Information Retrieval from Biological Databases

Information Retrieval from Biological Databases

Andreas D. Baxevanis

Introduction

On April 14, 2003, the biological community celebrated the achievement of the Human

Genome Project’s major goal: the complete, accurate, and high-quality sequencing of the

humangenome(InternationalHumanGenomeSequencingConsortium2001;Schmutzetal.

2004). The attainment of this goal, which many have compared to landing a person on the

moon, has had a profound effect on how biological and biomedical research is conducted

and will undoubtedly continue to have a profound effect on its direction in the future. The

availabilityof not just human genome data, but also human sequence variation data, model

organism sequence data, and information on gene structure and function provides fertile

ground for biologists to better design and interpret their experiments in the laboratory,

fulfillingthepromiseofbioinformaticsinadvancingandacceleratingbiologicaldiscovery.

OneofthemostimportantdatabasesavailabletobiologistsisGenBank,theannotatedcol-

lectionofallpubliclyavailableDNAandproteinsequences(Bensonetal.2017;seeChapter1).

Thisdatabase,maintainedbytheNationalCenterforBiotechnologyInformation(NCBI)atthe

NationalInstitutesofHealth(NIH),representsacollaborativeeffortbetweenNCBI,theEuro-

pean Molecular Biology Laboratory (EMBL), and the DNA Data Bank of Japan (DDBJ). At

thetimeofthiswriting,GenBankcontainedover200millionsequencesandover300trillion

nucleotide bases. The completion of human genome sequencing and the sequencing of an

ever-expandingnumberofmodelorganismgenomes,aswellastheexistenceofagargantuan

numberofsequencesingeneral,providesagoldenopportunityforbiologicalscientists,owing

to the inherent value of these data. However, at the same time, the sheer magnitude of data

presents a conundrum to the inexperienced user, resulting not just from the size of the “se-

quenceinformationspace”butfromthefactthattheinformationspacecontinuestogetlarger

and larger – by leaps and bounds – at a pace that will continue to accelerate, even though

humangenomesequencinghaslongbeen“completed.”

TheeffectoftheHumanGenomeProjectandothersystematicsequencingprojectsonthe

continuedaccumulationofsequencedataisillustratedbythegrowthofGenBank,asshown

inFigure2.1;theexponentialgrowthrateillustratedinthefigureisexpectedtocontinuefor

sometimetocome.Thecontinuedexpansionofnotjustthesequencespacebutofthemyriad

biologicaldatanowavailable becauseoftheexpansionofthesequencespaceunderscoresthe

necessity for all biologists to learn how to effectively navigate this information for effective

use in their work – even allowing investigators toavoid performing expensive experiments

themselvesbasedonthedatafoundwithinthesevirtualtreasuretroves.

GenBank(oranyotherbiologicaldatabase,forthatmatter)serveslittlepurposeunlessthe

datacanbeeasilysearchedandentriesretrievableinauseful,meaningfulformat.Otherwise,

sequencingeffortssuchasthosedescribedabovehavenousefulend–withouteffectivesearch

andretrievaltools,thebiologicalcommunityasawholecannotmakeuseoftheinformation

hiddenwithinthesemillionsofbasesandaminoacids,muchlessthestructurestheyformor

Bioinformatics,FourthEdition.EditedbyAndreasD.Baxevanis,GaryD.Bader,andDavidS.Wishart.

©2020JohnWiley&Sons,Inc.Published2020byJohnWiley&Sons,Inc.

CompanionWebsite:www.wiley.com/go/baxevanis/Bioinformatics_4e

20 Information Retrieval from Biological Databases

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

2016

Base pairs

(squares, billions)

Sequences

(circles, millions)

Figure 2.1 The exponential growth of GenBank in terms of number of nucleotides (squares, in millions)

and number of sequences submitted (circles, in thousands). Source data for the figure have been obtained

from the National Center for Biotechnology Information (NCBI) web site. Note that the period of accel-

erated growth after 1997 coincides with the completion of the Human Genome Project’s genetic and

physical mapping goals, setting the stage for high-accuracy, high-throughput sequencing, as well as the

development of new sequencing technologies (Collins et al. 1998, 2003; Green et al. 2011).

themutationstheyharbor.Muchefforthasgoneintomakingsuchdataaccessibletothebiolo-

gist,andaselectionoftheprogramsandinterfacesresultingfromtheseeffortsarethefocusof

thischapter.ThediscussionwillcenteronqueryingdatabasesmaintainedbyNCBI,asthese

more“general”repositoriesarefarandawaytheonesmostoftenaccessedbybiologists,but

attentionwillalsobegiventospecializeddatabasesthatprovideinformationnotnecessarily

foundthroughtheuseofEntrez,NCBI’sintegratedinformationretrievalsystem.

Integrated Information Retrieval: The Entrez System

One of the most widely used interfaces for the retrieval of information from biological

databasesistheNCBIEntrezsystem.Entrezcapitalizesonthefactthattherearepre-existing,

logicalrelationshipsbetweentheindividualentriesfoundinnumerouspublicdatabases.For

example,apaperinPubMedmaydescribethesequencingofagenewhosesequenceappears

inGenBank.Thenucleotidesequence,inturn,maycodeforaproteinproductwhosesequence

is stored in NCBI’s Protein database. The three-dimensional structure of that protein may

be known, and the coordinates for that structure may appear in NCBI’s Structure database.

Finally, there may be allelic or structural variants documented for the gene of interest,

catalogedindatabasessuchastheSingleNucleotidePolymorphismDatabase(calleddbSNP)

ortheDatabaseofGenomicStructuralVariation(calleddbVAR),respectively.Theexistence

ofsuchnaturalconnections,allhavingabiologicalunderpinning,motivatedthedevelopment

ofamethodthroughwhichalloftheinformationaboutaparticularbiologicalentitycouldbe

foundwithouthavingtosequentiallyvisitandqueryindividualdatabases,onebyone.

Entrez,tobeclear,isnotadatabaseitself.Rather,itistheinterfacethroughwhichitscom-

ponentdatabasescanbeaccessedandtraversed–anintegratedinformationretrievalsystem.

