Statistical evaluation for detection of peptide specific interferon-γ secreting T-cells induced by HIV vaccine determined by ELISPOT assay

Abstract

Interferon-gamma (IFN-γ) secreting T-cells play an important role in immunodeficiency virus (HIV) disease pathogenesis. A common technique to determine the efficacy of the HIV vaccines in murine models is to compare the number of IFN-γ secreting T-cells induced by HIV vaccine to a control group. The measurement of IFN-γ secreting T-cells relies on an ELISPOT assay. This assay is carried out in triplicate wells from the same mouse on the same ELISPOT plate. The numbers of spot forming cells (SFC) from these wells are correlated counts. Traditionally, simple statistical methods, such as ANOVA or Kruskall–Wallis tests, are performed on means by mouse. These approaches ignore the fact that the data are correlated counts. Count data are usually assumed to follow the Poisson distribution. However, some count data exhibit overdispersion that can affect the test statistics. The negative binomial distribution is an alternative to the Poisson distribution in the presence of overdispersion. Hence, negative binomial regression is a more suitable approach for overdispersed count data. In this study, we used a negative binomial regression to determine that IL-12 was a good adjuvant. The results of the study using a negative binomial regression are discussed.