Graduate Studies

 

First Advisor

Stephen Kachman

Second Advisor

Qi Zhang

Degree Name

Doctor of Philosophy (Ph.D.)

Committee Members

Kent Eskridge, Susan VanderPlas, Toshihiro Obata

Department

Statistics

Date of this Version

8-2025

Document Type

Dissertation

Citation

A dissertation presented to the Graduate College of the University of Nebraska in partial fulfillment of requirements for the degree of Doctor of Philosophy

Major: Statistics

Under the supervision of Professors Stephen Kachman and Professor Qi Zhang

Lincoln, Nebraska, August 2025

Comments

Copyright 2025, Pahalapathirage Dona Kalani Hasanthika. Used by permission

This file may be distributed and/or modified under the conditions of the LATEX Project Public License, either version 1.3c of this license or (at your option) any later version. The latest version of this license is in: http://www.latex-project.org/lppl.txt and version 1.3c or later is part of all distributions of LATEXversion 2006/05/20 or later

Abstract

We introduced couple different novel approaches to incorporate latent variable information to multivariate mixture regression models with both Gaussian and count data. We also evaluated the performance of these models with existing best approaches with simulated data from various sampling structures and also evaluated one of the models performance with rice metabolite data that provided some novel insights as well as validating existing literature about performance and behavior of these metabolites. We validated the method using extensive simulations and a real-world application. In both quantitative covariate designs and complex treatment design simulations, our method consistently outperformed established tools like limma, edgeR, and DESeq2, demonstrating a superior balance of sensitivity and precision. Application to a rice metabolite dataset confirmed its practical utility, successfully identifying key, biologically-relevant compounds. Through comprehensive simulations, we show this method consistently outperforms established tools in challenging low-replicate studies and complex, multi-group (ANOVA-type) designs, demonstrating superior sensitivity and overall performance.

Advisors: Stephen Kachman and Qi Zhang

Share

COinS