The Entrez information space includes PubMed records, nucleotide and protein sequence

data, information on conserved protein domains, three-dimensional structure information,

andgenomicvariationdatawithpotentialclinicalrelevance,agoodnumberofwhichwillbe

toucheduponinthischapter.ThestrengthofEntrezliesinthefactthat allofthisinformation,

across a large number of component databases, can be accessed by issuing one – and only

[EOF - 小节结束：下一小节为 Integrated Information Retrieval: The Entrez System]

20 Information Retrieval from Biological Databases

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

2016

Base pairs

(squares, billions)

Sequences

(circles, millions)

Figure 2.1 The exponential growth of GenBank in terms of number of nucleotides (squares, in millions)

and number of sequences submitted (circles, in thousands). Source data for the figure have been obtained

from the National Center for Biotechnology Information (NCBI) web site. Note that the period of accel-

erated growth after 1997 coincides with the completion of the Human Genome Project’s genetic and

physical mapping goals, setting the stage for high-accuracy, high-throughput sequencing, as well as the

development of new sequencing technologies (Collins et al. 1998, 2003; Green et al. 2011).

themutationstheyharbor.Muchefforthasgoneintomakingsuchdataaccessibletothebiolo-

gist,andaselectionoftheprogramsandinterfacesresultingfromtheseeffortsarethefocusof

thischapter.ThediscussionwillcenteronqueryingdatabasesmaintainedbyNCBI,asthese

more“general”repositoriesarefarandawaytheonesmostoftenaccessedbybiologists,but

attentionwillalsobegiventospecializeddatabasesthatprovideinformationnotnecessarily

foundthroughtheuseofEntrez,NCBI’sintegratedinformationretrievalsystem.

Integrated Information Retrieval: The Entrez System

One of the most widely used interfaces for the retrieval of information from biological

databasesistheNCBIEntrezsystem.Entrezcapitalizesonthefactthattherearepre-existing,

logicalrelationshipsbetweentheindividualentriesfoundinnumerouspublicdatabases.For

example,apaperinPubMedmaydescribethesequencingofagenewhosesequenceappears

inGenBank.Thenucleotidesequence,inturn,maycodeforaproteinproductwhosesequence

is stored in NCBI’s Protein database. The three-dimensional structure of that protein may

be known, and the coordinates for that structure may appear in NCBI’s Structure database.

Finally, there may be allelic or structural variants documented for the gene of interest,

catalogedindatabasessuchastheSingleNucleotidePolymorphismDatabase(calleddbSNP)

ortheDatabaseofGenomicStructuralVariation(calleddbVAR),respectively.Theexistence

ofsuchnaturalconnections,allhavingabiologicalunderpinning,motivatedthedevelopment

ofamethodthroughwhichalloftheinformationaboutaparticularbiologicalentitycouldbe

foundwithouthavingtosequentiallyvisitandqueryindividualdatabases,onebyone.

Entrez,tobeclear,isnotadatabaseitself.Rather,itistheinterfacethroughwhichitscom-

ponentdatabasescanbeaccessedandtraversed–anintegratedinformationretrievalsystem.

The Entrez information space includes PubMed records, nucleotide and protein sequence

data, information on conserved protein domains, three-dimensional structure information,

andgenomicvariationdatawithpotentialclinicalrelevance,agoodnumberofwhichwillbe

toucheduponinthischapter.ThestrengthofEntrezliesinthefactthat allofthisinformation,

across a large number of component databases, can be accessed by issuing one – and only

Integrated Information Retrieval: The Entrez System 21

one–query.Thisverypowerful,integratedapproachismadepossiblethroughtheuseoftwo

generaltypesofconnectionsbetweendatabaseentries: neighboringandhardlinks.

Relationships Between Database Entries: Neighboring

The concept of neighboring enables entrieswithin a given database to be connected to one

another. If a user is lookingat a particular PubMed entry, the user can then “ask” Entrez to

findalloftheotherpapersinPubMedthataresimilarinsubjectmattertotheoriginalpaper.

Likewise,ifauserislookingatasequenceentry,Entrezcanreturnalistofallothersequences

thatbearsimilaritytotheoriginalsequence.Theestablishmentofneighboringrelationships

withinadatabaseisbasedonstatisticalmeasuresofsimilarity,someofwhicharedescribedin

moredetailbelow.Whiletheterm“neighboring”hastraditionallybeenusedtodescribethese

connections,theterminologyontheNCBIwebsitedenotesneighborsas“relateddata.”

BL AST Biologicalsequencesimilaritiesaredetectedandsequencedataarecomparedwithone

another using the Basic Local Alignment Search Tool, or BLAST (Altschul et al. 1990). This

algorithmattemptstofindhigh-scoringsegmentpairs–pairsofsequencesthatcanbealigned

with one another and, when aligned, meet certain scoring and statistical criteria. Chapter 3

discussesthefamilyofBLASTalgorithmsandtheirapplicationatlength.

VAST Molecular structure similarities are detected and sets of coordinate data are com-

pared using a vector-based method known as VAST (the Vector Alignment Search Tool;

Gibrat et al. 1996). This methodology uses geometric criteria to assess similarity between

three-dimensionaldomains,andtherearethreemajorstepsthattakeplaceinthecourseofa

VASTcomparison:

• First,basedon knownthree-dimensionalcoordinatedata,thealphahelicesandbetastrands

thatconstitutethestructuralcoreofeachproteinareidentified.Straight-linevectorsarethen

calculatedbasedonthepositionofthesesecondarystructuralelements.VASTkeepstrack

of how one vector is connected to the next (that is, how the C-terminal end of one vector

connects to the N-terminal end of the next vector), as well as whether each vector repre-

sents an alphahelix or a beta strand. Subsequent comparison steps useonlythese vectors

inassessingstructuralsimilaritytootherproteins–so,ineffect,mostofthepainstakingly

deducedatomiccoordinatedataarediscardedatthisstep.Thereasonforthisapparentover-

simplificationissimplyduetothescaleoftheproblemathand;withthe150000structures

intheMolecularModelingDatabase(MMDB;Madejetal.2014)availableatthetimeofthis

writing,thetimethatitwouldtaketodoanin-depthcomparisonofeachandeveryoneof

thesestructureswithalloftheotherstructuresinMMDBwouldmakethecalculationsboth

impracticalandintractable.

• Next, the algorithm attempts to optimally align these sets of vectors, looking for pairs of

structuralelementsthatareofthesametypeandrelativeorientation,withconsistentcon-

nectivity between the individual elements. The object is to identify highly similar “core

substructures,”pairsthatrepresentastatisticallysignificantmatchabovethatwhichwould

beobtainedbycomparingrandomlychosenproteinswithoneanother.

• Finally,arefinementisdoneusingMonteCarlo(randomsearch)methodsateachresidue

positiontooptimizethestructuralalignment.Theresultantalignmentneednotbeglobal,

asmatchesmaybebetweenindividualstructuraldomainsoftheproteinsbeingcompared.

