ABSTRACT
The Mouse Genome Database (MGD) is a comprehensive community resource of mouse genetic and biological information populated both with data from published literature and with data electronically submitted from the research community. MGD stores genetic, physical and comparative mapping data, clones/probes/PCR information, and phenotype descriptions for genes, mutations and mouse strains. Supporting software for importation, analysis, display and distribution of mouse genetic data have been developed. User support is provided through dedicated staff providing documentation, training, and response to individual user queries. MGD is accessible over the Internet at URL http://www.informatics.jax.org.
INTRODUCTION
The Mouse Genome Informatics project at the Jackson Laboratory (TJL) is developing and maintaining the Mouse Genome Database (MGD), a comprehensive integrated database for genetic and biological data on the mouse. The focus of MGD is to represent reference-supported elemental genomic data suitable for analysis and integration with other biological data, and to provide easy-to-use electronic interfaces for the scientific community to access these data. MGD is one of several Mouse Genome Informatics (MGI) projects underway at the Jackson Laboratory (TJL). Other projects include the Gene Expression Database (1 ) and the Encyclopedia of the Mouse Genome, a graphical display tool for representing mouse genetic maps. User support is provided for all TJL bioinformatics projects.
BACKGROUND
MGD was first released on the World Wide Web (WWW) in June of 1994. The initial release incorporated data structures and information previously contained in independent, task-specific databases maintained at TJL. Among these was GBASE (Genome Database of the Mouse) (2 ), an entry interface into several databases holding locus mapping data (LocusBASE), gene descriptions (Mouse Locus Catalog) and a strain-by-allele table (MATRIX) that permitted researchers to compare strain differences or similarities across a series of loci. Other separate databases at TJL that were incorporated into MGD included HMDP, containing homologies and comparative mapping data; MLDP, containing experimental mapping data and analysis programs; and MusProb and MusPCR, containing probes, PCR primers and clone information.
MGD continues to reflect, to some extent, the historical legacy of all the databases incorporated during the initial implementation (3 ). Much of the redundancy has been removed and efforts continue to replace remaining textual descriptions with more controlled vocabularies and syntax. Additionally, the scope of the database has been expanded to include multiple links to other biological databases and to enhance the representation of cytogenetic, genetic and physical maps.
SCOPE
The Mouse Genome Database project is based upon the premise that holding primary research data is essential to the long-term use and understanding of the genetics and biological information described by the data. Elemental data permits re-examination and analysis of the data within the context of new information. In addition, conflicting data can be maintained and represented to the research community. To this end, all the data contained within the database are associated with their primary data source, a publication or electronic reference. Data acquisition currently focuses on information pertaining to genetic markers, experimental mapping data, clones/probe/PCR primer data, mammalian homology determinations, bibliographic references, gene descriptions and phenotype/strain characteristics. Each of these areas is integrated with the others as described below.
In addition to the annotation and accumulation of data pertaining to the genetics of the mouse, MGD strives to provide suitable user interfaces for the recovery and analysis of the data. Genetic maps are interactively displayed using the Encyclopedia of the Mouse Genome graphic display tool that can be invoked within the Web interface. Map Manager, a Macintosh application accessible via MGD, is a tool for statistical analysis of genetic mapping experiments using backcrosses, intercrosses, or recombinant inbred (RI) strains. It includes tools for statistical analysis of the experimental results (4 ).
The MGD Home Page
The MGD Home Page provides several points of entry into the database (Fig. 1 ). The user chooses an area of the database to initiate a search. The home page also connects to other resources available from this site. The query forms return a summary of results that then link to more detailed information.
Figure 1 .
MGD home page. Upper Left, Mouse Genome Informatics Home PAge, Lower front, Mouse Genome Database Home PAge.
Data sources
Data for the Mouse Genome Database come from multiple sources. The majority of the references are obtained through regular scanning of ~350 scientific journals. Relevant papers are added to MGD's bibliography and submitted to data editors for annotation and entry into the database. Additional references of interest are obtained from searches of Current Contents and MEDLINE, and from editorial requests. As of July, 1996, MGD contained >32 000 citations from ~1400 different journals. Roughly 500 -600 references are added to MGD each month.
We continue to develop better mechanisms for electronic submission of data sets. For example, haplotype datasets from a mapping experiment can be electronically submitted to MGD and the accession number for the dataset published with a summary of results in the subsequent scientific article. Currently, electronic data submissions are either the sporadic data submission of small data sets in conjunction with a publication, or the transfer of large data sets from individual laboratories maintaining large mapping panels that are used by the scientific community. Electronic data submissions undergo a series of internal checks before being posted.
