Osaka Prefecture University

A Toolkit for Correlation Network Analysis – a Tailored-Size Community Detection- Is Now Available

LastUpDate: October 24, 2018

Dr. Yoshiyuki Ogata, Graduate School of Life and Environmental Sciences, Osaka Prefecture University, Dr. Hideyuki Suzuki, Department of Research and Development, Kazusa DNA Research Institute and other members of their research group released ConfeitoGUI, a toolkit that users can adjust network module size during the detection from a correlation network. This toolkit can be useful for analyzing large data as local communities with preferred size can be obtained by simply changing size parameters. The original algorithm for a correlation network is patent program (Intellectual Property No. 6318334). The research findings were published in online journal of PLOS ONE.

The main point of research

  • figure, imageWhile ConfeitoGUI offers a function to adjust the size of network modules and avoid elements having redundancy in their memberships, users can focus on a reasonable number of candidate elements desired for actual verification.
  • The Confeito algorithm is applicable to a network with even less distinctive communities. In the field of molecular biology, it is used for validation of the grouping in which the size of network module is tailored for user’s capacity of performance.
  • The tool name “Confeito” actually refers to Portuguese traditional sugar candies and the algorithm was also inspired by the recipe.

Overview of research

There are a variety of approaches to extract network modules from a large network of elements and the links in between. However, despite of extracting a focused element, the number of redundant elements are inadvertently included. As it is difficult to adjust the size of network modules, the modules are often much smaller or larger than expected. In the field of molecular biology, for example, the candidate number of elements such as genes and metabolites is so vast that further analysis is required to verify the predictions made by the community assignment. When conducting such analysis, a reasonable number of candidate elements is desired for actual verification.

With a function to remove such redundancies in its backend process, ConfeitoGUI is a user-friendly tool offering the adjustment of network module size by simply changing the size parameters. It can identify elements related to the network modules even when they are weakly correlated.

The toolkit has a great command of accepting any type of multivariate data and is size sensitive. The size of the individual network module is synchronized with each other and is automatically adjusted according to the size parameters. As the research group was inspired with Portuguese tiny colorful candies, this toolkit offers similar appearance for network modules – colorful and all perfectly uniformed size.

The research findings were published in online journal of PLOS ONE.

Article title: A toolkit for size-sensitive community detection from a correlation network.

Contact:

Graduate School of Life and Environmental Science
Dr. Yoshiyuki Ogata

E-mail ogata [at] plant.osakafu-u.ac.jp *Please change [at] to @.