You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
125<br />
In February 1998, a ‘cDNA <strong>and</strong> EST Expression’ search was added. Unlike the<br />
MGD ‘Molecular Probes <strong>and</strong> Segments’ search, the GXD format enables better<br />
searching on source data associated with expressed sequences. Particularly, tissue,<br />
age, cell line, <strong>and</strong> library are prominent search fields.<br />
In June 1998, the first large-scale expression data set was searchable through the<br />
‘Gene Expression Data’ query form. These data, generated by T. Freeman (Sanger<br />
Centre, U.K.) include RT-PCR assays for 517 genes in 45 mouse tissues from 6-8<br />
week old animals <strong>and</strong> from 15 day old embryos. Assay data include primary data on<br />
each sample prepared, expression profiles for each sample in each experimental gel,<br />
<strong>and</strong> images of the gels.<br />
Importantly, although only limited expression data are available at this writing,<br />
GXD is poised to undergo rapid data expansion. The database is now implemented to<br />
capture <strong>and</strong> make available other datasets <strong>and</strong> several other types of expression<br />
information. Expression patterns are described by a comprehensive dictionary of<br />
anatomical terms that has been developed in collaboration with Drs. Bard <strong>and</strong><br />
Kaufmann of the University of Edinburgh <strong>and</strong> Drs. Davidson <strong>and</strong> Baldock of the<br />
MRC, Western General Hosptial, Edinburgh. Further, editorial interfaces for<br />
capturing <strong>and</strong> updating assay data from publications <strong>and</strong> image scanning are in place.<br />
Some expression data will be acquired from annotation of the literature by<br />
database editors. However, because only a fraction of the gene expression data that a<br />
laboratory generates actually appears in published form, it is anticipated that GXD<br />
data largely will come from electronic data submissions. The Gene Expression<br />
Annotator (GEA) has been developed for this purpose <strong>and</strong> is currently being tested<br />
by several laboratories. The Annotator prototype provides important features for<br />
capturing st<strong>and</strong>ardized descriptions of gene expression data, validating data, <strong>and</strong><br />
submitting data files. GEA supports a drag-<strong>and</strong>-drop facility for importation <strong>and</strong><br />
indexing of image files, a hierarchical look-up list for embryonic anatomy, <strong>and</strong> links<br />
to resources such as MGD <strong>and</strong> GenBank for verifying nomenclature <strong>and</strong> describing<br />
probes [ 18].<br />
With the sum of these developments, GXD will quickly exp<strong>and</strong> <strong>and</strong> new data sets<br />
<strong>and</strong> features can be expected to appear regularly on the Mouse Genome Informatics<br />
WWW site.<br />
The Mouse Genome Informatics (MGI) WWW site<br />
MGD <strong>and</strong> GXD together provide a unique resource for analyzing how the structural<br />
genome, through developmental pathways, produces observed phenotypes (Figure 3).<br />
These databases are tightly integrated to enable comprehensive analysis of genotype,<br />
expression, <strong>and</strong> phenotype data. Further, as a practical consideration, because MGD<br />
already contains many data types that need to be shared with GXD ( e.g., data on<br />
genes, molecular probes, inbred <strong>and</strong> mutant strains, references), efficiency is gained<br />
by coordinating maintenance of these data.