In 2014, a significant improvement to VAST was introduced. This new approach, called

VAST+ (Madej et al. 2014), moves beyond assessing structural similarity by comparing

individual three-dimensional domains with one another; instead, it considers the entire set

of three-dimensional domains within a macromolecular complex. This approach essentially

movesthecomparisonfromthetertiarystructuretothequaternarystructurelevel,enabling

the identification of similar functional, multi-subunit assemblies. In the VAST+ parlance,

macromolecularcomplexesarereferredtoasa“biologicalunit”andcanincludenotjustthe

22 Information Retrieval from Biological Databases

proteinsthatconstitutethecomplex,butalsonucleotidesandchemicalswheresuchstructural

informationisavailable.TheVAST +comparisonbeginsasdescribedaboveforVASTandthen

marchesthroughanumberofstepsthatinvolvetheidentificationofbiologicalunitsthatcan

be superimposed, calculation of root-mean-square deviations (RMSDs) of the superimposed

structuresasaquantitativemeasureofthesuperposition(seeBox12.1),and,finally,performs

arefinementsteptoimprovetheRMSDvaluesforthesuperposition.Theresultofthisprocess

is a global structural alignment where both the most and least similar parts of the aligned

moleculescanbeidentifiedand,fromabiologicalstandpoint,comparisonsbetweensimilarly

shapedproteinscanbefacilitated;itcanalsobeusedinthecontextoflookingatconforma-

tionalchangesofasinglecomplexundervaryingconditions.WhileVAST +isnowthedefault

methodforidentifyingstructuralneighborswithintheEntrezsystem,keepinmindthatthe

algorithm depends on biological units being explicitly identified within the source Protein

Data Bank (PDB) coordinate data records that form the basis for MMDB records; if no such

biologicalunitsaredefined,theoriginalVASTalgorithmisthenusedforthecomparisons.

ByusingapproachessuchasVASTandVAST +,itispossibletofindstructuralrelationships

betweenproteinsincaseswheresimplylookingatsequencesimilaritymaynotsuggestrelat-

edness–informationthatcould,withadditionaldataandinsights,beusedtohelpinformthe

questionoffunctionalsimilarity.Moreinformationonadditionalstructurepredictionmeth-

ods based on X-ray or nuclear magnetic resonance (NMR) coordinate data can be found in

Chapter12.

Weighted Key T erms The problemofcomparingsequence or structure data somewhatpales

nexttothatofcomparingPubMedentries,whichconsistoffreetextwhoserulesofsyntaxare

notnecessarilyfixed.Giventhatnotwopeople’swritingstylesareexactlythesame,findinga

waytocompareseeminglydisparateblocksoftextposesasubstantialproblem.Entrezemploys

amethodknownastherelevancepairsmodelofretrievaltomakesuchcomparisons,relying

onweightedkeyterms(WilburandCoffee1994;WilburandYang1996).Thisconceptisbest

describedbyexample.Considertwomanuscriptswiththefollowingtitles:

BRCA1asaGeneticMarkerforBreastCancer

GeneticFactorsintheFamilialTransmissionoftheBreastCancerBRCA1Gene

Bothtitlescontaintheterms BRCA1,Breast,and Cancer,andthepresenceofthesecommon

termsmayindicatethatthemanuscriptsaresimilarinsubjectmatter.Theproximitybetween

thewordsisalsoconsidered,sothatwordscommontotworecordsthatareclosertogetherare

scoredhigherthancommonwordsthatarefurtherapart.Intheexample,theterms Breastand

Cancerarealwaysnexttoeachother,sotheywouldscorehigherbasedonproximitythaneither

ofthosewordswouldagainst BRCA1.Commonwordsfoundinatitlescorehigherthanthose

foundinanabstract,sincetitlewordsarepresumedtobe“moreimportant”thanthosefound

in the body of an abstract. Overall, weighting depends inversely on the frequency of a given

word among all the entries in PubMed, with words that occur infrequently in the database

assignedahigherweightwhilecommonwordsaredown-weighted.

Hard Links

The hard link concept is simpler and much more straightforward than the neighboring

approachesdescribedabove.Hardlinksareappliedbetweenentriesin differentdatabasesand

existwhereverthereisalogicalconnectionbetweenentries.Forinstance,ifaPubMedentry

describesthesequencingofachromosomalregioncontainingageneofinterest,ahardlinkis

establishedbetweenthePubMedentryandthecorrespondingnucleotideentryforthatgene.

If an open reading frame in that gene codes for a known protein, a hard link is established

betweenthenucleotideentryandtheproteinentry.Iftheproteinentryhasanexperimentally

deducedstructure,ahardlinkwouldbeplacedbetweentheproteinentryandthestructural

entry.

SearchescanbeginanywherewithintheEntrezecosystem–therearenoconstraintsonthe

user as to where the foray into this information space must begin. However, depending on

whichdatabaseisusedasthejumping-offpoint,differentdatabasefieldswillbeavailablefor

Integrated Information Retrieval: The Entrez System 23

searching.Thisstandstoreason,astheentriesindifferentdatabasesarenecessarilyorganized

differently,reflectingthebiologicalnatureoftheentitiesthateachdatabaseistryingtocatalog.

The Entrez Discovery Pathway

The best way to illustrate the integrated nature of the Entrez system and to drive home the

powerofneighboringisbyconsideringsomebiologicalexamples.Thesimplestwaytoquery

Entrezisthroughtheuseofindividualsearchterms,coupledtogetherbyBooleanoperators

such as AND, OR, or NOT. Consider the case in which one wants to retrieve all available

information on a gene namedDCC (deleted in colorectal carcinoma), limiting the returned

informationtopublicationswhereaninvestigatornamedGuyA.Rouleauisanauthor.There

isaverysimplequeryinterfaceatthetopoftheNCBIhomepage,allowingtheusertoselect

whichdatabasetheywanttosearchfromapull-downmenuandatextboxwherethequery

termscanbeentered.Inthiscase,tosearchforpublishedpapers,PubMedwouldbeselected

fromthepull-downmenuand,withinthetextboxtotheright,theuserwouldtype DCC AND

"Rouleau GA" [AU] .The [AU]qualifyingthesecondsearchtermindicatestoEntrezthat

this is anauthor term, so only the author field in entries should be considered when evalu-

atingthispartofthesearchstatement.TheresultofthequeryisshowninFigure2.2.Here,

Figure 2.2 Results of a text-based Entrez query against PubMed using Boolean operators and field delimiters. The initial query ( DCC AND

"Rouleau GA" [AU] ) is shown in the search box near the top of the window, with the three papers identified using this query following

below. Each entry gives the title of the manuscript, the names of the authors, and the citation information. The actual record can be

retrieved by clicking on the name of the manuscript.

24 Information Retrieval from Biological Databases

T able 2.1 Entrez Boolean search statements.

General syntax:

search term [tag] Boolean operator search term [tag] ...

where [tag] =

[ACCN] Accession

[AD] Affiliation

[ALL] Allfields

[AU] Authorname

Lentz R [AU] yieldsallof LentzRA,LentzRB,etc.

