Automatic Cell Segmentation with Zymomonas mobilis

Undergraduate Honors Thesis

Citations

Citeable URL: https://scholar.colorado.edu/concern/undergraduate_honors_theses/6q182m64g

Abstract

Zymomonas Mobilis ZM4 is a well-known industrial gram-negative bacterium known for its high sugar uptake bioethanol production. With single-cell imaging and machine learning algorithms advancing rapidly, it will be much easier to understand the metabolic system of the bacterium. A pre-cursor to metabolic manipulation of the bacterium is to evaluate the cell growth of various populations by using an inducible CRISPR-interference (CRISPRi) toolset on fluorescent and metabolic targets to determine the construction of strains capable of manipulating cell metabolism to increase bioethanol production. To evaluate cell growth, we implemented single-cell time-lapse microscopy, which revealed population heterogeneity in knockdown rate. These observations led to using a cell segmentation program (CyAn) to analyze biomass and growth rates generated from metabolic manipulations. This program was coupled with the development of cell tracking software for Z. Mobilis using dilated Convolutional Neural Networks (dCNNs) to simplify the cell tracking pipeline and to open new avenues of tracking for the future.

Creator