Date of Award

1-1-2017

Language

English

Document Type

Master's Thesis

Degree Name

Master of Science (MS)

College/School/Department

Department of Chemistry

Content Description

1 online resource (ii, v, 63 pages) : illustrations (some color)

Dissertation/Thesis Chair

Daniele Fabris

Committee Members

Daniele Fabris, Cara Pager, Alexander Shekhtman

Keywords

direct infusion electrospray ionization, machine learning, mass spectrometry, post-transcriptional modification, RNA, Salmonella, Mass spectrometry, Microorganisms, Genetic transcription, Machine learning, Mass Spectrometry, Transcription, Genetic

Subject Categories

Biochemistry

Abstract

RNA post-transcriptional modifications (PTMs) are dynamic features that can be up- or down-regulated by the health and metabolic state of a cell. These covalent modifications are installed and removed on RNA nucleosides by enzymes controlled by the activation and deactivation of specific genes. The goal of this research was to demonstrate that RNA PTMs can serve as a unique feature for the classification/identification of microorganisms. We utilized a scheme based on electrospray ionization mass spectrometry (ESI-MS) to obtain global PTM profiles from total RNA extracted from various microorganisms in optimal growth conditions as well as Salmonella typhimurium (S. typhimurium) spiked into a complex biological matrix, raw milk. The mixture of S. typhimurium and raw milk presented a challenge due to the presence of mammary gland somatic cells and bacteria present a priori to spiking. A gradient boosting machine learning algorithm was used to rationalize the experimental data provided by our ESI-MS platform by unambiguously differentiating, classifying, and identifying organisms based on their global PTM profiles. The performance of machine learning framework was measured by leave-one-out cross-validation. We demonstrated that the dynamic nature of PTMs can provide a new outlook on organism classification/identification that is typically performed with highly conserved features, such as the sequence for the small rRNA subunit. Our results showed that uncontaminated raw milk can be differentiated Salmonella contaminated raw milk based on global PTM profiles. This provides strong evidence that the technology employed in this study can lay the groundwork for uses in food safety. Although, this exploratory work only considered Salmonella contaminated milk, the methods promoted here can be extended to interrogate various organisms in many food sources.

Download

Included in

Biochemistry Commons

Share

COinS