"Lentz R" [AU] yieldsonly LentzR

[AUID] Uniqueauthoridentifier,suchasanORCIDID

[ECNO] EnzymeCommissionnumbers

[EDAT] Entrezdate

YYYY/MM/DD, YYYY/MM,or YYYY;insertacolonfordaterange,

e.g. 2016:2018

[GENE] Genename

[ISS] Issueofjournal

[JOUR] Journaltitle,officialabbreviation,orISSNnumber

Journal of Biological Chemistry

J Biol Chem

0021-9258

[LA] Language

[MAJR] MeSHmajortopic

Oneofthe majortopicsdiscussedinthearticle

[MH] MeSHterms

Controlledvocabularyofbiomedicalterms( subject)

[ORGN] Organism

[PDAT] Publicationdate

YYYY/MM/DD, YYYY/MM,or YYYY;insertacolonfordaterange,

e.g. 2016:2018

[PMID] PubMedID

[PROT] Proteinname(forsequencerecords)

[PT] Publicationtype,includes:

Review

Clinical Trial

Lectures

Letter

Technical Report

[SH] MeSHsubheading

UsedtomodifyMeSHTerms

stenosis [MH] AND pharmacology [SH]

[SUBS] Substancename

Nameofchemicaldiscussedinarticle

[SI] SecondarysourceID

Namesofsecondarysourcedatabanksand/oraccessionnumbersof

sequencesdiscussedinarticle

[TITL] Titleword

Onlywordsinthedefinitionline(notavailableinStructuredatabase)

[WORD] Text words

Allwordsandnumbersinthetitleandabstract,MeSHterms,

subheadings,chemicalsubstancenames,personalnameassubject,

andMEDLINEsecondarysources

[VOL] Volumeofjournal

andBooleanoperator =

AND, OR,or NOT

Integrated Information Retrieval: The Entrez System 25

threeentriesmatchingthequerywerefoundinPubMed.Theusercanfurthernarrowdown

thequerybyaddingadditionaltermsiftheuserisinterestedinamorespecificaspectofthis

geneoriftherearequitesimplytoomanyentriesreturnedbytheinitialquery.Alistofavailable

fielddelimitersisgiveninTable2.1.

ForeachofthefoundpapersshownintheSummaryviewinFigure2.2,theuserispresented

with the title of the paper, the authors of that paper, and the citation. To look at any of the

papers resulting from the search, the user can simply click on any of the hyperlinked titles.

For this example, consider the third reference in the list, by Srour et al. (2010). Clicking on

thetitletakestheusertotheAbstractviewshowninFigure2.3.Thisviewpresentsthename

of the paper, the list of authors, their institutional affiliation, and the abstract itself. Below

theabstractisagraybarlabeled“MeSHterms,Substances”;clickingontheplussignatthe

endofthegraybarrevealscataloginginformation(MeSHterms,for medicalsubjectheadings)

and indexed substances related to the manuscript. Several alternative formats are available

for displaying this information, and these various formats can be selected using the Format

pull-downmenufoundintheupperleftcornerofthewindow.SwitchingtoMEDLINEformat

Figure 2.3 An example of a PubMed record in Abstract format, as returned through Entrez. This Abstract view is for the third reference

shown in Figure 2.2. This view provides connections to related articles, sequence information, and the full-text journal article through the

Discovery Column that runs down the right-hand side of the page. See text for details.

26 Information Retrieval from Biological Databases

produces the MEDLINE layout, with two-letter codes corresponding to the contents of each

fieldgoingdowntheleft-handsideoftheentry(e.g.theauthorfieldisagaindenotedbythe

code AU).Listsofentriesinthisformatcanbesavedtothedesktopandeasilyimportedinto

third-partybibliographymanagementprograms.

The column on the right-hand side of this window – aptly named the Discovery

Column – provides access to the full-text version of the paper and, more importantly,

containsmanyusefullinkstoadditionalinformationrelatedtothismanuscript.TheSimilar

articlessectionprovidesoneoftheentrypointsfromwhichtheusercantakeadvantageofthe

neighboring and hard link relationships described earlier and, in the examples that follow,

we will return to this page several times to illustrate a selected cross-section of the kinds of

information available to the user. To begin this journey, if the user clicks on theSeeall link

atthebottomofthatsection,Entrezwillreturnalistof104referencesrelatedtotheoriginal

Rouleaupaperatthetimeofthiswriting;thefirstsixofthesepapersareshowninFigure2.4.

ThefirstpaperinthelististhesameRouleaupaperbecause,bydefinition,itismostrelatedto

itself(the“parent”entry).Theorderinwhichtherelatedpapersfollowisbasedonstatistical

Figure 2.4 Neighbors to an entry found in PubMed. The original entry from Figure 2.3 (Srour et al. 2010) is at the top of the list, indicating

that this is the parent entry. Additional neighbors to each of the papers in this list can be found by clicking the Similar articles link found

below each entry. See text for details.

Integrated Information Retrieval: The Entrez System 27

Figure 2.5 The Entrez Gene page for the DCC (deleted in colorectal carcinoma) netrin-1 receptor from human. The entry indicates that this

is a protein-coding gene at map location 18q21.2, and information on the genomic context of DCC, as well as alternative gene names and

information on the encoded protein, is provided. An extensive collection of links to other National Center for Biotechnology Information

(NCBI) and external databases is also provided. See text for details.

similarity. Thus, the entry closest to the parent is deemed to be the closest in subject matter

to the parent. By scanning the titles, the user can easily find related information on other

studies, as well as quickly amass a bibliography of relevant references. This is a particularly

usefulandtime-savingfunctionwhenoneiswritinggrantsorpapers,asabstractscaneasily

bescannedandpapersofrealinterestcanbeidentifiedquickly.

Returning to the Abstract view presented in Figure 2.3, at the bottom of the Discovery

Column is a series of hard-link connections to other databases within the Entrez system

thatcantaketheuserdirectlytoanextensivesetofinformationrelatedtothecontentofthe

publicationofinterest.Here,selectingtheGenelinktakestheusertoEntrezGene,afeatureof

Entrezthatprovidesawealthofinformationaboutthegeneinquestion(Figure2.5).Thedata

aregatheredfromavarietyofsources,includingRefSeq.Here,weseethatDCCistheofficial

symbol of a protein-coding gene for a netrin-1 receptor in humans. The Genomic context

sectionofthispageindicatesthattheDCCisaprotein-codinggeneatmaplocation18q21.2.

