Graduate Seminar

When

November 19, 2014    
1:10 pm - 2:00 pm

Where

3043 ECpE Building Addition
Coover Hall, Ames, Iowa, 50011

Event Type

Title: DiffSLc and VisPNet: a graph centrality, and a linear network layout viewer

Speaker: Divya Mistry, ECpE Graduate Student

Advisor: Julie Dickerson, Professor

Abstract: This seminar presents our work on a graph centrality measure geared towards biological networks, and a linear network layout viewer that utilizes the said work.

Graph centralities rank nodes and edges based on their topological importance. Centralities, such as node degree, closeness, and betweenness, have been useful in finding critical genes and proteins in a gene co-expression network and a protein interaction network. We have developed an information diffusion based centrality measure, called DiffSLc, geared towards similar biological networks. DiffSLc is designed to provide an insight into roles of genes or proteins in their respective network neighborhoods. Results of DiffSLc prioritization is compared to degree, closeness, betweenness, and subgraph centrality using a Saccharomyces cerevisiae gene coexpression network. Top-ranked genes from each method are compared against the known essential genes from Saccharomyces Gene Deletion Project. DiffSLc is implemented in R using igraph package, and will be available in near future via http://git.io/diffslc. We have also developed a linear network layout viewer, called VisPNet, which utilizes DiffSLc and other centralities to help visually mine a graph topology. VisPNet is built using D3, JavaScript, and related HTML5 technologies. VisPNet will be available in near future via http://git.io/vispnet.

Loading...