January 1st, 2012

Speaker: Richard Park
Advisers: Peter Park & Simon Kasif
Title: Seqeyes: A multi-scale interactive visualization tool for structural variations

ABTRACT
Genomic structural variations (SV) are known to play an important role in cancer and other diseases. Next-generation sequencing is a key technology for the identification of such variations, but current data and algorithms yield many false positives and false negative predictions. We have created a visualization tool called Seqeyes to explore and interpret predicted structural variations to help guide experimental validations. Users can sort, filter, and aggregate samples based on clinical attributes, which facilitates the association of phenotypes with specific patterns of structural variation. Our tool provides a combination of linear and circular representations. Two genome browser views show detail at multiple locations concurrently, while the circular ideogram view provides a global summary. Multiple molecular data types including copy number, gene expression, and methylation microarrays for each sample are integrated as additional genomic tracks. We leverage advanced open-source indexes available from Postgresql and Postbio to greatly enhance the speed and amount of data available to visualize. Seqeyes is a novel multi-scale visualization that can interactively navigate dozens of genomes down to individual sequencing reads within a web browser.