My research area of interest in Bioinformatics which special
considerations given to the mathematical discovery of biomarkers. Using
the tools of mathematical modeling and machine learning, I am able to
create classifiers to accurately predict disease outcome, discover the
best treatment options for disease, or to examine disease prognosis. The
methods of machine learning allow for multiple types of input, which
make it extremely well suited for use in a clinical/translational
setting. By utilizing both a training set of data and a testing set of
data, one can check the accuracy of the classifier or model to build the
best possible biomarker. Once the classifier has been created,
researchers can use it to predict the outcome or best treatment options
for future patients. I am also actively involved in developing methods
to combine various types of data as well as optimizing the kernel
functions which are at the heart of the machine learning algorithms.
- Renwick, A., Davison, L., Spratt, H., King, P., and Kimmel, M.
(2001) DNA Dinucleotide Evolution in Humans: Fitting Theory to Facts.
- Brasier, A., Spratt, H., Wu, Z., Boldogh, I., Zhang, Y.,
Garogalo, R., Casola, A., Pashmi, J., Haag, A., Luxon, B., and Kurosky,
A. (2004) Nuclear Heat Shock Response and Novel Nuclear Domain 10
Reorganization in Respiratory Syncytial Virus-Infected A549 Cells
Identified by High Resolution 2D Gel Electrophoresis, Journal of
- Forbus, J., Spratt, H., Wiktorowicz, J., Wu, Z., Boldogh, I.,
Denner, L., Kurosky, A., Brasier, R., Luxon, B., and Brasier, A. (2006)
Functional Analysis of the Nuclear Proteome of Human A549 Alveolar
Epithelial Cells by HPLC-High Resolution 2D Gel Electrophoresis,
Proteomics, 6(9): 2656 – 2672.