The NIDDK Information Network: A Community Portal for Finding Data, Materials, and Tools for Researchers Studying Diabetes, Digestive, and Kidney Diseases

Abstract

The NIDDK Information Network (dkNET; http://dknet.org) was launched to serve the needs of basic and clinical investigators in metabolic, digestive and kidney disease by facilitating access to research resources that advance the mission of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). By research resources, we mean the multitude of data, software tools, materials, services, projects and organizations available to researchers in the public domain. Most of these are accessed via web-accessible databases or web portals, each developed, designed and maintained by numerous different projects, organizations and individuals. While many of the large government funded databases, maintained by agencies such as European Bioinformatics Institute and the National Center for Biotechnology Information, are well known to researchers, many more that have been developed by and for the biomedical research community are unknown or underutilized. At least part of the problem is the nature of dynamic databases, which are considered part of the "hidden" web, that is, content that is not easily accessed by search engines. dkNET was created specifically to address the challenge of connecting researchers to research resources via these types of community databases and web portals. dkNET functions as a "search engine for data", searching across millions of database records contained in hundreds of biomedical databases developed and maintained by independent projects around the world. A primary focus of dkNET are centers and projects specifically created to provide high quality data and resources to NIDDK researchers. Through the novel data ingest process used in dkNET, additional data sources can easily be incorporated, allowing it to scale with the growth of digital data and the needs of the dkNET community. Here, we provide an overview of the dkNET portal and its functions. We show how dkNET can be used to address a variety of use cases that involve searching for research resources.

Fig 1. Schematic overview of dkNET Navigation that illustrates the progressive refinement of information within the system.

At the top level, dkNET organizes resources under 3 views: Community Resources, More Resources and Literature. Within each view are specific categories, e.g., for Community Resources, the categories comprise Materials, Funding, Protocols, Data and Organism. Each of these, in turn, may be expanded into several subcategories, e.g., Materials is subdivided into Antibodies, Vectors, Embryonic Stem Cell (ESC) lines and PCR primers. Once a Subcategory is selected, a set of relevant sources are provided, each with a set of appropriate facets. Not all levels may be available for all subsections, e.g., a category may not have subcategories or a given source may not have facets.

Fig 2. Main results page of dkNET search portal as of June, 2015.

Search results are organized within 3 data views, indicated by the categories in the upper right menu (rectangle), and also via categories (rectangle, left side), subcategories and sources (left navigation menu). Results can be displayed as snippets (main window) or as a table (lower right) by selecting the Options menu next to each result. Results sets may be saved (see text) or exported as CSV files by selecting “Download first 1000 results” (oval) accessed via the tool icon in the upper right corner of the table view (red arrow).

Fig 3. Summary of the current coverage of the data content available through the dkNET More Resources tab (Oct, 2014).

The major categories are displayed on the Y axis, while the number of records for each category of data is displayed on the X axis. For legibility, only some of the categories are shown.

Fig 4. Query expansion.

The included terms are additional terms used to query the data sources based on the term expansion process using knowledge from the ontology. The most common expansion includes synonyms and common abbreviations of a term.

Fig 5. Example of faceted data in dkNET.

Data columns that contain repeated values are selected by the data curators to be faceted. This is one type of ‘filter’ to help users narrow search results.

Fig 6. Collections and Save Searches Users can create collections to store data records.

A record can be added to one or more collections. All records on a page can be added to a collection by selecting the “Add all records…” option in the left navigation menu. Searches are saved by selecting the icon to the left of the search bar. A menu is then displayed (inset, upper right) that shows previously saved searches, along with functions to save the current search.

Fig 7. Workflow to search for funding opportunities related to drosophila and diabetes.

In the first search using the terms “drosophila diabetes funding opportunities” a limited number of results are returned. Since dkNET includes a funding category and an opportunities subcategory, theses terms can be removed as search terms, therefore expanding the search and returns 223 (6/2015) records. This expanded set of results can then be further filtered in the table view.