Metabolic engineering has been very successful in addressing pathway optimization for localized branches, but multigenic or poorly understood traits have remained a challenge. To overcome these limitations, we have successfully delivered a method that introduces global transcriptomic changes, thereby generating diversity at the phenotypic level. This method contrasts with others for its easy implementation and for the tractability and transferability of the modifications. Furthermore, the global nature of the generated diversity imparts significant versatility to the resulting libraries. In addition, we have worked on quantifying the potential for finding improved strains in transcription factor and similar libraries. Having such objective measures will aid immensely in prioritizing the screening of different libraries, especially when combinatorial studies are considered. These metrics are based on evolution principles. To date, we have used these methods to measure the effect of mutation frequency of sigma factor libraries on phenotypic diversity, and to compare these libraries with those constructed through NTG-mutagenesis (commonly used in classical strain improvement).