COMPUTATIONAL TOOLS 77
It is also desirable to exploit information in the database that is not explicitly represented in the
schema.^49 Thus, the CCDB project team is developing specific data types around certain classes of seg-
mented objects contained in the CCDB. For example, the creation of a “surface data type” will enable users
to query the original surface data directly. The properties of the surfaces can be determined through very
general operations at query time that allow the user to query on characteristics not explicitly modeled in
the schema (e.g., dendrites from striatal medium spiny cells where the diameter of the dendritic shaft
shows constrictions of at least 20 percent along its length). In this example, the schema does not contain
explicit indication of the shape of the dendritic shaft, but these characteristics can be computed as part of
the query processing. Additional data types are being developed for volume data and protein distribution
data. A data type for tree structures generated by Neurolucida has recently been implemented.
The CCDB is being designed to participate in a larger, collaborative virtual data federation. Thus, an
approach to reconciling semantic differences between various databases must be found.^50 Scientific
ProjectExperimentSubjectTissue
ProductTissue
ProcessingFixation Protein
Localization Staining EmbeddingMicroscopy
ProductMicroscopy
Image DetailsAnatomical
DetailsRegion of
Interest
Reconstruction Reconstruction
Image DetailsSegmentationTree Tracking MorphometricsFIGURE 4.1 The schema and entity relationship in the Cell Centered Database.
SOURCE: See http://ncmir.ucsd.edu/CCDB.
(^49) Z. Lacroix, “Issues to Address While Designing a Biological Information System,” pp. 4-5 in Bioinformatics: Managing Scien-
tific Data, Z.T. Lacroix , ed., Morgan Kaufmann, San Francisco, 2003.
(^50) Z. Lacroix, “Issues to Address While Designing a Biological Information System,” pp. 4-5 in Bioinformatics: Managing Scien-
tific Data, 2003.