Abstract
Molecularly imprinted polymers have been shown to be useful in competitive biomimetic binding assays. Recent developments in materials science have further enhanced the capabilities of imprinted polymers. Binding assays, biological and biomimetic alike, owe their usefulness to their selectivity. The selectivity of competitive binding assays has been characterized with the cross-reactivity, which is usually expressed as the ratio of the measured IC50 concentration values of the interferent and the analyte, respectively. Yet this cross-reactivity is only a rough estimate of analytical selectivity. The relationship between cross-reactivity and analytical selectivity has apparently not been thoroughly investigated. The present work shows that this relationship depends on the underlying model of the competitive binding assay. For the simple but widely adopted model, where analyte and interferent compete for a single kind of binding site, we provide a simple formula for analytical selectivity. For reasons of an apparent mathematical problem, this formula had not been found before. We also show the relationship between analytical selectivity and cross-reactivity. Selectivity is also shown to depend on the directly measured quantity, e.g., the bound fraction of the tracer. For those cases where the one-site competitive model is not valid, a practical procedure is adopted to estimate the analytical selectivity. This procedure is then used to analyze the example of the competitive two-site binding model, which has been the main model for describing molecularly imprinted polymer behavior. The results of this work provide a solid foundation for assay development.