A VISUAL ANALYTICS APPROACH TO FEATURE DISCOVERY AND SUBSPACE EXPLORATION IN PROTEIN FLEXIBILITY MATRICES

Search for this publication on Google Scholar
Barlowe, S. (2011). A VISUAL ANALYTICS APPROACH TO FEATURE DISCOVERY AND SUBSPACE EXPLORATION IN PROTEIN FLEXIBILITY MATRICES. Unc Charlotte Electronic Theses And Dissertations.
Download PDF

Analytics

88 views ◎
32 downloads ⇓


Abstract

The vast amount of information generated by domain scientists makes the transitionfrom data to knowledge difficult and often impedes important discoveries. Forexample, the knowledge gained from protein flexibility data sets can speed advancesin genetic therapies and drug discovery. However, these models generate so muchdata that large scale analysis by traditional methods is almost impossible. This hindersbiomedical advances. Visual analytics is a new field that can help alleviate thisproblem. Visual analytics attempts to seamlessly integrate human abilities in patternrecognition, domain knowledge, and synthesis with automatic analysis techniques. Ipropose a novel, visual analytics pipeline and prototype which eases discovery, comparison,and exploration in the outputs of complex computational biology datasets.The approach utilizes automatic feature extraction by image segmentation to locateregions of interest in the data, visually presents the features to users in an intuitiveway, and provides rich interactions for multi-resolution visual exploration. Functionalityis also provided for subspace exploration based on automatic similarity calculationand comparative visualizations. The effectiveness of feature discovery and subspaceexploration is shown through a user study and user scenarios. Feedback from analystsconfirms the suitability of the proposed solution to domain tasks.

Details
Author

Barlowe, Scott

Title

A VISUAL ANALYTICS APPROACH TO FEATURE DISCOVERY AND SUBSPACE EXPLORATION IN PROTEIN FLEXIBILITY MATRICES

Physical Description

1 online resource (120 pages) : PDF

Date

2011

Degree Granting Institution

University of North Carolina at Charlotte

Abstract

The vast amount of information generated by domain scientists makes the transitionfrom data to knowledge difficult and often impedes important discoveries. Forexample, the knowledge gained from protein flexibility data sets can speed advancesin genetic therapies and drug discovery. However, these models generate so muchdata that large scale analysis by traditional methods is almost impossible. This hindersbiomedical advances. Visual analytics is a new field that can help alleviate thisproblem. Visual analytics attempts to seamlessly integrate human abilities in patternrecognition, domain knowledge, and synthesis with automatic analysis techniques. Ipropose a novel, visual analytics pipeline and prototype which eases discovery, comparison,and exploration in the outputs of complex computational biology datasets.The approach utilizes automatic feature extraction by image segmentation to locateregions of interest in the data, visually presents the features to users in an intuitiveway, and provides rich interactions for multi-resolution visual exploration. Functionalityis also provided for subspace exploration based on automatic similarity calculationand comparative visualizations. The effectiveness of feature discovery and subspaceexploration is shown through a user study and user scenarios. Feedback from analystsconfirms the suitability of the proposed solution to domain tasks.

Genre

doctoral dissertations

Subjects--Topics

Computer science
Bioinformatics

Degree

Ph.D.

Keywords

Cluster
Feature Extraction
Protein Flexibility
Subspace Exploration
Visual Analytics
Visualization

Subject Area

Information Technology

Advisor(s)

Yang, Jing

Committee Members

Livesay, Dennis
Jacobs, Donald
Lipford, Heather
Willis, Andrew

Degree Note

Thesis (Ph.D.)--University of North Carolina at Charlotte, 2011.

Rights Statement

This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s). For additional information, see http://rightsstatements.org/page/InC/1.0/.

Rights Holder Information

Copyright is held by the author unless otherwise indicated.

Identifier

Barlowe_uncc_0694D_10258

Permalink

http://hdl.handle.net/20.500.13093/etd:138