Chromosome Committee Reports
Additional data sources for MGD include external annotations, most notably the annual Chromosome Committee Reports. The Chromosome Committees consist of mouse genetic researchers with special interest in a particular chromosome. Each Chromosome Committee submits an annual electronic report that includes a list of genes and analyzed map locations for that chromosome. Also included are known human-mouse homologies, additional data concerning gene families and discussion of genome organization for the chromosome. The complete reports are posted on the WWW with hypertext links to MGD data.
MGD accession numbers
Each entity in MGD has a permanent, unique accession number. All entities are supported by links to the reference paper, as well as to other information in the database. In collaboration with the I.M.A.G.E. Consortium and the Washington University /Howard Hughes Medical Institute mouse EST project, MGD accession numbers are being assigned to all I.M.A.G.E. clones and these are subsequently associated with dbEST mouse ESTs. In addition, MGD project leaders continue, in collaboration with the managers of other databases including the major sequence databases, to support appropriate links between biological database resources.
At the time of submitting a manuscript for publication, scientists can submit data files directly to MGD to obtain accession numbers that can then be included in the publication. This allows direct submission of data to the database with subsequent links to the citation and other information. This implementation is being encouraged by several journals, and instructions are included in their Instructions to Authors.
Genes, markers and alleles
All genetic entities that have been mapped in the mouse genome
are assigned unique symbols and names (e.g. Hba , Fig. 2 ; Pgm1 ,
Lepr , Tyrp1 , Lama2 ). Genetic marker designations are applied to
genes, phenotypic mutations, chromosomal aberrations, anonymous DNA segments, and quantitative trait loci (QTLs). Over 20 000 valid symbols are represented in MGD, including ~12 000 mapped anonymous loci and >5000 mapped genes. Of the >7000 genes identified in the database, >4500 are represented by molecular probes or clones. In addition, many alleles are reported on in both molecular and phenotypic terms.
Figure 2 .
Genetic Marker Web report for Hba . The Genetic Marker Information page details include alleles, history of the use of the symbol, links to experiments involving this marker, references, and information about mapped location.
Alleles and polymorphisms
Alleles are alternate forms of a gene occupying the same locus on a particular chromosome. MGD maintains records of alleles and uses standardized allelic symbols to differentiate between them. Polymorphisms as would be revealed by RFLPs and PCR length variants and phenotypic alleles are emphasized, although increasing representation of other gene mutations and biochemical alleles is anticipated. Mutant phenotype descriptions are converted into allele descriptions when the gene for the mutant phenotype is cloned. The distribution of polymorphic alleles by strain has been the focus on previous work at TJL, and these associations will be more fully represented in future releases of MGD. Over 1l00 alleles representing >500 genes are described in MGD.
Gene descriptions
The Mouse Locus Catalog (MLC) presents concise descriptions of genes, their transcripts and alleles, and any expression and phenotypic information that is available (Fig. 3 ). Descriptions also detail some of the history of the discovery of the gene and discussion of any conflicting data. These gene descriptions are presented as text and graphical representations of mice showing the different phenotypes ( i.e. manifestion of the mutation). We continue to incorporate an increasingly structured syntax to facilitate links to other curated information about the locus. Over 5000 of the genetic loci in MGD are represented by MLC descriptions.
Figure 3 .
Mouse Locus Catalog entry for Hba . The Locus Catalog page contains a description of the gene including links to information about other related genes.
Nomenclature and synonyms
Each genetic marker is provided with a unique symbol and name that can be used for identifying, comparing and linking multiple sources of information about the marker. These symbols are assigned by the MGD nomenclature committee which utilizes the guidelines and policies set by the International Committee on Standardized Genetic Nomenclature for Mice. In addition, aliases, synonyms, withdrawn symbols and names can be searched for through the MGD WWW query interface.
With the advent of new technologies that speed gene discovery, nomenclature issues have become a complex and detailed component of databases containing information about genes. Over 3000 new loci were added to the genetic marker listin MGD within the last year; >1000 of these were new genes (the rest being anonymous DNA markers). MGD works closely with other organism-based genetic databases, particularly GDB (Human Genome Database) (5 ) to coordinate symbol assignments for mouse and human genes.
