Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Kari Lavikka, Jaana Oikkonen, Rainer Lehtonen, Johanna Hynninen, Sakari Hietanen, Sampsa Hautaniemi (University of Helsinki, Faculty of Medicine, Helsinki, Finland)
Many visualization tools for genomic data are built from the ground up and in an ad-hoc manner for specific purposes. Much effort is put into implementing the basic functionality, but the visualizations rarely exploit modern computers' graphics capabilities, namely the GPU (Graphics Processing Unit). GenomeSpy is a visualization tool with a Vega-Lite-inspired declarative grammar, allowing users to specify bespoke genomic and other data visualizations. With combinatorial building blocks such as graphical marks, scales, and transforms, users can author novel visualizations with little effort. Using the GPU for all scale transformations and rendering, the visualizations have high performance and smoothly animated interactions such as zooming and panning. GenomeSpy allows visualization of up to thousands of samples and supports their interactive filtering, grouping, and aggregation. We show that the declarative approach can be applied to data such as GWAS, multiple-sequence alignment, structural variation, and copy-number segmentation. We use GenomeSpy to explore and analyze the copy-number landscape and mutations of high-grade serous ovarian cancer.