Theses and Dissertations

ORCID

https://orcid.org/0009-0007-5276-2102

Advisor

Kouba, Andrew

Committee Member

Griffin, Matthew

Committee Member

Allen, Peter

Committee Member

Hanson, Larry

Date of Degree

5-15-2026

Original embargo terms

Visible MSU Only 1 year

Document Type

Dissertation - Campus Access Only

Major

Forest Resources (Wildlife, Fisheries & Aquaculture)

Degree Name

Doctor of Philosophy (Ph.D.)

College

College of Forest Resources

Department

Department of Wildlife, Fisheries and Aquaculture

Abstract

Near-infrared spectroscopy (NIRS) is a rapid, non-invasive, and non-destructive vibrational technique that has been widely applied in food science, pharmaceuticals, agriculture, clinical microbiology, wildlife conservation, and fisheries. In NIRS, infrared light in the range of 700–2500 nm interacts with biological samples, generating spectra that reflect overtone and combination bands of molecular vibrations associated primarily with O–H, C–H, and N–H bonds, thereby providing a “fingerprint” of the sample. This fingerprint allows us to predict qualitative and quantitative properties of a biological sample. NIRS has been successfully used in live-animal studies to determine biological sex, species identity, physiological status, and disease presence, and to differentiate pathogens at the genus, species, and strain levels using chemometric and machine learning models. Despite these advances, the application of NIRS to catfish aquaculture, particularly for in vivo phenotyping and health assessment, remains limited. The aim of this research was to establish proof of principle and evaluate the feasibility of in vivo NIRS for catfish aquaculture. Specifically, the objectives were to: (i) develop and optimize spectral acquisition protocols for live catfish at different life stages and under varying handling conditions; (ii) assess the ability of NIRS combined with multivariate and machine learning methods to differentiate catfish types; (iii) evaluate the performance of NIRS-based models in distinguishing healthy fish from those experimentally infected with Edwardsiella ictaluri; and (iv) investigate the potential of NIRS, coupled with machine learning, to differentiate bacterial pathogens relevant to catfish production at both species and strain levels under laboratory conditions. Collectively, this work provides a foundational framework for applying NIRS in catfish type differentiation and health monitoring. The developed methodology can also be extended to other fish and aquatic animals for diverse physiological, diagnostic, and management applications.

Sponsorship (Optional)

This research was funded by Biophotonics project #6066–31,000-015-00D. The facility was funded by the United States Department of Agriculture (USDA) Agricultural Research Service (ARS) grant#58-6066-5-042 and the Mississippi Agriculture and Forestry Experiment Station (MAFES). The USDA National Institute for Food and Agriculture (Grant: 1005154) provided partial support for PJA. This work also utilized resources provided by the Mississippi State University College of Veterinary Medicine and the USDA-ARS Catfish Health Initiative (6066-31320-006-000D).This work is a contribution of MAFES and the Forest and Wildlife Research Center (FWRC) at Mississippi State University.

Download

Share

COinS