Abstract
Random searches have been the hallmark of directed evolution and have been extensively employed in the improvement of complex or poorly understood phenotypes such as tolerance to toxic compounds in the context of cellular engineering. While genome-wide mutagenesis followed by selection or screening has been a traditional means of phenotype improvement, the list of experimental methods for cellular engineering based on random searches is rapidly expanding. Adding to the confusion is the element of chance, which lengthens the process and most notably adds to the cost of phenotypic improvement programs. Here we present a method to systematize the effort of finding superior mutants by successively improving random libraries. The method, based on the quantification of phenotypic diversity, is then used to isolate more-robust strains.