Herbert H. Tsang:: Research

Research

My research interests are in the broad areas of bioinformatics/biomedical engineering, scientific visualization, signal processing, education technologies, and mobile application development. Also, these research projects has been supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and Mitacs.

a. Bioinformatics / Biomedical Engineering

RNA Structure Prediction:

In my PhD research, I focused on the development of a new heuristic algorithm for the prediction of RNA secondary structure. My algorithm SARNA-Predict has shown surpassing results in terms of correctly predicted base pairs when compared with nine other existing methods: mfold, P-RnaPredict, pknotsRE, NUPACK, pknotsRG-mfe, Sfold, HotKnots, ILM, and STAR. These algorithms are from three difference classes: heuristic, dynamic programming, and statistical sampling techniques. Furthermore, SARNA-Predict-pk is an extension of SARNA-Predict which can also predict pseudoknotted RNA structures.

An evaluation for the performance of the new algorithm in terms of prediction accuracy was verified with native structures. Experiments on thirty three individual known structures from eleven RNA classes (tRNA, viral RNA, anti-genomic HDV, telomerase RNA, tmRNA, rRNA, RNaseP, 5S rRNA, Group I intron 23S rRNA, Group I intron 16S rRNA, 16S rRNA) were performed. The results demonstrated that SARNA-Predict can out-perform other state-of-the-art algorithms in terms of prediction accuracy.

Sequence Analysis

Large scale small RNA sequencing projects generate large amount of data. Generally, only perfectly matched sequences are further analyzed. In some cases 61.8 % of data are unattended. We are investigating a novel approach to decipher sequencing errors from RNA editing events in small RNA sequence analysis. A software tool Ebbie-(mis)match was developed and to aid the exploration of biochemist in small RNA. Ebbie-MM is a computational algorithm that analyzes single mismatches of small RNAs to the genome of origin.

Multi-Imaging Mode Confocal Optical Microscope:

imgcogEunKhaV.jpg

Multi-Imaging Mode Confocal Optical Microscope (MIMCOM) is an advanced digital imaging microscope providing chemical composition of the target material. The images will be chemically analyzed with a selection of spectral and spatial feature identification tools, including artificial intelligence methods. Project work completed includes optical system design, microscope control and image analysis tool development. More information - link, also view the Networks of Centres of Excellence "Intelli-HEALTH" press releases, link

Biomolecule Visualization

RNA-DV

RNA-DV is a flexible and powerful tool targeted at visualization and modification of RNA secondary structure. Compare to existing software applications, RNA-DV set itself apart by not only providing an editing and visualization interface but also integrates thermodynamic energies for the current structure design. In RNA-DV, structural modification is achieved through a combination of mouse and keyboard operations. The direct and intuitive interface allows the structure prediction researcher to visualize the structure. Furthermore, RNA-DV provides the necessary tool for structure designers to explore the effect of various structural elements in terms of thermodynamic contributions. Four different thermodynamic energy models (efn2, HotKnots, Major and RNAeval) were incorporated allowing users to observe real-time energy changes while designing the structure. Visualizing and editing of pseudoknotted structure is also supported in RNA-DV.

  • Download source link
  • Online demo link
  • View a demonstration video (click on the photo below)

b. Multidimensional Data Visualization / Signal Processing

Agent-Based Modeling and Visualization

Developing an agent based model simulation framework to model scenario planning and analysis in an urban settings. The model is grounded on the population based stochastic optimization technique Particle Swarm Optimization (PSO) and System Dynamics models.


Hyperspectral Image Browser

Hyperspectral Image Browser (HIBR) was eventually incorporated into the commercial release of PCI’s Geomatica 9. Typical hyperspectral data size in the order of several hundred megabytes, which impose stringent requirements in data storage, data transmission and processing time. Using a patented compression technologies, this product allow scientist to explore and analyze hyperspectral data cube in a very effective manner. Also, experiments were performed on medical imaging data with encouraging results.

News items:

  • PCI Geomatics Joins The Canada Centre for Remote Sensing (CCRS) at the Opening of the Innovation Acceleration Centre (IAC) link
  • "Viewing Hyperspectral Data", Space Technologies: Investing in Out Future, Canadian Speace Agency, 2000 p. 25 link

Visualization of Medical Images

I had worked in the research lab where we performed research in the areas of medical imaging, image compression and audio engineering application. In particular, I was involved in developing applications for helping surgeons visualize medical images (Magnetic Resonance Imaging) while performing frameless image-guided neurosurgery. These experiences have given me a solid foundation for performing research in multi-disciplinary areas (mathematics, medicine and engineering) and for the successful development of a commercial product.

c. Education Technologies

At Simon Fraser University, besides performing research in the area of Bioinformatics, I also developed research interest in education technologies. In particular, I was involved projects in the analysis, design, development and evaluation of training and education materials and programs in eLearning environments. Specific areas include developing algorithms for enabling self-regulated learning. In collaboration with SFU's University/Industry Liaison Office, a spin-off company MusicCentric Technologies Inc. (http://www.music-centric.com) was formed.

d. Modelling of Complex Social Systems

In the Interdisciplinary Research in the Mathematical and Computational Sciences (IRMACS) Centre, the Modelling of Complex Social Systems (MoCSSy) program brings together extensive expertise in criminology, health science, urban dynamics, computer science, and mathematical modelling, under the unifying theme of modelling the complex dynamics that drive the linked epidemiologies of crime, disease, homelessness and other social ills in urban neighborhoods. See http://mocssy.irmacs.sfu.ca/

At MoCSSy, one of the projects that I am involve with is to examine the "HIV Transmission Rate through Sexual Contact with the Cellular Automata Model" MoCSSy-HIV_Poster_20090831.png

e. Mobile Application Development

I have several projects centered around mobile application development.

OSDFA "One Size Doesn't Fit All"

Project centered on “Mobile computing for user with accessibility issues.”

Location-based Application and Games

To see some of the project, visit here.


© Herbert H. Tsang, http://www.herberttsang.org

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License