Design Optimization Algorithms for Synthetic Biology Applications
Open Access
- Author:
- Halper, Sean
- Graduate Program:
- Chemical Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 04, 2019
- Committee Members:
- Howard M Salis, Dissertation Advisor/Co-Advisor
Howard M Salis, Committee Chair/Co-Chair
Andrew Zydney, Committee Member
Phillip E Savage, Committee Member
Reka Z Albert, Outside Member
Phillip E Savage, Program Head/Chair - Keywords:
- Synthetic Biology
Metabolic Engineering
CRISPR
Design Optimization
Gene Synthesis - Abstract:
- One of the central challenges facing the efforts of metabolic engineers to produce value-added biochemicals via microbial cell factories is the time and resource costs of pathway optimization. When a metabolic pathway is taken from its native host, recoded and expressed in another host, the enzymes may not be in a productive expression configuration initially, leading to low yields. Retuning these pathways scales combinatorially with the pathway length, resulting in substantial time and experimental costs when trying to produce more complex metabolites. In the cases where a good chassis organism already has the biochemical pathway of interest natively expressed, a new set of challenges arise. Redirecting metabolic flux toward a desired internal metabolite without hampering cell growth requires substantial genome edits to change the expression of multiple native genes, which imposes strain development overhead, Additionally, this genomic refactoring is usually irreversible and not tunable, limiting applicability to pathways overlapping with essential genes. A third, hidden cost to crop up during pathway optimization is the time and lost productivity cost of synthesis failures, which can make heterologous gene expression and other genetic system design substantially more complex. In this dissertation, we seek to eliminate these costly experimental hurdles via design optimization algorithms. With the Pathway Map Calculator, an automated kinetic modeling algorithm, we can replace the large experimental hurdle of exhaustive heterologous pathway characterization with a smaller mixed experimental and computational approach to develop an accurate pathway model, and use the namesake Pathway Maps to guide downstream engineering efforts toward the optimal expression of the pathway’s enzymes. In order to allow rapid, tunable regulation of native enzymes, we developed extremely long sgRNA arrays (ELSAs) in order to leverage the advantages of CRISPR-Cas9 as a means of regulating several genes simultaneously via a single genetic construct that can be ordered directly from gene synthesis providers, eliminating experimental overhead. To that end, we developed the ELSA Calculator, a design automation algorithm for producing synthesizable ELSAs, and the Synthesis Success Calculator, a classifier for predicting gene synthesis delays, to ensure rapid, reliable synthesis of ELSAs, metabolic pathways and other genetic systems.