Interpreting ELISA analyses from wild animal samples: some recurrent issues and solutions

Abstract

Many studies in disease and ecological immunology rely on the use of assays that quantify the amount of specific antibodies (immunoglobulin) in samples. Enzyme‐linked immunosorbent assays (ELISAs) are increasingly used in ecology due to their availability for a broad array of antigens and the limited amount of sampling material they require. Two recurrent methodological issues are nevertheless faced by researchers: (1) the limited availability of immunological assays and reagents developed for non‐model species, and (2) the statistical determination of the cut‐off threshold used to distinguish individual samples that are likely to have or not to have antibodies against a specific antigen. Here, we outline two solutions to deal with these issues. First, we show that implementing two assays with differing detection methods can help validate the use of reagents, such as antibodies, in species different from their intended target. We illustrate this by comparing the quantification of specific vaccinal antibodies against Newcastle disease virus (NDV) using two ELISA approaches in four seabird species (Cory's shearwater, European shag, European storm petrel and Southern rockhopper penguin). Second, we provide a simple way to determine from the distribution of ELISA values whether the assayed samples are likely to be made of a single group of individuals (likely negative) or of two groups of individuals (negative and positive). We illustrate the use of this approach with two independent datasets: NDV antibody levels following vaccination and anti‐Borrelia antibody levels following natural exposure. The practical implementation of these methodological approaches could provide a way to efficiently apply ELISAs and other immune‐based assays to address questions in the growing fields of ecological immunology and disease ecology.