Educational Resources
Computational Biology and Biostatistics Summer Research Program
- Research Projects - 2005-
The following research topics were from the Summer 2005 Research Program in Biostatistics. Research project descriptions from previous years are also available. Next year's program will probably have different topics, but on the same level as the ones listed below. Click on title of project for further information on each research project.
Student
Mentor(s)
Title of Project
Huy Vuong
Moo Chung
Correlation, Permutation, and the Brain
Malisha Pattanaik
Jens Eickhoff
Prediction of Cardiorespiratory Fitness (CRF) in Obese Children
Joanne Sklodowski
Brian Yandell
Insulin Production in Diabetes Model:
Spatial Comparison of Strains
Basic R
programming, nonparametric tests, Monte-Carlo simulation.
Abstract:
In brain imaging studies, it is necessary to correlate psychological measures with neuroanatomical measures and compare the correlation measure between groups while removing the effect of gender and age. Partial correlation can be used for this purpose. We will test if the partial correlation measure is different between the groups of autistic and normal subjects using the permutation test. Since it is required to test the correlation in more than 40,000 sites in the brain, we will also extend the
permutation test to the problem of testing multiple hypotheses simultaneously.
Title of Project:
Prediction of cardiorespiratory fitness (CRF)
in obese children
Development of a model to predict cardiorespiratory fitness
Knowledge/Skills needed::
Basic concepts of probability and statistics
Knowledge/Skills obtained:
Statistical modeling (regression analysis, principal
component analysis, etc.); use of statistical software (SAS and R).
Abstract:
Low cardiorespiratory fitness (CRF) in children can lead to heart disease, stroke diabetes and hypertension in adult life. CRF is typically assessed by measuring maximum oxygen uptake (VO2max) which is performed in a laboratory setting (using graded exercise tests) and may require considerable expense in terms of time, trained personnel and medical supervision (American College of Sports Medicine, 2000). Moreover, accurate assessment of VO2max may be difficult in obese children. However, less expensive standard measurements, e.g., BMI, lean body mass, fat percentage, etc., can be used to predict VO2max. In this project, we will develop a model for predicting VO2max in obese children using various standard demographic and body composition variables.
Title of Project:
Insulin Production in Diabetes Model:
Spatial Comparison of Strains
Improve methods to distinguish quality of insulin production between two strains of mice based on 3-color images of islets in pancreas slices.
Knowledge/Skills needed::
Basic statistics (t-tests, ideally some nonparametrics)
R programming (a la biostat training)
Knowledge/Skills obtained:
Communication skills with scientist (lab biochemist), image analysis basics, nonparametric density estimation (mechanics at least), experimental design considerations (nesting, balance, etc.)
Abstract:
Insulin is produced in the pancreas, specifically in the Islets of Langerhans by beta cells. (Yes, there are alpha cells as well.) Scientists in biochemistry have images of hundreds of these islets for two types of mice. One strain, BTBR.ob, mimics Type II (mature onset) diabetes. A second strain, B6.ob, appears healthy. We have roughly a hundred islets per mouse, with 3 mice per strain. We can visually see differences in the production of insulin in the islet beta cells in these images. And we have summary measurements on each islet that we can use to compare the two types of mice. Further, there is opportunity to consider how we might improve on the measurement process and the experimental design. The goal is to clearly infer differences in the
quality of insulin production between BTBR.ob and B6.ob. You can/will meet with the scientists and work with the measurement system as well.
