With a research
theme in applied probability and statistical bioinformatics,
several ongoing funded projects can be classified into a trio of
major theoretical concepts: scan
statistics, Poisson
approximation, and excursion
theory. The most exciting
investigation is in the overlapping area of all three concepts
for our research work on prediction algorithms. The prediction
can be very computationally intensive for longer structural
sequences, while much mathematical groundwork needs to done on
this concept trio, especially for the new excursion theory and
the concept of compound Poisson approximation. On the other hand,
scan statistics
with a better mathematical foundation is now a popular concept
in genomics studies and health data surveillance. Here are brief
descriptions of various research interests at different levels:
-
Undergraduate Participation in Bioinformatics Training
(UPBiT):
Individualized 3-year training programs with activities
including communication workshops, lab training, research
rotations, field trips, conferences, and specialized
bioinformatics research projects that will typically involve
techniques in discrete math, linear algebra, optimization,
probability, and statistics.
-
Genomics and
Sequence Analysis (Master's
in Bioinformatics):
Development
and application of
more specific bioinformatics computing tools to predict
genomic structures and analyze molecular sequences. Students
are encouraged to modify the computer programming for
existing
tools and to integrate with the wet-lab equipment involved.
-
Applied
Probability and Biostatistics (Master's in
Statistics or
Mathematics):
Investigation
into probabilistic models and other techniques (e.g.,
Poisson approximation) for
statistical analysis of biomolecular sequences. The emphasis will vary for
students interested in mathematics, applied mathematics, or
statistics.
-
Probability
and Computational Biology (Ph.D.
in Computational Science):
Development
of new probabilistic models (e.g., excursions) and
distributed computing techniques to establish new approaches
to computationally intensive problems stemming from
biomolecular sequence and structure analyses. A current
problem of interest is RNA structure prediction.
-
Prediction
Algorithms (Postdoctoral in Science and Engineering):
Research into
new mathematical and computational frameworks for developing
prediction algorithms and data mining systems to address emerging
bioinformatics problems.
|
|