Translational science is very complex and can be characterized in a multitude of ways. Publications, grant applications, progress reports, funding, specific aims, study designs, hypotheses, outcomes, scientific evidence, experts, collaborations, organizations, and networks, as well as many other factors reflect the complex activity of translational and clinical research. These elements are interlinked and often are represented by massive heterogeneous data, e.g. thousands of papers are produced weekly. To make sense of these data and to aid in seeing the big picture, new methods are needed. Data visualization techniques (e.g. maps, networks, clusters, time series) expose patterns, trends and correlations and are proven to be useful for extracting information from abundant data. NCATS invites SBIR proposals that will facilitate the introduction of visualization technology into understanding the big picture of translational science, the evidence behind the knowledge about human health, and interdisciplinary communication of complex scientific information.
The outcome of this contract is expected to be software that assists in exploring multidimensional data and understanding complex concepts. The software should:
• visualize high-dimensional data from potentially diverse data sources
• enable data exploration, change of displayed dimensions, and semantic zooming
• create personalized views
• work with complex data dimensions, utilizing the elements of principal component analysis (PCA) or other appropriate techniques
• have transparent, validated, and well-documented protocols for all steps of data processing (cleansing, filtering, analysis, visualization, personalization, etc.)
• be accompanied by documentation of data processing algorithms, data accuracy, precision, and other features necessary for the most accurate interpretation of the produced visualization
• take advantage of the existing tools and technologies whenever possible • have Application Programming Interface (API) that does not require programming skills
Sample areas of interest to NCATS include
• Landscape of knowledge about human health with underlying evidence
• Knowledge gaps, discrepancies
• Comparative evidence
• Provenance of information about human health, therapeutics and diagnostics
• Translational Pathways: from discovery to clinical practice
• The patterns of self-care and health literacy in various cultures and communities
• Uncertainty of information about human health coming from clinical practice, research and consumers
• Propensity scores in observational studies
• Human subject research design
• Complex and distributed resources and on-going research activities, e.g. among Clinical and Translational Science Award (CTSA) institutions or NIH Institutes
The deliverable of Phase I is a visualization of a test dataset(s), which is made meaningful and valuable to NCATS through the process of interactive learning with minimal burden on the NCATS experts. It is envisioned that the offeror’s representative will gather initial information from publically available sources, and then fine-tune it via observing NCATS activities and interactions with NCATS staff to ensure that the presentation of data and analysis is tailored to NCATS interests and facilitates actions, discussions, feedback, and further learning.
The Phase II deliverable is web-enabled software that can be used for multidimensional data exploration and analysis, can work with multiple data sources, and can be personalized to the customer needs via generalized interactive learning methodology of Phase I.
The analysis should be done using a number of various data sources, e.g., publications, social media, NIH databases. The identification of appropriate sources is determined by the offeror. The choices must be justified, analyzed, and well documented with advantages and limitations of every source.
Other project clarifications
The offerors are encouraged to utilize the multiple principal investigator option to bring in experts from academia http://grants.nih.gov/grants/multi_pi/