28 Information Retrieval from Biological Databases

Immediately below, summary information on the genomic region, transcripts, and products

of theDCC gene are presented graphically, with genomic coordinates provided. Additional

content not shown in the figure can be found by scrolling down the Gene page, where the

user will find relevant functional information (such as gene expression data), associated

phenotypes,informationonprotein–proteininteractions,pathwayinformation,GeneOntol-

ogyassignments,andhomologiestosimilarsequencesinselectedorganisms.Shortcutlinksto

thesesectionscanbefoundintheTableofcontentsatthetopoftheDiscoveryColumn.Fur-

therdowntheDiscoveryColumnareextensivelistsoflinkstoadditionalresourcesprovided

throughNCBIandothersources.Onelinkofnoteisthe SNP:GeneView link,takingtheuser

todataderivedfromdbSNP(Figure2.6).TheinformationfoundwithindbSNPgoesbeyond

justsingle-nucleotidepolymorphisms(SNPs),includingdataonshortgeneticvariationssuch

as short insertions and deletions, short tandem repeats, and microsatellites. Here, we will

focus on the table shown in Figure 2.6, which is a straightforward way to view information

aboutindividualSNPs.EachSNPentryoccupiestwoormorelinesofthetable,withoneline

showingthecontigreference(themorecommonallele)andtheothershowingtheSNP(the

Figure 2.6 A section of the Database of Single Nucleotide Polymorphisms (dbSNP) GeneView page providing information on each SNP

identified within the human DCC gene. See text for details.

Integrated Information Retrieval: The Entrez System 29

lesscommonallele).Considerthefirstthreelinesofthetable,showingacontigreferenceGfor

whichtherearetwodocumentedSNPs,changingtheGatthatpositiontoeitheranAoraC.At

theproteinlevel,thischangestheaminoacidatposition2oftheDCCproteinfromglutamic

acidtolysine(fortheG-to-Asubstitution)ortoglutamine(fortheG-to-Csubstitution).These

rowsarecoloredredsincetheseare“non-synonymousSNPs”–thatis,theSNPproducesadis-

cretechangeattheaminoacidlevel.Incontrast,considerthefirstsetofgreenrowsinthetable,

with the green indicating that this is a “synonymous SNP,” where the codons for the contig

reference(G)andtheSNPallele(A)ultimatelyproducethesameaminoacid(Glu);thisisnot

altogethersurprising,withtheSNPbeinginthewobblepositionofthecodon,wherethereis

oftenredundancyinthegeneticcode.AdditionalinformationonhumanSNPscanbefoundin

Chapter15.

StartingagainfromtheAbstractviewshowninFigure2.3,proteinsequencesfromRefSeq

that have been linked to this abstract can be found by clicking on theProtein (RefSeq)link

found in the Related information section on the right-hand side of the page, producing the

viewshowninFigure2.7.Notethatallbutoneoftheentriesismarkedas“predicted”;thefinal

Figure 2.7 Entries in the RefSeq protein database corresponding to the original Srour et al. (2010) entry in Figure 2.3. Entries can be

accessed and examined by clicking on any of the accession numbers. See text for details.

30 Information Retrieval from Biological Databases

Figure 2.8 The RefSeq entry for the netrin receptor, the protein product of the human DCC gene. The FASTA link at the top of the entry

provides quick access to the protein sequence in FASTA format, while the Graphics link provides access to a graphical view of all of the

individual elements captured within the entry’s feature table (see Figure 2.9). See text for details.

entry in the list has an accession number beginning with NP, indicating that it contains an

experimentallydeterminedorverifiedsequence(seeBox1.2).Clickingonthefirstlineofthat

entry(number6)takestheusertotheviewshowninFigure2.8,theRefSeqentryforthenetrin

receptor,theproteinproductofthe DCCgene.Thefeaturetable–thesectionoftheGenBank

entry listing the location and characteristics of each of the documented biological features

found within this protein sequence, such as post-translational modifications, recognizable

repeat units, secondary structural regions, and clinically relevant variation – is particularly

long in this case. This makes it difficult to determine the relative orientation of the features

tooneanotherandmayleadtheusertomissimportantinteractionsorrelationshipsbetween

biologicalfeatures.Fortunately,aviewerthatprovidesabird’seyeviewoftheelementsfound

within the feature table is available by clicking on the Graphics link at the top of the entry,

producingthemoreaccessibledisplayshowninFigure2.9.Zoomcontrolsareprovided,and

Integrated Information Retrieval: The Entrez System 31

Figure 2.9 The same RefSeq entry for the netrin receptor shown in Figure 2.8, now rendered in graphical format. The user can learn more

about individual elements displayed in this view by simply hovering the cursor over any of the elements in the display; one such example

is shown in the pop-up box at the bottom right, for the phosphorylation site at position 1267 of the sequence. Zoom and navigational

controls are at the top of the view window, allowing the user to understand this gene within its broader genomic context.

hovering over any of the elements in the display produces a pop-up containing the specific

informationforthatfeaturefromtheGenBankentry.

Fromhere,theusercanalsoenterthestructuralrealmbyexaminingtheproteinstructures

that are available through the Discovery Column. Clicking on theSee all 9 structureslink

takestheusertotheviewshowninFigure2.10,listingstructuralentriesrelatedtothenetrin

receptor.Thesecondentryisforthecrystalstructureofafragmentofnetrin-1complexedwith

theDCCreceptor(PDB:4URT;Fincietal.2014),andclickingonthetitleofthatentrytakesthe

usertothestructuresummarypageshowninFigure2.11.Startingontheright,theInterac-

tionswindowshowstherelationshipsbetweentheindividualelementsinthisbiologicalunit,

hereconsistingofthenetrin-1protein(circleA),theDCCreceptor(circleB),andfivedifferent

chemicalentities(diamonds1–5).Thethree-dimensionalstructureisshownintheleftpanel,

32 Information Retrieval from Biological Databases

Figure 2.10 Protein structures associated with the RefSeq entry for the human netrin receptor shown in Figures 2.8 and 2.9. The description

of each structure is hyperlinked, allowing the user to access the structure summary page for that entry (see Figure 2.11). Individual links

below each entry allow quick access to related structures and proteins, information on conserved domains, and the iCn3D viewer.

and the structure can be further interrogated by clicking on the square with the diagonal

arrowinthebottomleftofthatpanel.ThisactionwilllaunchiCn3D(for“Iseeinthree-D”),

a web-based viewer that allows the structure to be rotated, provides coloring and rendering

optionstoenhancevisualization,andprovidesawidevarietyofadditionaloptions;thereader

is referred to the iCn3D online documentation for specifics. In the upper right of the 4URT

structuresummarypageisalinktosimilarstructures,asdeterminedbyVAST +.Clickingon

the VAST+link produces the output shown in Figure 2.12, here showing the first 10 of 256

structuresdeemedtohavesimilarbiologicalunitstothequery(4URT);thetableshownhere

issortedbyRMSDofallalignedresidues(inÅ),fromsmallesttolargest.

===== PDF page 53 (Figure 2.11 caption portion) =====

Medical Databases 33

Figure 2.11 The structure summary page for pdb:4URT , the crystal structure of a fragment of netrin-1 complexed with the DCC receptor

(Finci et al. 2014). The entry shows header information from the corresponding Molecular Modeling Database (MMDB) entry, a link to the

paper reporting this structure, and the methodology used to determine this structure (here, X-ray diffraction with a resolution of 3.1 Å).

