms from the ‘Disease

Category’ associated to publications in OMIM are

retrieved. These terms are then associated with

chemical, drugs and molecular functions by using

GO. Thedetectedmolecular functionsareusedto

identifyasequenceofDNAbyqueryingtheRefSeq

proteindatabase.Thissequenceisintegratedwitha

chromosomallocationforthegivendiseaseprovided

bytheOMIMdatabaseinordertoobtainalistof

genesrelatedtotheanalyzeddisease.G2Dwasori-

ginallydevelopedonlyfor Mendeliandiseases;cur-

rently, it t also o works for complex genetic diseases

[107].Likewise,Tiffinetal.[108]proposegenespri-

oritization according to the relationship disease-

affected tissue. The method integrates literature

discoveries (co-occurring disease and tissue names

in MEDLINE) and human gene expression data

from the Ensemble database [109] to link gene

expressionstodiseasesbyusingananatomicalontol-

ogy.First,thetoolassociatesanatomicaltermsfrom

anontologyforhumananatomicalsystemsandcell

types(eVOC[110])todiseasesnames,basedonthe

co-occurrenceinPubmedabstracts.Eachtermofthe

eVOContologyisthenrankedaccordingtothefre-

quency of annotation. . The top-scoring terms are

comparedwiththetermsalreadyannotatedtocan-

didate disease genes s using the Ensemble database.

Thegenesthatmismatchwiththealreadyannotated

genesrepresentthelistofgenestobeexplored.

The second approaches (Figure 3) generate

hypotheses(intheformofassociationsgene-disease)

byminingliterature,thentheyvalidatesuchhypoth-

esesbycheckingifthereisanyevidenceofthedis-

covered relationships in the experimental data. To

thebestofourknowledge,fewmethodsuseknow-

ledge gathered from m the literature for hypotheses

generation

and

validate

these

sets using

high-throughputmethods,thusallowingtheidenti-

ficationofnovelbiologicalentitiesrelationships.

Anapproachinthisdirectionistheoneproposed

by Faro et t al. [11], where hypothesis generation

about gene–diseases relationships is made by

miningspecializedliteratureusingtheco-occurrence

processing approach described in [10] and the

inferredrelationshipsareselectedandthenvalidated

by means of microarray data analysis. The used

text-miningalgorithmtendstoprovidebetterrecall

than precision, i.e. it provides more relationships

Figure 2: Hypothesisgenerationbymicroarraydataanalysis.Thefirstapproachfollowsthisworkflow:givenaset

ofgenes(thehypothesis)eithergathereddirectlyfrommicroarraydataanalysisorpreviouslystoredinpublicdata-

bases,theliteratureismined,startingfromthissetofgenes,inordertoelicitageneprioritizationforthegivendis-

easeortofindoutotherbiologicalconceptsinvolvedinthesamedisease(Diseasemodeling).

Faroetal.

by guest on May 19, 2016

http://bib.oxfordjournals.org/

Downloaded from

Documents you may be interested

windows form application in c# with database pdf : Add hyperlink to pdf online SDK application service wpf azure web page dnn 61.full0-part1940