There is a great deal of interest in developing companion diagnostics for many therapies, including (eg) oncology due to the side effects of chemotherapy and other therapeutics. Many of these assays may be complex by nature due to the multi-faceted nature of related pathways and also because of anticipated robustness. In developing these complex assays (which may involve microarrays), there are many important considerations, some of which have a statistical component: the assay’s prospective performance compared to its performance retrospectively during initial development; the types of tests and processes required to validate its performance; measurement platform technology and its stability over time.