Skip to main content
National Cancer Institute

Network Analytics to Assess Team Science

Edit resource

Nagarajan R. Network Analytics to Assess Team Science. Oral presentation at 2017 SciTS Conference. Clearwater Beach, FL. Jun 13, 2017. Multi-Site Health Research Collaborations Online at:

Recent studies have clearly demonstrated an increasing shift towards team science potentially attributed to growing interdisciplinary, multidisciplinary and transdisciplinary research. Grants from federal agencies (e.g. Clinical Translational Science Awards, NCATS) have especially emphasized the importance of translational research that in turn demand team science approaches. Our recent studies have successfully used network analytics of grant collaborations to objectively quantify and assess team science in translational settings in an evidence-based/data-driven manner (Nagarajan R et al., J. Biomedical Informatics 2013; Nagarajan R et al. Clinical Translational Science 2015). Collaborative grants are often the culminating point or outcome of successful and sustained research collaborations. Grant collaboration data sets are accessible and curated diligently with minimal errors making them a useful resource for investigating team science efforts. Grant collaboration networks (GCNs) provide a convenient abstraction of collaborations that can be studied in a controlled and cost-effective manner in-silico. In this presentation we show that GCNs can provide insights into inherent nontrivial community structures, cross-talk between communities and their temporal evolution. The strength of these communities as a function of time is also investigated by using synthetic surrogate network models (e.g. random graphs) as internal controls. Understanding the temporal evolution of these communities and their deviation from random graphs can especially be useful in evaluation in pre-/post-intervention settings and has the potential to serve as evaluation metrics. Inherent community structures and cross-talk between communities in the GCN can also assist in targeted resource allocation that can impact policy. However, GCNs are opensystems and prone to external perturbations and confounders that demand careful interpretation of the results. Forecasting of the GCNs can also be challenging as the nodes as well as the edges are not conserved as a function of time. Universality of the findings presented will demand repeating the exercise across diverse settings.



Type of Publication:

Oral presentation


scits 2017 conference, presentation, network analytics, team science

Addresses these goal(s):

  • Conduct research on/evaluate team science

Comments (0 comments)

Add Comments
Resource created by Jane Hwang on 10/5/2017 4:06:05 PM.

Narrow your search by
resource type or goal:

Advanced Search



Email this page