S e et e x tf o rd e t a i l s .

Text-Specific Search Options: Limits and Advanced Search

实际小节名：Medical Databases

用途：

Medical Databases 33

Figure 2.11 The structure summary page for pdb:4URT , the crystal structure of a fragment of netrin-1 complexed with the DCC receptor

(Finci et al. 2014). The entry shows header information from the corresponding Molecular Modeling Database (MMDB) entry, a link to the

paper reporting this structure, and the methodology used to determine this structure (here, X-ray diffraction with a resolution of 3.1 Å).

S e et e x tf o rd e t a i l s .

Medical Databases

Althoughthefocusofmanyinvestigatorsisonsequence-baseddata,databasecatalogingand

organizingsequenceinformationarenottheonlykindsofdatabasesusefultothebiomedical

research community. An excellent example of such a database that is tremendously useful

in genomicsis calledOnlineMendelianInheritancein Man(OMIM), the electronicversion

of the venerable catalog of human genes and genetic disorders originally founded by Victor

McKusick and first published in 1966 (McKusick 1966, 1998; Amberger et al. 2014). OMIM,

34 Information Retrieval from Biological Databases

Figure 2.12 A list of structures deemed similar to pdb:4URT using VAST +. The table is sorted by the root-mean-square deviation of all

aligned residues (in Å), from smallest to largest. Details on each individual structure in the list can be found by clicking on its Protein Data

Bank (PDB) ID number.

which is authored and maintained at The Johns Hopkins University School of Medicine,

providesconcisetextualinformationfromthepublishedliteratureonmosthumanconditions

havingageneticbasis,aswellaspicturesillustratingtheconditionordisorder(whereappro-

priate), full citation information, and links to a number of useful external resources, some

of which will be described below. As will become obvious through the following example, a

basicknowledgeofOMIMshouldbepartofthearmamentariumofphysician-scientistswith

aninterestintheclinicalaspectsofgeneticdisorders.

OMIMhasadefinednumberingsysteminwhicheachentryisassignedauniquenumber–a

“MIM number” – that is similar to an accession number, with certain positions within that

number indicating information about the genetic disorder itself. The first digit represents

the mode of inheritance of the disorder: 1, 2, and 6 stand for autosomal loci or phenotypes,

3 for X-linked loci or phenotype, 4 for Y-linked loci or phenotype, and 5 for mitochondrial

lociorphenotypes.Anasterisk(*)precedingaMIMnumberindicatesagene,ahashsign(#)

indicates an entry describing a phenotype, a plus sign (+) indicates that the entry describes

Medical Databases 35

Figure 2.13 Online Mendelian Inheritance in Man (OMIM) entries related to the DCC gene. The hash sign (#) preceding the first entry

indicates that it is an entry describing a phenotype – here, mirror movements. The second entry is preceded by an asterisk (*), indicating

that it is a gene entry – here, for the DCC gene.

a gene of known sequence and phenotype, and a percent sign (%) describes a confirmed

Mendelian phenotype or locus for which the underlying molecular basis is unknown. If no

Mendelian basis has been clearly established for a particular entry, no symbol precedes the

MIMnumber.

Here,wewillcontinuetheEntrezexamplefromtheprevioussection,followingthe OMIM

(cited)linkfoundintheDiscoveryColumnshowninFigure2.3.Anintermediatelandingpage

will then appear listing two entries, one for theDCC gene, the other for a phenotype entry

describingmirrormovements(Figure2.13).Clickingonthesecondentryleadstheusertothe

OMIMpageforthe DCCgeneshowninFigure2.14,withtheTextsectionoftheentryproviding

acomprehensiveoverviewofseminaldetailsregardingtheidentificationofthegene,itsstruc-

ture,relevantbiochemicalfeatures,mappinginformation,anoverviewofthegene’sfunction

andmoleculargenetics,andstudiesinvolvinganimalmodels.Forindividualsstartingworkon

anewgeneorgeneticdisorder,thisexpertlycuratedsectionoftheOMIMentryshouldbecon-

sidered“requiredreading,”asitpresentsthemostimportantaspectsofanygivengene,with

36 Information Retrieval from Biological Databases

Figure 2.14 The Online Mendelian Inheritance in Man (OMIM) entry for the DCC gene. Each entry in OMIM includes information such as

the gene symbol, alternative names for the disease, a description of the disease, a clinical synopsis, and references. See text for details.

links to the original studies cited within the narrative embedded throughout. A particularly

usefulfeatureisthelistofallelicvariants(Figure2.15);ashortdescriptionisgivenaftereach

allelicvariantoftheclinicalorbiochemicaloutcomeofthatparticularmutation.Atthetime

of this writing, there are over 5200 OMIM entries containing at least one allelicvariant that

eithercausesorisassociatedwithadiscretephenotypeinhumans.Notethattheallelicvari-

ants shown in Figure 2.15 produce significantly different clinical outcomes – two different

types of cancer as well as the motor disorder used throughout this example – an interesting

casewheredifferentmutationsinthesamegeneleadtodistinctgeneticdisorders.

The studies leading to these and similar observations described in a typical entry often

provide the foundation for clinical trials aimed at translating this knowledge into new

preventionandtreatmentstrategies.NIH’scentralinformationsourceforclinicaltrials,aptly

named ClinicalTrials.gov, contains data on both publicly and privately funded clinical trials

Medical Databases 37

Figure 2.15 An example of a list of allelic variants that can be found through Online Mendelian Inheritance in Man (OMIM). The figure

shows three of the four allelic variants for the DCC gene. T wo of the documented variants lead to cancers of the digestive tract, while

two are associated with a movement disorder. The description under each allelic variant provides information specific to that particular

mutation.

beingconductedworldwide.Figure2.16showsthefirsteightofmorethan4600clinicaltrials

actively recruiting patients with colorectal cancer at the time of this writing, and clicking

on the name of a protocol will bring the user to a page providing information on the study,

includingtheprincipalinvestigator’snameandcontactinformation.Clickingthe OnMap tab

atthetopofthepageproducesaclickablemapoftheworldshowinghowmanyclinicaltrials

arebeingconductedineachregionorcountry(Figure2.17);thisviewisusefulinidentifying

trialsthataregeographicallyclosetoapotentialstudysubject’shome.Whilewe,asscientists,

tend to focus on the types of information discussed throughout the rest of this chapter, the

clinical trials site is, unarguably, the most important of the sites covered in this chapter, as

it provides a means through which patients with a given genetic or metabolic disorder can

38 Information Retrieval from Biological Databases

Figure 2.16 The ClinicalT rials.gov page showing all actively recruiting clinical trials relating to colorectal neoplasms. Information on each

trial, including the principal investigator of the trial and qualification criteria for participating in the trial, can be found by clicking on the

name of the trial.

receivethelatest,cutting-edgetreatment–treatmentthatmaymakeasubstantialdifference

totheirqualityoflife.