Maps and mapping data
Experimental genetic mapping data are extensively represented in MGD (Fig. 4 ). Major categories of experimental mapping data include experimental crosses, in situ and FISH hybridization, recombinant inbred and recombinant congenic strains, somatic cell hybrid analyses and physical mapping data. Information specific to each data type is represented (e.g. cytogenetic band hybridized to for in situ hybridization experiments). Each experiment is described in terms of the materials/resources used, the method used for marker identification (e.g. Southern analysis, or visible phenotype), and the raw data mapping results (e.g. haplotypes for progeny from linkage crosses). MGD supports the generation of statistical analysis of mapping cross data and the results are displayed with the raw data.
Figure 4 .
Web Page of representative Experimental Mapping Data fo Hba . Details of a experimental cross shows a matrix of the haplotypes and statistical analysis of the results as well as information about the cross itself.
MGD currently includes >10 000 mapping experiments. Almost half of the data comes from experimental crosses. Haplotype data are recorded, as well as source information about the parental mice and type of cross employed in the experiment.
Mapping panels
Genetic mapping panels were first developed in the 1980s with the recognition that a high degree of DNA polymorphism exists between the laboratory mouse and interbreeding, closely related members of the Mus species group (Mus spretus and M.castaneus ) (6, 7 ). DNA isolated from the backcross or intercross offspring of interspecies matings provide a large resource against which multiple genes can be mapped. A dense genetic map created from a single progeny set emerges as increasing numbers of genes are mapped with a given set of DNAs. Datasets from several major mouse mapping panels have been downloaded into MGD and are accessible from a MGD query form. The six panels represented in MGD include The Jackson Laboratory BSS and BSB panels (8 ), EUCIB BSS and BSB panels (9 ), the Copeland-Jenkins (7 ) and Seldin (10 ) backcross panels and the MIT intercross panel (11 ). These panels receive update information from the researchers maintaining the panels on a regular basis.
Probes and clones
Clones, probes and PCR data are represented in MGD. These genomic segments are linked to the loci that they contain or hybridize to. Characteristics of the probe or clone (e.g. the species of origin, the type of segment such as cDNA or genomic, the insert size and vector, and the sequence database accession number) are recorded. This information is structured so that presentation of genetic and physical maps is facilitated. Over 21 000 probes and clones are recorded in MGD representing >12 500 distinct genetic markers. PCR records include the sequence of the primer pair as well as the characteristics of the amplified sequence and the sequence accession numbers, if applicable.
Mammalian homology
Homology associations between genes of different species are often used to find genes in related species, for studies of the evolution of nucleic acids and proteins, or for comparative genome analysis. Determination of homology in MGD is based on experimental analysis and requires citation support for the assertion (12, 13 ). For a given mammalian homology, MGD contains, for each species, the gene symbol, chromosomal location and supporting citations (Fig. 5 ). Over 50 mammalian species are represented in MGD. The homology definition used to define relationships conforms to the guidelines set forth by the Comparative Gene Mapping Committee criteria of HUGO (14 ). The data are obtained from the mouse Chromosome Committee reports, scientific literature and personal communications. Interactive Oxford Grids (15 ) displaying comparative mapping between two species are available for mouse, human, rat, cow, pig, sheep and cat. Results from queries feature hypertext links to other on-line species databases such as GDB (5 ), SheepGBASE (16 ), PiGBASE (17 ) and RatMAP (18 ). MGD will establish hypertext links to additional species databases as they become available.
Figure 5 . Web Page of Homology results for Hba . Details of the Mammalian Homology page include hotlinks to the gene records in other databases.
There are four different routes to retrieve homology data contained in MGD. (i) Searching by a gene symbol using the Marker Query form will retrieve a genetic marker screen that includes links to any mammalian homology for that gene. (ii) Through the mammalian Homology Query form, an Oxford Grid format for data retrieval can be invoked. Two species are chosen for comparison and a grid is retrieved that shows areas of genomic similarity. Links to specific homology records for regions of genomic similarity are maintained. (iii) Searching by using the Homology Query form retrieves one or more species comparisons for the chromosomal region specified by the user. (iv) Finally, the MGD Home Page contains a link to MGD Data and Statistical Reports. Reports are generated nightly and contain complete listings of either mouse/human homologies or mouse/rat homologies.
