ORCID

https://orcid.org/0000-0003-1928-7913

Date of Award

Fall 2024

Language

English

Embargo Period

10-16-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Chemistry

Program

Chemistry

First Advisor

Igor K. Lednev

Second Advisor

Alexander Shekthman

Third Advisor

Jeremy Feldblyum

Committee Members

Entesar Al-Hetlani

Keywords

Raman Spectroscopy, Machine Learning, Medical Diagnostics, PFAS Environmental Contamination, Species Authentication, Forensic Analysis

Subject Categories

Analytical Chemistry | Data Science | Multivariate Analysis

Abstract

There is an urgent need for rapid and on-site analytical methods to address various environmental and forensic challenges. This dissertation explores the application of Raman spectroscopy combined with machine learning for three critical areas: evaluating perfluoroalkyl or polyfluoroalkyl substances (PFAS) exposure in wildlife, authenticating tuna species in the seafood industry, identifying body fluids at crime scenes, and medical diagnostics.

First, Raman spectroscopy and support vector machine discriminant analysis (SVMDA) were employed to detect PFAS levels in the blood plasma of smallmouth bass from two different tributaries, rivers, or streams that flow into larger bodies of water. The method achieved 100% accuracy in differentiating between samples containing either low or high PFAS concentrations, highlighting its potential as a diagnostic tool for fish health.

Second, the capability of Raman spectroscopy to authenticate three commonly sold tuna species—yellowfin, bluefin, and escolar was assessed. The developed method accurately identified the species using partial least squares discriminant analysis (PLSDA), demonstrating its value as a quality control tool in the seafood industry to prevent fraudulent labeling.

Third, Raman spectroscopy combined with Random Forest machine learning was utilized to identify six main body fluids and distinguish them from 49 potential environmental interferences that may be discovered at crime scenes. The method achieved 100% accuracy in external validation, indicating its effectiveness as a rapid, non-destructive technique for body fluid detection, which is crucial for criminal investigations.

Lastly, the detection of Alzheimer’s disease (AD) was explored using a novel screening method combining Raman spectroscopy with deep learning, specifically Artificial Neural Networks (ANNs), to analyze blood plasma. The ANN model demonstrated high accuracy, sensitivity, and specificity in distinguishing between moderate AD patients and healthy controls. This integrative approach offers a promising, non-invasive, rapid, and accurate screening method for early AD detection, potentially enhancing diagnostic capabilities and paving the way for earlier and more effective therapeutic interventions.

Overall, this research showcases the versatility and efficacy of Raman spectroscopy and machine learning in providing reliable, accurate, and rapid analysis across different fields. Thus offering significant advancements in environmental monitoring, food authentication, and forensic science.

License

This work is licensed under the University at Albany Standard Author Agreement.

Available for download on Friday, October 16, 2026

Share

COinS