ABSTRACT
The reproducibility of SELDI (surface-enhanced laser desorption/ionization) profiling is an issue in any research that envisions SELDI profiling as the basis for development of a biomarker. How this issue is to be approached is a question with many facets. One facet is dependence of the approach on the conception of the biomarker. A second facet is adaptation of the approach to the high dimensionality of SELDI profiles. A third facet is inclusion in the approach of concepts familiar from the univariate case. A fourth facet is consideration of physical sources of variation in the measurement system.