Organismal Sequence Databases Beyond NCBI

Although it may appear from this discussion that NCBI is the center of the sequence uni-

verse,manyspecializedgenomicdatabasesthroughouttheworldservespecificgroupsinthe

scientificcommunity.Often,thesedatabasesprovideadditionalinformationnotavailableelse-

where,suchasphenotypes,experimentalconditions,straincrosses,andmapfeatures.These

dataareofgreatimportancetothesecommunities,notonlybecausetheyinfluenceexperimen-

tal design and interpretation of experimental results but also because the kinds of data they

containdonotalwaysfitneatlywithintheconfinesoftheNCBIdatamodel.Developmentof

specialized databases necessarily ensued (and continues), and these databases are intended

tobeusedasanimportantadjuncttoGenBankandsimilarglobaldatabases.Itisimpossible

38 Information Retrieval from Biological Databases

Figure 2.16 The ClinicalT rials.gov page showing all actively recruiting clinical trials relating to colorectal neoplasms. Information on each

trial, including the principal investigator of the trial and qualification criteria for participating in the trial, can be found by clicking on the

name of the trial.

receivethelatest,cutting-edgetreatment–treatmentthatmaymakeasubstantialdifference

totheirqualityoflife.

Organismal Sequence Databases Beyond NCBI

Although it may appear from this discussion that NCBI is the center of the sequence uni-

verse,manyspecializedgenomicdatabasesthroughouttheworldservespecificgroupsinthe

scientificcommunity.Often,thesedatabasesprovideadditionalinformationnotavailableelse-

where,suchasphenotypes,experimentalconditions,straincrosses,andmapfeatures.These

dataareofgreatimportancetothesecommunities,notonlybecausetheyinfluenceexperimen-

tal design and interpretation of experimental results but also because the kinds of data they

containdonotalwaysfitneatlywithintheconfinesoftheNCBIdatamodel.Developmentof

specialized databases necessarily ensued (and continues), and these databases are intended

tobeusedasanimportantadjuncttoGenBankandsimilarglobaldatabases.Itisimpossible

Organismal Sequence Databases Beyond NCBI 39

Figure 2.17 A clickable map showing where actively recruiting clinical trials relating to colorectal neoplasms are being conducted. This

map-based view of the information presented in Figure 2.15 is useful in identifying trials that are within a reasonable distance of a potential

study participant’s home.

to discuss the wide variety of such value-added databases here but, to emphasize the sheer

numberofsuchdatabasesthatexist,thejournal NucleicAcidsResearch devotesitsfirstissue

everyyeartopapersdescribingthesedatabases(Galperinetal.2017).

An excellent representative example of a specialized organismal database is the Mouse

Genome Database (MGD; Bult et al. 2016). Housed at the Jackson Laboratory in Bar Har-

bor, ME, the MGD provides a curated, comprehensive knowledgebase on the laboratory

mouse and is an integral part of its overall Mouse Genome Informatics (MGI) resource.

The MGD provides information on genes, genetic markers, mutant alleles and phenotypes,

and homologies to other organisms, as well as extensive linkage, cytogenetic, genetic, and

physical mapping data. A cross-section of these data is shown in Figure 2.18, providing

information on theDcc gene in mouse, the ortholog to the humanDCC gene from the

examples discussed earlier in this chapter. This page can be accessed either by directly

searching for the gene name or, in this case, by following links found within the Animal

Model section of the OMIM entry for DCC discussing seminal discoveries made using

40 Information Retrieval from Biological Databases

Figure 2.18 The Mouse Genome Informatics (MGI) entry for the Dcc gene in mouse. The entry provides information on the ortholog to the

human DCC gene, including data on mutant alleles and phenotypes, mapping data, single-nucleotide polymorphisms, and expression data.

In the Mutations, Alleles, and Phenotypes section, the phenotype overview uses blue squares to indicate which phenotypes are due to

mutations in the Dcc gene. In the Expression section, blue squares indicate expression in wild-type mice in the designated tissues, organs,

or systems.

mouse models that, in turn, informed our understanding of the effect ofDCC mutations

inhumans.

Another long-standing resource devoted to a specific organism is the Zebrafish Model

Organism Database, or Zebrafish Information Network (ZFIN) (Howe et al. 2012) – a

particularly attractive animal model given the experimental tractability of zebrafish in

studying a wide variety of questions focused on vertebrate development, regeneration,

inflammation, infectious disease, and drug discovery, to name a few. ZFIN provides a very

simple search interface that allows free-text searches using any term. UsingDCC once

again as our search term brings the user to the summary page for the zebrafishdcc gene

Organismal Sequence Databases Beyond NCBI 41

Figure 2.19 The Zebrafish Information Network (ZFIN) gene page for the dcc gene in zebrafish. This entry provides information on the

ortholog to the human DCC gene. See text for details.

(Figure 2.19), providing information on zebrafish mutants, sequence targeting reagents,

transgenic constructs, orthology to other organisms, data on protein domains found within

the Dcc protein product, and annotated gene expression and phenotype data derived from

the literature or from direct submissions by members of the zebrafish community. Here,

by following the link to the 19 figures in the Gene Expression section, one can examine

full-sizeimagesillustratingexpressionpatternsfor dccundervariousexperimentalconditions

(Figure2.20).

While MGD and ZFIN are excellent examples of model organism databases, every major

model organism community maintains such a resource. These groups also collaborate to

developcentralportalstoeaseinformationretrievalacrossmanyoftheseresourcesthrough

theAllianceofGenomeResources.

42 Information Retrieval from Biological Databases

Figure 2.20 An example of gene expression data available through the Zebrafish Information Network (ZFIN), here showing the expression

patterns for the zebrafishdcc gene under various experimental conditions. The inset displays a full-size image of data from Gao et al. (2012),

showing the expression pattern for dcc in (panel A) and the co-expression of dcc and the Lim homeobox 5 gene ( lhx5; panel B).

Summary

As alluded to in the introduction to this chapter, the information space available to investi-

gatorswillcontinuetoexpandatbreakneckspeed,withthesizeofGenBankalonedoubling

every year. Although the sheer magnitude of data can present a conundrum to the inexpe-

rienced user, mastery of the techniques covered in this chapter will allow researchers in all

biologicaldisciplinestomakethebestuseofthesedata.Themovementofmodernscienceto

“big data” approaches underscores the idea that both laboratory and computationally based

strategieswillbenecessaryincarryingoutcutting-edgeresearch.Inthesamewaythatinves-

tigatorsaretrainedin,forexample,basicbiochemistryandmolecularbiologymethodologies,

a basic understanding of bioinformatic techniques as part of the biologist’s arsenal will be

indispensable in the future. As is undoubtedly apparent at this point, there is no substitute

for placing one’s hands on the computer keyboard to learn how to search and use genomic

sequencedataeffectively.Readersarestronglyencouragedtotakeadvantageoftheresources

References 43

presentedhere,togrowinconfidenceandcapabilitybyworkingwiththeavailabletools,and

tobegintoapplybioinformaticmethodsandstrategiestowardadvancingtheirownresearch

interests.

