Use of multivariate statistical methods to identify immunochemical cross‐reactants |
| |
| Authors: | Alexander E. Karu Tony H. Lin Leo Breiman Mark T. Muldoon Jean Hsu |
| |
| Affiliation: | 1. Hybridoma Facility, College of Natural Resources , University of California , Berkeley, CA, 94270, USA;2. Department of Mathematics , University of California , Los Angeles, CA, 90024, USA;3. Department of Statistics , University of California , Berkeley, CA, 94270, USA;4. Department of Statistics , University of California , Berkeley, CA, 94270, USA;5. US Department of Agriculture , Pesticide Degradation Laboratory , BARC Building 050, Room 100, Beltsville, MD, 20705, USA;6. CDFA Chemistry Laboratory Services , 3292 Meadowview Road, Sacramento, CA, 95832, USA |
| |
| Abstract: | Quantitative competition immunoassays with appropriate combinations of antibodies give consistent dose‐response patterns which may be used to identify and estimate amounts of cross‐reacting compounds. Previously reported methods of analyzing cross‐reaction patterns include multiple regression, principal components analysis and minimum estimates of variance (MEV). Four other techniques which are preferable in theory have been surveyed: discriminant analysis (DA), maximum likelihood estimates (MLE), classification and regression trees (CART), and computational neural networks (NN). MLE and simple back‐propagation neural networks can estimate the concentration, as well as the identity, of individual compounds. These four methods worked well with unfitted, unscaled data from monoclonal assays of triazines, phenylureas and avermectins. Immunoassays must be properly designed to provide adequate data for pattern recognition. Cross‐reactivity pattern analysis will make multi‐analyte, multi‐antibody immunoassays feasible for many applications in toxicology and hazard assessment. |
| |
| Keywords: | Multi‐analyte immunoassay immunoassay statistics pattern recognition |
|
|
