where x is the unsealed data and a, b, and c are constants estimated empirically. The entire sample is then rescaled using the obtained model. LISS therefore serves as a method to both adjust for technical variation and to convert between measured Luminex intensity and more traditional Affymetrix log2-expression values.
