Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/33862
Title: | Visualizing High Dimensional Datasets Using Parallel Coordinates: Application to Gene Prioritization | Authors: | Boogaerts, T Tranchevent, LC Pavlopoulos, GA AERTS, Jan Vandewalle, J |
Issue Date: | 2012 | Publisher: | IEEE | Source: | IEEE 12TH INTERNATIONAL CONFERENCE ON BIOINFORMATICS & BIOENGINEERING, IEEE, p. 52 -57 | Series/Report: | IEEE International Conference on Bioinformatics and Bioengineering | Abstract: | In this paper, we introduce a visualization tool for interactive and efficient exploration of high dimensional data using parallel coordinates. An algorithm is developed to find an optimal permutation of dimensions, which allows the data miner to immediately see the most important features or irregularities in the dataset. This is implemented as a genetic algorithm based on the travelling salesman problem using maximal correlation as fitness. Other features of the tool include selection operators to group the data such as selection by intersection or by angle, orthogonal and density plots complementing the parallel coordinates plot, manual arrangement of permutation order of the dimensions, possibility to show all plots necessary to see all dimensional relations and displaying a certain number of standard deviations for each dimension separately. The tool is applied to multiple gene prioritization cases in search of genes that are relevant to certain genetic disorders. The used datasets are obtained with the MerKator and Endeavour tools and include a Breast cancer, Cataract, Charcoth-Marie-Tooth and Cardiomyopathy dataset, as well as a dataset relating 29 diseases with 22206 genes. Our tool, manual and data can be downloaded from http://www.toomas.be/parcoord/. | Keywords: | data visualization;parallel coordinates;genetic algorithm;gene prioritization | Document URI: | http://hdl.handle.net/1942/33862 | ISBN: | 978-1-4673-4358-9 978-1-4673-4357-2 978-1-4673-4356-5 |
ISI #: | WOS:000315332900015 | Category: | C1 | Type: | Proceedings Paper |
Appears in Collections: | Research publications |
Show full item record
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.