Improvement of the Visualization of Stanford WebFEATURE System for Function Prediction in Protein Structures


Nilaykumar Patel

Oral Defence Date: 



HH 301


Professors Petkovic and Wong, & Mike Wong, CCLS


Very important component of all bioinformatics software for analyzing biologic structures is the Web-based presentation and visualization of results for further analyses by researchers and students. Stanford WebFEATURE (http://feature.stanford.edu/webfeature/) developed by Helix group is a one such application. It gives scientists the ability to scan protein structures for functional sites. WebFEATURE provides an online service with visualization of scan results (e.g. hits) superimposed on 3-D molecular structure models. The primary goal of this work was to redesign current WebFEATURE results permeation interface and subsystem based on needs of real users and future enhancements of WebFEATURE like incorporation of multiple machine learning methods. In the course of the work two additional goals were identified and addressed: novel functionality for archiving scan results with subsequent visualization via same WebFEATURE results interface, and general improvements on WebFEATURE infrastructure. This work is part of ongoing collaboration between Stanford Helix Group and SFSU FEATURE group. The contributions of this work addressed the above there goals as follows: a) WebFEATURE result user interface and subsystem were completely redesigned and implemented: results page now allows for display of results of two machine learning algorithms (Naïve Bayesian and Support Vector Machines); layout is greatly improved; better interactivity in choosing variety of display options and selections was introduced; b) new archiving functionality which allows archiving of scan results from WebFEATURE and standalone FEATURE systems, and their display suing WebFEATURE was designed, developed and tested; and c) WebFEATURE configuration management was improved. Modern software engineering methods were applied in developing this work such as agile and SCRUM project management, iterative SW design and development and User Centric Design (UCD) for user-intensive component of the systems. UCD process was implemented in the form of focus groups at SFSU and Stanford.


visualization, bioinformatics, WebFEATURE


Nilaykumar Patel