Publication Date

2025

Document Type

Thesis

Committee Members

Tomojit Ghosh, Ph.D. (Advisor); Krishnaprasad Thirunarayan, Ph.D. (Committee Member); Tanvi Banerjee, Ph.D. (Committee Member)

Degree Name

Master of Science (MS)

Abstract

This thesis investigates the application of Generative AI models, mainly Generative Adversarial Network (GAN) models to high dimensional and low sample size biological datasets like Motion Sickness, Breast Cancer, Crohn, and Melanoma. We utilized and compared three generative AI frameworks: Vanilla GAN, Wasserstein GAN (WGAN), Locality-Sensitive Hashing GAN (LSH-GAN) and Omics GAN. To address the challenges associated with high-dimensionality and low sample size, which was leading to very poor outputs of biological synthetic samples, we came up with an approach to stop the model when it reaches its saturation level. That is, we printed the loss plots to see where the model was getting the common issue of mode-collapse in order to better identify how many numbers of epochs a particular model would need for a particular dataset. We validated the quality of the synthetic samples by projecting them into 2/3-dimensional Principal Component Analysis (PCA) space along with the original samples and visually checking the closeness of the generated samples with the original ones. We analyzed the effect of training epochs on these High-dimension and low-sample-size datasets and found that early stopping is helpful to generate good quality synthetic samples. Results demonstrate that the best performing generative AI model was WGAN, showing superior results in preserving the biological structure and creating high quality synthetic data with the help of PCA plots by visually comparing the original and synthetic generated samples. This work provides a comprehensive framework for generative modeling in computational biology and applications in data augmentation (synthetic sample generation) for high-dimensional, low sample size omics datasets.

Page Count

64

Department or Program

Department of Computer Science and Engineering

Year Degree Awarded

2025


Share

COinS