Selection is powerful yet poorly understood, which limits strain engineering efforts. Laboratory selections are based on enrichment for the fittest members of a population and dilution of the less fit members of the population. The complication here lies in defining fitness, where multiple phenotypes can fractionally contribute to overall fitness, the relative weight of each phenotype in defining fitness can change as a function of environment, and multiple genotypes are capable of conferring the same phenotype. These complications underlie the recognized complexity of traditional strain selection approaches. New genomics tools can be used to address this issue. We have developed one such tool, SCalar Analysis of Library Enrichments (SCALES), which is capable to assessing greater than 10^6 enrichment patterns corresponding to each individual clone contained within comprehensive genomic libraries. That is, SCALEs allows one to track the relative concentration of each clone, with specific genetic information on each clone, to assess enrichment and dilution patterns throughout a broad range of selections. We have used this approach to improve understanding of how i) different genotypes are capable of conferring the same phenotype, ii) different phenotypes combine to contribute to overall fitness, iii) genetic adaptations that are beneficial in one environment can be costly in another (tradeoffs), and iv) selection strategy can dictate enrichments directed at the same phenotype. Based on these efforts, we began the development of a general strategy for using SCALEs to inform the development of strain selection strategies. In this approach, we first perform a general selection that attempts to identify all of the different mechanisms that may be a work in altering fitness in a relevant host and environment. We then use SCALEs to track specific fitness altering mechanisms throughout the selection, which can be used to inform the design of a directed selection strategy that attempts to enrich for those fitness altering mechanisms most relevant to the strain engineering objective. We have demonstrated this approach within the context of engineering improved 3-hydroxypropionic acid (3HP) production in E. coli. Our general selection indicated that increased copy of carbon catabolism, transporters, or biofilm mediation genes were the primary means for increasing fitness in selections performed in a continuous flow reactor. We then redesigned our selections to enrich for transporter functions, which may work to decrease intracellular accumulation of 3HP, as opposed to the less desirable catabolic or biofilm functions. In summary, we have developed a new genomics strategy for assessing library enrichments and applied it to improve abilities to direct strain selections for the enrichment of targeted fitness altering mechanisms.

Extended Abstract Status: Not Uploaded