Internet Resources

AllianceofGenomeResources www.alliancegenome.org

BasicLocalAlignmentSearchTool(BLAST) ncbi.nlm.nih.gov/BLAST

ClinicalTrials.gov clinicaltrials.gov

DNADataBankofJapan(DDBJ) www.ddbj.nig.ac.jp

EuropeanMolecularBiologyLaboratory–European

BioinformaticsInstitute(EMBL-EBI)

www.ebi.ac.uk

GenBank www.ncbi.nlm.nih.gov/genbank

iCn3D www.ncbi.nlm.nih.gov/Structure/

icn3d/docs/icn3d_about.html

MouseGenomeDatabase(MGD) informatics.jax.org

OnlineMendelianInheritanceinMan(OMIM) omim.org

ProteinDataBank(PDB) www.rcsb.org/pdb

RefSeq ncbi.nlm.nih.gov/refseq

SingleNucleotidePolymorphismDatabase(dbSNP) www.ncbi.nlm.nih.gov/SNP

VectorAlignmentSearchTool(VAST) www.ncbi.nlm.nih.gov/Structure/VAST

ZebrafishInformationNetwork(ZFIN) zfin.org

Further Reading

Baxevanis,A.D.(2012).SearchingOnlineMendelianInheritanceinMan(OMIM)forinformation

ongeneticlociinvolvedinhumandisease. Curr.Protoc.Hum.Genet. Chapter9,Unit9.13.1–10.

Aprotocol-drivendescriptionofthebasicmethodologyforformulatingOMIMsearchesanda

discussionofthetypesofinformationavailablethroughOMIM,includingdescriptionsof

clinicalmanifestationsresultingfromgeneticabnormalities.

Galperin,M.Y.,Fernández-Suárez,X.M.,andRigden,D.J.(2017).The24thannual NucleicAcids

Researchdatabaseissue:alookbackandupcomingchanges. NucleicAcidsRes. 45:D1–D11.A

curated,annualreviewofspecializeddatabasesofinterestandimportancetothebiomedical

researchcommunity.

References

Altschul,S.,Gish,W.,Miller,W.etal.(1990).Basiclocalalignmentsearchtool. J.Mol.Biol. 215:

403–410.

Amberger,J.S.,Bocchini,C.A.,Schiettecatte,F.etal.(2014).OMIM.org:OnlineMendelian

InheritanceinMan,anonlinecatalogofhumangenesandgeneticdisorders. NucleicAcidsRes.

43:D789–D798.

Benson,D.A.,Cavanaugh,M.,Clark,K.etal.(2017).GenBank. NucleicAcidsRes. 45:D37–D42.

Bult,C.J.,Eppig,J.T.,Blake,J.A.etal.(2016).Mousegenomedatabase2016. NucleicAcidsRes. 44:

D840–D847.

Collins,F.S.,Patrinos,A.,Jordan,E.etal.,andMembersoftheDOEandNIHPlanningGroups

(1998).NewgoalsfortheU.S.HumanGenomeProject:1998–2003. Science.282:682–689.

Collins,F.S.,Green,E.D.,Guttmacher,A.E.,andGuyer,M.S.,onbehalfoftheU.S.National

HumanGenomeResearchInstitute(2003).Avisionforthefutureofgenomicsresearch.

Nature.422:835–847.

44 Information Retrieval from Biological Databases

Finci,L.I.,Krüger,N.,Sun,X.etal.(2014).Thecrystalstructureofnetrin-1incomplexwithDCC

revealsthebifunctionalityofnetrin-1asaguidancecue. Neuron.83:839–849.

Galperin,M.Y.,Fernández-Suárez,X.M.,andRigden,D.J.(2017).The24thannual NucleicAcids

Researchdatabaseissue:alookbackandupcomingchanges. NucleicAcidsRes. 45:D1–D11.

Gao,J.,Zhang,C.,Yang,B.etal.(2012).Dccregulatesasymmetricoutgrowthofforebrainneurons

inzebrafish. PLoSOne. 7:e36516.

Gibrat,J.-F.,Madej,T.,andBryant,S.(1996).Surprisingsimilaritiesinstructurecomparison. Curr.

Opin.Struct.Biol. 6:377–385.

Green,E.D.andGuyer,M.S.,andTheNationalHumanGenomeResearchInstitute(2011).

Chartingacourseforgenomicmedicinefrombasepairstobedside. Nature.470:204–213.

Howe,D.G.,Bradford,Y.M.,Conlin,T.etal.(2012).ZFIN,theZebrafishModelOrganism

Database:increasedsupportformutantsandtransgenics. NucleicAcidsRes. 41:D854–D860.

InternationalHumanGenomeSequencingConsortium(2001).Initialsequencingandanalysisof

thehumangenome. Nature.409:860–921.

Madej,T.,Lanczycki,C.J.,Zhang,D.etal.(2014).MMDBandVAST +:trackingstructural

similaritiesbetweenmacromolecularcomplexes. NucleicAcidsRes. 42:D297–D303.

McKusick,V.A.(1966). MendelianInheritanceinMan:CatalogsofAutosomalDominant,

AutosomalRecessive,andX-LinkedPhenotypes .Baltimore,MD:TheJohnsHopkinsUniversity

Press.

McKusick,V.A.(1998). OnlineMendelianInheritanceinMan:CatalogsofHumanGenesand

GeneticDisorders,12e.Baltimore,MD:TheJohnsHopkinsUniversityPress.

Schmutz,J.,Wheeler,J.,Grimwood,J.etal.(2004).Qualityassessmentofthehumangenome

sequence.Nature.429:365–368.

Srour,M.,Rivière,J.B.,Pham,J.M.T.etal.(2010).MutationsinDCCcausecongenitalmirror

movements.Science.328:592.

Wilbur,W.J.andCoffee,L.(1994).Theeffectivenessofdocumentneighboringinsearch

enhancement.Inf.Process.Manag. 30:253–266.

Wilbur,W.J.andYang,Y.(1996).Ananalysisofstatisticaltermstrengthanditsuseintheindexing

andretrievalofmolecularbiologytexts. Comput.Biol.Med. 26:209–222.

This chapter was written by Dr. Andreas D. Baxevanis in his private capacity. No official support

or endorsement by the National Institutes of Health or the United States Department of Health

and Human Services is intended or should be inferred.