Data Centralization - Biomedical Informatics in Translational Research - page 147

Table 8.1 (continued)
Name and Reference Description and URL

Atlas [33] Atlas is an integration of the public GenBank, RefSeq, UniProt, Human Protein Ref-

erence Database (HPRD), BIND, DIP, Molecular Interactions Database (MINT),
IntAct, NCBI Taxonomy, GO, OMIM, Entrez Gene, and HomoloGene databases.
The authors classified these databases into four main groups: sequence, molecular
interactions, gene-related resources, and ontology. A set of APIs written in C++,
Java, or Perl is provided in the toolbox to allow users to retrieve the data based on
what they need.
URL: http://bioinformatics.ubc.ca/atlas

ONDEX [34] ONDEX presents biological data in graphs, with nodes for entities and edges for

relationships that are defined at the semantic level. The generalized data structure,
also called the entity-attribute-value (EAV) model, which is discussed in Section
8.3.7, is used for data storage. Currently parsers of some of the public databases
that are of interest to the developer have been developed, including AraCyc,
BRENDA, Drastic, TRANSFAC, EC Nomenclature, WorldNet, MeSH, and NCBI
TAXONOMY. The system allows for importing and integration of any of the pub-
lic databases as long as a parser has been developed.
URL: http://sourceforge.net/projects/ondex

Entrez [26] This is a web-based integrated database retrieval system developed by NCBI. It

enables text searching of a diverse set of databases using simple Boolean queries.
Thirty such databases had been included by 2005 and the number is still increasing.
URL: http://www.ncbi.nlm.nih.gov/Database/index.html

SRS [27, 28] The EBI SRS is an integration system that allows for both data retrieval and pro-

vides applications for data analysis. Users can query across databases quickly and
efficiently. As of 2002, the server contained more than 130 biological databases and
integrated more than 10 applications.
URL: http://srs.ebi.ac.uk

lection of patient-centric data across these platforms is not available, at least not to
the public. There are a few organizations in the world that are at the cutting edge of
collecting and producing such data sets. In this section we provide an overview of
the types of data that we think are important to translational biomedical informa-
tics research. This discussion is followed by a section on selected public repositories
for these types of data.

8.2.1 in-house Patient-Centric Clinical, Genomic, and Proteomic Data
A comprehensive body of clinical data is important to the analysis and interpreta-
tion of genomic and proteomic data. Unfortunately, the types of collected clinical
data for this purpose vary from project to project, and standards governing the col-
lection of these data have not been developed. As a result, although genomic and
proteomic data from clinical specimens are available in the public domain, the
information about the specimens and the donors (normal subjects or patients) is
very limited. To collect relatively complete clinical, genomic, and proteomic data
sets, researchers will generally almost have to start at the beginning by designing a
proper project for clinical data and specimen collection, and then continue to con-
duct molecular studies using the collected specimens. In Chapters 2 and 3, we dis-
cussed clinical operations, and the following types of data obtained from the clinic
are considered to be important to biomedical informatics activities.