Serial dilution is a crucial step that is widely used when measuring the concentrations of unknown compounds in biological samples. Measurements below detection limits are a persistent problem in these assays. We will use hierarchical Bayes inference-a statistical approach for estimating groups of parameters in the presence of uncertainty-to improve estimation for serial dilution assays, thus allowing estimation of concentrations that would previously have been identified as """"""""below detection limits."""""""" We will develop a program using open-source software so that researchers from laboratories around the world can evaluate and use the new method. We will perform a laboratory validation study evaluating the new method under known conditions. We will immediately apply the methodology to laboratory studies of allergens in dust samples collected from the homes of children who are at risk for asthma. By extending limits of detection, the improved estimation procedure will be particularly helpful in the study of childhood asthma, where even very low allergen concentrations are hypothesized to have adverse health effects. We will also undertake a series of experiments and data analyses to extend the model to allow for changes in the calibration curve due to contamination of the samples, which is a common problem in the study of environmental samples, and in bioassays more generally. We will explore possibilities of more efficient designs of serial dilution experiments using our estimation procedure. Our method will be developed in the context of our laboratory studies of allergens and asthma; however, we anticipate it will be applicable much more generally to serial dilution assays in many different biological contexts. Measurement of low levels of exposure is critical in public health problems in general, hence this project, in which sources of uncertainty are explicitly modeled, leading to more precise estimates at low levels, will potentially lead to more effective bioassays generally. ? ? ?

Agency
National Institute of Health (NIH)
Institute
National Institute of General Medical Sciences (NIGMS)
Type
Research Project (R01)
Project #
1R01GM074806-01A1
Application #
7093264
Study Section
Biostatistical Methods and Research Design Study Section (BMRD)
Program Officer
Singh, Shiva P
Project Start
2006-07-01
Project End
2009-06-30
Budget Start
2006-07-01
Budget End
2007-06-30
Support Year
1
Fiscal Year
2006
Total Cost
$225,400
Indirect Cost
Name
Columbia University (N.Y.)
Department
Biostatistics & Other Math Sci
Type
Other Domestic Higher Education
DUNS #
049179401
City
New York
State
NY
Country
United States
Zip Code
10027