SLIM: Stochastic Lineage-based Iterative Minimization

Citation

Lipschitz, Mikel (2026) SLIM: Stochastic Lineage-based Iterative Minimization. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/q80m-3t75. https://resolver.caltech.edu/CaltechTHESIS:08212025-162748066

Abstract

The bacterial genomes we encounter today have been shaped by billions of years of genome altering events which involve rewriting, addition, and removal of genomic elements. The resulting product is a complex network of interactions composed of elements which are defined by contextual necessity. The elucidation of a minimal set of elements, comprised of just those essential for sustaining life, has long been sought after. This set, or minimal genome, has been proposed to be a representation for the foundation of life itself. Genome minimization, the pursuit of this foundation, is a process by which genomic segments deemed unnecessary, or non-essential, in an environmental context dependent manner are identified and removed, leaving only DNA that provides the cell with the resources and processes it needs to stay alive and reproduce. Numerous genome minimization efforts have been undertaken previously. However, each of these studies has resulted in the generation of a single genome-reduced strain derived from a single wild-type bacteria in a single environment. While these methods have shown a great deal of promise in their ability to identify foundational genomic pieces in this extremely narrow context, they lack the throughput and generalizability to identify foundational pieces of all bacterial life.

Building upon prior genome minimization efforts, we developed SLIM (Stochastic Lineage-based Iterative Minimization), a modular genome reduction system designed for unbiased DNA removal, high-throughput parallelization and cross-species compatibility. In this study we utilize SLIM to generate a library of ten genome-reduced E. coli strains. We then rigorously interrogate the library to identify patterns in deleted segments. We assess the effects that these deletions have on remaining genomic components and explore how these effects can result in substantial fitness changes in different environments. Finally, we demonstrate the modularity of SLIM by generating two additional libraries of genome-reduced strains from two phylogenetically distinct parent bacteria, S. flexneri and P. putida. This work highlights the power and promise of generating diverse libraries of genome-reduced strains; substantially expanding the number of minimized genomes that can be achieved, while simultaneously reducing the time to generation.