-
-
Development of methods for identifying genetic
networks from DNA-probe-array generated gene expression data.
-
Mathematical model of apoptosis and resistance
mechanisms in cancer chemotherapy
-
Evaluation of prognostic importance of molecular
markers (e.g. p53, erbB2 , Ki67 etc) in a large series of
node negative breast cancer patients with long follow-up.
-
Development of improved methods of prognostic
prediction: neural network models for time-to-event data,
regression splines for studying how prognosis varies with
marker level, methods for selection and evaluation of optimal
cut-point for markers, cross-validation for model selection,
Bayesian proportional hazards models, measures of explained
variation for time-to-event data.
-
Analyses of factors which effect responses
of tumor cell lines to a wide range of chemical compounds,
e.g. mechanism of action, histology, MDR, P53. Development
and application of pattern recognition analytic and visualization
methods for studying multivariate relationships in this large
database.
-
Development of Bayesian methods for the design
and analysis of 2 x 2 factorial trials.
-
Development of Bayesian methods for dealing
with multiple comparison problems. Applications to evaluating
associations between tamoxifen and second cancers of G.I.
tract. Application to analysis of genetic linkage studies.
Application to analysis of cytogenetic abnormalities in patients
with melanoma or ovarian cancer.
-
Development and application of randomized
selection designs as alternatives to phase III designs for
therapeutic development in rare diseases
-
Development of Bayesian sequential designs
for monitoring efficacy and toxicity in phase II trials of
drug combinations using historical controls.
-
Development of more flexible sequential monitoring
designs to adapt the sample size of a randomized trial to
the magnitude of the treatment effect.
-
New methods for the analysis of complex health
survey data.
-
New methods for the analysis of quality of
life data in clinical trials.
-
Mathematical models of the relationship between
chemotherapy schedules and dose intensities and effects.
-
New methods for the analysis of spatially
correlated binary and count data.
-
Research in molecular statistics and bioinformatics
section