Database cross references
MGD supports hypertext links to other biological resources. In particular, links are supported between records of identified homologs between mouse and other mammalian species (as described above). Through the references, MGD links directly to Medline-Entrez. Probes and PCR amplification products associated with sequences contain links to the major sequence databases; NCBI GenBank (19 ), EMBL (20 ), DDBJ (21 ) and GSDB (22 ). An integrated OMIM/MLC searching capability is included in MGD so that one may enter a search phrase to execute a query in both MLC and Online Mendelian Inheritance in Man (OMIM) (23, 24 ).
COMMUNITY OUTREACH
User support
An important component of a community database is outreach to users through the development of user manuals and documentation tools, through demonstrations and training in the use of the database and through individual user assistance. The Mouse Genome Informatics user group includes biologists to assist users in answering biology-related questions and computer professionals to assist in answering technical questions. Individual user queries are received either by email, phone, fax or written correspondence. Announcements and PostScript technical documentation can be viewed on the WWW site and are available as PostScripttm files. Members of the bioinformatics staff regularly attend scientific meetings where they demonstrate MGD and provide training in the use of MGD and other Jackson Laboratory bioinformatics databases and tools.
Mailing lists
The MGI project maintains several electronic mailing lists to facilitate communication with the genome community (see below for contact information).
These lists use the ListProcessor software system. Mailing lists available to the community include `mgi-list', for announcements about bioinformatics at TJL, `mgd-list' for discussion about MGD, and `nomen-list' for mouse genome and genetic nomenclature discussions.
IMPLEMENTATION
MGD is implemented in the Sybase relational database system. We are currently running Sybase System 10.0.2, but will be moving shortly to Sybase System 11. The database consists of >150 tables in which information about each type of entity (e.g. a mapping experiment) is cross-linked to related information (e.g. the markers used) through unique identifiers. A large (and growing) set of stored procedures and triggers insures continued integrity of the data in the face of updates.
There are three major interfaces to MGD: editorial, WWW and SQL. The editorial interface is used by the MGI staff for entering and modifying database records. It is implemented with the TeleUSE user interface management system, a general-purpose tool for building X/Motif interfaces. The WWW interface comprises a set of static HTML forms and other supporting documents, and a large set of CGI (Common Gateway Interface) scripts that mediate the user's interaction with the database. Essentially, these scripts convert a completed query form or a selected hypertext link into an SQL query, execute the query, format the results as HTML, and return the result to the user. These scripts are written in Python, an object-oriented, interpreted language, available free on the Internet (http://www.python.org ).
Finally, users may access MGD directly via SQL. Users wishing an SQL account may contact MGD User Support.par
FUTURE DEVELOPMENT
Current development and enhancement of MGD includes the restructuring of the database to integrate MGD further with the Gene Expression Database (GXD) and to enhance the representation of physical mapping presentations. This reorganization will further integrate tables related to various constructs of genomic segments, and will support the display of physical mapping data. Procedures for facilitating direct electronic submission and updating of data sets are also being implemented. In addition to maintaining and updating MGD, future releases will enhance the representation of gene families in mice and enunciate the details of genetic markers implicated in complex traits. The distribution of alleles by strain, and complete strain descriptions and histories will be incorporated.
REFERENCING MGD
We suggest the following citation format when referring to MGD: Mouse Genome Database (MGD), Mouse Genome Informatics, The Jackson Laboratory, Bar Harbor, Maine. World Wide Web
(URL: http://www.informatics.jax.org/ ) (date when information retrieved).
To reference the database itself, please cite this article and refer to
others found at: http://www.informatics.jax.org/doc/citation.html
ADDRESSES
WWW URLs
MGD can be accessed at The Jackson Laboratory site (URL: http://www.informatics.jax.org/ ). Mirror sites have been established in Europe at the UK HGMP Resource Center (URL:
http://mgd.hgmp.mrc.ac.uk/ ) and in Ja pan at National Institute of Animal Industry/Japan Animal Genome Database
(http://mgd.niai.affrc.go.jp ). The mirror sites are updated weekly.
FTP
Chromosome committee reports, MGD Documentation, nomenclature guidelines and other MGD resources can be obtained at the Jackson Laboratory Informatics FTP site (ftp.informatics. jax.org/pub/ ).
User support
Email to: mgi-help@informatics.jax.org . In addition to email queries, user support staff can be reached at the following voice number: +1 207 288 6445 or FAX number: +1 207 288 6132.
Mailing lists
These include a list for announcements and general information exchange (mgi-list@informatics.jax.org) and a list specific to the Mouse Genome Database (mgd-list@informatics.jax.org). Users can subscribe electronically by sending a message such as (subscribe <list name> <your full name>) to `listserver@ informatics.jax.org' or by registering with the User Support Registration form (http://www.informatics.jax.org/doc/support.html ).
