Jim Hogan: Finding Friends Outside the Species: Making Sense of Large Scale BLAST Results with Silvermap


Presentation slides

Extended abstract PDF

Authors

Peter Ansell, Lawrence Buckingham, Xin-Yi Chua, James Hogan, Scott Mann, and Paul Roe (Queensland University of Technology)

Abstract

Bioinformatics is increasingly dominated by comparative investigations of coding and regulatory function across genome and species boundaries. Among the most important tasks in this enterprise is the discovery of similarity between fragments from different sequences, and this is normally accomplished through the use of BLAST, the well-known alignment search tool provided by NCBI. Regardless of the tool employed, the n vs. n-1 comparisons characteristic of such studies may produce literally thousands of numeric similarity scores and candidate alignments, making comprehension of all but the most clear cut results time consuming and error-prone, and discovery of more subtle relationships all but impossible. Some visualisation options are available, but the traditional, alignment oriented views are severely limited in their capacity to handle complex relationships, and the approach plainly cannot scale to accommodate the datasets being processed.

Silvermap is a web-based tool for visualising the results of comparative studies in genomics and proteomics, showing the relationships between items of interest based on the distance between them - essentially the inverse of their similarity - as though the genes or proteins are cities on a map. In this way, researchers may discover and explore relationships through the familiar visual metaphor of maps and navigation. Currently, the system presents a radar screen style view, with distance shown with respect to a primary feature, located centrally. Animation facilities allow an almost instant re-organisation of the map around another feature of interest, enabling rapid clustering by eye, and exploration of groupings based on distance threshold. Silvermap is currently available in beta form with  representative similarity databases from the prokaryotes: based on some 70000 genes from 14 E. Coli genomes; and the eukaryotes: in excess of 50000 genes drawn from each of the mouse and human genomes. Work is currently underway to extend the coverage of the tool to broader data sets, and to support easy deployment to individual laboratories.

Silvermap is written in the Microsoft Silverlight presentation environment, and runs on most common browsers. The tool is freely available under an open source licence.

About the speaker

James M. Hogan is Associate Professor of Computer Science at QUT and a project leader within the Microsoft QUT eResearch Centre, where he heads the Smart Tools for Bioinformatics initiative. Most of his research efforts in bioinformatics are focused on problems related to gene regulation, and on novel machine learning methods and software tools to allow discovery and management of promoter and transcription factor binding sites, and to understand the relationships between them. Among other achievements, this group has produced the state of the art methods for promoter prediction in bacteria, the SilverMap and SilverGene visualisation tools, the BioPatML pattern description language and tools, and a series of mashup components supporting bioinformatic exploration, shortly to be available as a JavaScript library. See www.mquter.qut.edu.au/bio for details, demos and downloads.