Bounds on the efficiency of two-stage group testing

Abstract

In biotechnology screening applications for peptide and cDNA libraries and for monoclonal antibody generation, there are severe limitations on the number of stages of group testing that can be performed. The authors provide exact formulas for the screening efficiency of several such assaying strategies, including the popular rows-and-columns technique. Also, they upper bound the maximum efficiency of any two-stage strategy and use random test selection to show that there exist two-stage group tests that perform close to the authors' upper bounds.