ACKNOWLEDGEMENT
The Mouse Genome Informatics project is supported by NIH grant P01 HG00330.
REFERENCES
1 Ringwald,M., Baldock,R., Bard,J., Eppig,J.T., Kaufmann,M., Nadeau,J.H., Richardson,J.E., Davidson,D. (1994 ) Science 265 2033 -2034.
2 Doolittle,D.P., Hillyard,A.L., Davisson,M.T., Roderick,T.H., Guidi,J.N. (1991 ) Fifth International Workshop on Mouse Genome Mapping
3 Richardson,J.E., Eppig,J.T., Nadeau,J.H. (1995 ) IEEE Eng. Med. Biol. 14 718 -724.
4 Manly,K. (1993 ) Mamm. Genet 4 303 -313.
5 Fasman,K.H., Letovsky,S.I., Cottingham,R.W., Kingsbury,D.T. (1996 ) Nucleic Acids Res. 24 57 -63.
6 Avner,P., Amar,L., Dandolo,L., Guenet,J.L. (1988 ) Trends Genet. 4, 18 -23.
7 Copeland,N.G., Jenkins,N.A. (1991 ) Trends Genet. 7 113 -118.
8 Rowe,L.B., Nadeau,J.H., Turner,R., Frankel,W.N., Letts,V.A., Eppig,J.T.,
Ko,M.S., Thurston,S.J., Birkenmeier,E.H. (1994 ) [published erratum
appears in (1994 ) Mamm. Genome 5 Mamm. Genome 5 253 -274.
9 European Backcross Collaborative Group (1994 ) Hum. Mol. Genet. 3 621 -627.
10 Seldin,M.F., Howard,T.A., D'Eustachio,P. (1989 ) Genomics 5 24 -28.
11 Dietrich,W., Katz,H., Lincoln,S.E., Shin,H.S., Friedman,J., Dracopoli,N.C.,
Lander,E.S. (1992 ) Genetics 131 423 -447.
12 Eppig,J.T., Nadeau,J.H. (1995 ) Curr. Opin. Genet. Dev. 5 709 -716.
13 Nadeau,J.H., Grant,P.L., Mankala,S., Reiner,A.H., Eppig,J.T. (1995 ) Nature 373 363 -365.
14 First International Workshop on Comparative Genome Organization (in press) Genomics .
15 Edwards ,J.H. (1991 ) Ann. Hum. Genet. 55 17 -31.
16 SheepBase at the New Zealand Pastoral Agriculture Research Institute (URL-http://dirk.invermay.cri.nz/docs/sheepgbase/manager.html ) and at the Roslin Institute, Edinburgh, UK (URL-http://www.ri.bbsrc.ac.uk/genome_mapping.html ).
17 PiGBase at the Roslin Institute , Edinburgh, UK (URL-http://www.ri.bbsrc.ac.uk/pigmap/pigbase/pigbase.html ) or at the National Agriculture Library, Beltsville, USA (URL-http://tetra.gig.usda.gov:8400/ pigbase/manager.html ).
18 RatMAP at Goteburg University , Goteburg, Sweden (URL-http://ratmap.gen.gu.se/ ).
19 Benson ,D.A., Boguski,M., Lipman,D.J., Ostell,J. (1996 ) Nucleic Acids Res. 24 1 -5
20 Rodriguez-Tome ,P., Stoehr,P.J., Cameron,G.N., Flores,T.P. (1996 ) Nucleic Acids Res. 24 6 -12.
21 DDBJ; DNA DataBase of Japan at the National Institute of Genetics , Shizuoka, Japan (URL-http://www.ddbj.nig.ac.jb )
22 Keen ,G., Burton,J., Crowley,D., Dickinson,E., Espinosa-Lujan,A., Franks,E., Harger,C., Manning,M., March,S., McLeod,M., et al . (1996 ) Nucleic Acids Res. 24 13 -16.
23 McKusick,V.A. (1994 ) Mendelian Inheritance in Man. Catalogs of Human Genes and Genetic Disorders 11th edition
24 Online Mendelian Inheritance in Man, OMIM (TM). Center for Medical Genetics , Johns Hopkins University (Baltimore, MD) and National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD), 1996. (URL-http://www3.ncbi.nlm.nih.gov/omim/ ).
