Online Program

Friday, February 21
PS2 Poster Session II & Refreshments Fri, Feb 21, 4:45 PM - 6:15 PM
Bayshore II-IV

Practical Statistical Issues in Analyzing Immunoassay Data (302834)

Yueh-Yun Chi, University of Florida 
Maureen M. Goodenow, University of Florida 
John W. Sleasman, University of South Florida 
*Xinrui Zhang, University of Florida 

Keywords: Biomarkers, HIV, plate adjustment, below detection, multiple comparisons

Immunoassays, such as ELISA and Luminex, are important tools to measure abundance of immunological biomarkers. Laboratory practice and physical constraints of the assays generate data that require careful statistical planning. Three practical issues will be addressed. First, the 96-well format in standard immunoassays limits the number of samples that can be analyzed in any single plate. Consequentially, samples are clustered at the plate level. Mixed models with the random effects component can be used to adjust for the plate effect. Second, practice in calibrating abundance can lead to data with values below or above the detection limits. Sensitivity analysis can be performed to compare the results of different calibration methods and derive an approach for robust calibration. Third, simultaneous analyses of multiple biomarkers provide a comprehensive assessment of immune activation and require adjustment for multiple testing to ensure valid statistical inferences. These three issues will be discussed in the context of a funded, ongoing project comparing immune activation between a cohort of 78 HIV-infected adolescents and 56 healthy controls.