The National NeuroAIDS Tissue Consortium (NNTC) Database: an integrated database for HIV-related studies

Abstract

We herein present the National NeuroAIDS Tissue Consortium-Data Coordinating Center (NNTC-DCC) database, which is the only available database for neuroAIDS studies that contains data in an integrated, standardized form. This database has been created in conjunction with the NNTC, which provides human tissue and biofluid samples to individual researchers to conduct studies focused on neuroAIDS. The database contains experimental datasets from 1206 subjects for the following categories (which are further broken down into subcategories): gene expression, genotype, proteins, endo-exo-chemicals, morphometrics and other (miscellaneous) data. The database also contains a wide variety of downloadable data and metadata for 95 HIV-related studies covering 170 assays from 61 principal investigators. The data represent 76 tissue types, 25 measurement types, and 38 technology types, and reaches a total of 33,017,407 data points. We used the ISA platform to create the database and develop a searchable web interface for querying the data. A gene search tool is also available, which searches for NCBI GEO datasets associated with selected genes. The database is manually curated with many user-friendly features, and is cross-linked to the NCBI, HUGO and PubMed databases. A free registration is required for qualified users to access the database.

Figure 1.

Information and data management flow for the DCC. Investigator requests go through (1) the clinical data manager who works with the investigator and data available through the web site to refine the request, which is then transmitted to the clinical sites and/or (2) the bioinformatics data curator, who can also work with (3) the investigator to refine the request as well as coordinate bioinformatics and epidemiological assistance as needed. Requests are entered into the management system for approval processes, and shipment of specimens/access to data is tracked for completeness. Once studies are complete the data curator works with the bioinformatics data manager to capture the data into the bioinformatics database, which is linked to the clinical database as well as outside databases (4) containing information on NNTC (and potentially other related) experiments.

Figure 2.

Entity-Relationship Diagram for the experimental data tables in the NNTC-DCC database. The central table is the studies table which describes a given study characterized by a study_id (e.g. in the form of S0001.1), principal investigator, year, tissue request ID, measurement, technology, unit and tissue. A study may consist of multiple tissue requests, thus the study_id might be incremented (e.g. S0001.1 to S0001.2). The tables gene_expression, genotype, proteins, endo_exo_chemicals, morphometrics, and other are grouped around the studies table in a 1:N relationship signifying that a single study may study multiple genes, proteins, morphological characteristics, etc. The table gene_annotation contains HGNC ids, gene symbols, synonyms and annotations for genes and proteins, hence their connection to the tables gene_expression, genotype and proteins. The ER diagram was created by using Lucidchart at www.lucidchart.com.

Figure 3.

Venn diagram showing the number of patients with data points from different combinations of experimental data categories. These categories include gene_expression, genotype, proteins, endo_eco_chemicals (drugs, endotoxins, glycosaminoglycans and steroids) and morphometrics. The data category other was excluded, because the Venn diagram can depict five categories at most; the category other had few data points. The Venn diagram tool is available at http://bioinformatics.psb.ugent.be/webtools/Venn/.

Figure 4.

Navigation of the NNTC-DCC database. After signing up and logging in, the user arrives at the browser (a), where studies can be selected based on a freetext search by entering keywords, or by filtering based on tissues, measurement and/or technology. Studies which meet the search/filter criteria show up below the search bar, ordered by date. Each study is characterized by a title, short description and measurement/technology. By clicking on the title the user comes to the study page (b), where full details of the study are displayed, e.g. long description, design, contacts, etc. By clicking on the ‘Gene Search’ link on the top tab, the user will then come to the gene search page (c). The user can search for up to 5 genes from a drop-down box populated by a list of genes names from the HUGO database. When the Search button is clicked, an NCBI page pops up with GEO sets associated with these genes.