|
|
|
|
|
|
|
|
| About Bruce Kristal (Brigham and Women’s Hospital) |
|
Bruce Kristal received in BS in Life Sciences from the Massachusetts Institute of Technology in 1986 and his PhD from Harvard University Graduate School of Arts and Sciences, Division of Medical Sciences, Committee on Virology in 1991. He rose from post-doc to Research Assistant Professor at the University of Texas Health Science Center at San Antonio (1991-1996), then moved to Burke Medical Research Institute (1996) and the Departments of Biochemistry (1997) and Neuroscience (1998) at Weill Medical College of Cornell University, becoming an Associate Professor in Neuroscience in 2004. He joined Brigham and Women's Hospital's Department of Neurosurgery. Dr. Kristal's research has been recognized by awards from the American Federation for Aging Research, American Institute for Cancer Research, the Hereditary Disease Foundation, and others. He has served as a Contributing Editor for Science's Science of Aging Knowledge Environment. He currently serves on the editorial Boards of The Journals Gerontology Biological Sciences Section and Analytical Biochemistry. Dr. Kristal is the first secretary and a member of the Board of Directors of the Metabolomics Society.
|
|
Cross-Species Metabolomics Analysis of Biomarkers for Diet and Disease Risk
Bruce S. Kristal, Brigham & Women’s Hospital
Dietary or caloric restriction (DR) is the most potent and reproducible known means of increasing longevity and reducing morbidity in mammals. This data is directly analogous to studies in humans that link obesity with poor health outcomes. We therefore proposed to test the general concept that biomarkers of diet in rats will predict disease in humans. Exploratory studies identified 93 redox-active small molecules from sera (measured by HPLC coupled with coulometric detector arrays) with potential to distinguish dietary groups in both male and female rats. Classification and predictive power were addressed using megavariate data analysis approaches. The compounds weakly distinguished AL and DR samples using clustering or principal components analysis (PCA) due to noise resulting from inter-cohort sampling. Soft Independent Modeling of Class Analogy, which builds independent PCA models of each class of interest, distinguished groups with 95% accuracy, but overfit models built on single cohorts. Partial Least Squares Projection to Latent Structures Discriminant Analysis, a projection method optimized for class separation, in contrast, built models with >95% accuracy in distinguishing groups without obvious cohort interference. Data processing choices of transformation, scaling, and winsorizing (outlier removal) each affected strength of the models, and, in some cases, revealed distinct metabolites to be of importance in building these models, often in gender-specific ways. Diets varying in extent and duration of DR were used to develop models for intermediate caloric intakes, which are more relevant for human studies. Markers were adapted for human study, analytically validated at both the instrumentation and at the sample collection levels, then biologically validated. We will present these modelling approaches, the models and their ability to distinguish sera based on caloric intake, and the potential for moving these markers to epidemiological studies in human sera.
|
|
|
|
|
|
|
|
|
|
|
