Advisory Committee Chair
Jonas S Almeida
Advisory Committee Members
Malay K Basu
Emidio Capriotti
Allan C Dobbins
Seung L Park
Document Type
Dissertation
Date of Award
2015
Degree Name by School
Doctor of Philosophy (PhD) School of Engineering
Abstract
This work explores web computing as a solution to current problems in informatics, including algorithm delivery, distributed computation, and data visualization. A novel computing platform called QMachine has been developed for use in clinical research environments for computationally intensive applications like medical image processing and genomics sequence processing. In such applications, the emergence of Big Data as the inevitable reference for the analysis of local (“small”) data from individual experiments or patients has presented numerous challenges. In particular, computational workflows that handle Big Data need significantly greater processing and memory resources. Advances in virtualization – especially in cloud computing – provide opportunities to find novel solutions for this problem in biomedical and clinical research environments. This thesis work approaches the central architectural challenge of scaling biomedical workflows through the use of web computing with histology and genomics as validating use cases. Web computing is an extreme form of cloud computing in which computations are distributed across the clients of a web service – typically, the web browsers of a website’s visitors. It permits a natural extension to the life sciences’ long tradition of volunteer computing by projects such as BOINC and Folding@Home. The work here seeks to establish the feasibility of this approach to Big Computing for typical bioinformatics scenarios such as image analysis and genomics sequence processing. The social conventions and governance that characterize biomedical and clinical re- search environments were treated as primary design criteria for this work. Data locality and security are two of web computing’s most promising concerns for hospital environments. For example, web browsers’ strict constraints for data exchange actually imply “Beyond the Data Deluge” [3] solutions in which code travel to the data. This work also evaluates web computing for supplying “the right information, to the right person, in the right intervention format, through the right channel, at the right time in workflow”. The goal of a web computing solution for distributed computing was pursued through the development and validation of a computational platform, QMachine. A live software prototype is publicly available at https://www.qmachine.org, and its entire codebase is open-sourced, version-controlled, and available online at https://github.com/qmachine.
Recommended Citation
Wilkinson, Sean, "Web Computing for Bioinformatics Applications" (2015). All ETDs from UAB. 3329.
https://digitalcommons.library.uab.edu/etd-collection/3329