Proteomics Online Tool Development and Design: FEATUREViz and WebFEATURE
Nicholas C. Mew
Oral Defence Date:
Profs. Dragutin Petkovic and Ilmi Yoon
Proteomics analyses often involve noisy, high-dimensional, big data. The challenges of dimensionality, tractability, and noise can confound automated research. Effective research employs human-guided semi-automatic algorithms, combining the best of human curiosity and creativity with computational speed and accuracy. Therefore, applications that facilitate discovery through query, manipulation, and visualization are essential tools for proteomics. In this report we present new research on two such front-end systems: novel design and development of FEATUREViz, and improvements of WebFEATURE. FEATUREViz is a front-end application developed in collaboration with the Stanford Helix group, designed to assist in human visualization and interpretation of results from PocketFEATURE, an application which identifies novel protein-drug interactions by analyzing the similarity between pockets in proteins. WebFEATURE is a web application for interactive annotation of functional sites in molecules, hosted at Stanford University and co-developed at SFSU. We present focus group feedback that shows improved understanding of protein models as a consequence of using these tools. FEATUREViz is scheduled to be published as open source software on Stanford's SimTK website in December 2014 and hosted as an application in early 2015. Both FEATUREViz and WebFEATURE can be found online at the Stanford's Simulation Toolkit (SimTK) website http://simtk.org.
Protein, Proteomics, FEATURE, Micoenvironment, Molecular Discriptor, Protein Pocket, Pocket Alignment, Alignment Algorithm, Visualization
Nicholas C. Mew