The identification of adipose tissue transcriptomic biomarkers for metabolic conditions; The Clinical and Social Implications of Genetic Biomarker Research on Metabolic Disorders

In a world increasingly defined by precision and personalization, the intersection of engineering and ethics is not just inevitable, but it is essential. This Capstone Project focuses on developing a computational model that identifies clinical outcome markers using biological data from the METSIM (METabolic Syndrome in Men) study. This project aims to enhance the diagnostic potential of clinical research by improving how relevant genes are selected and analyzed, ultimately aiding personalized treatment strategies for metabolic disorders. In parallel, the STS research paper investigates how advancements in genetic biomarker research influence healthcare equity, with a specific focus on metabolic diseases such as Type 2 Diabetes (T2D). This paper seeks to uncover how innovations like the Capstone model can affect access, affordability, and fair implementation in medical care. Together, these projects reflect a commitment to technological advancement and maintaining proper ethical considerations.
The Capstone Project was undertaken to address a major bottleneck in the translational pipeline: the challenge of identifying which genes, among thousands, are relevant to clinical outcomes. By applying Partial Least Squares Regression (PLSR) with feature selection using the 'mixOmics' package in R, I aimed to create a predictive model that could sift through high-dimensional data to find meaningful associations between gene expression and clinical markers like BMI, waist-hip ratio, and insulin resistance. The model was tailored to avoid overfitting and emphasize interpretability, which is crucial for clinical translation.
The model yielded a focused set of gene candidates that showed strong correlation with key metabolic traits. These findings suggest that this approach can effectively narrow down potential targets for future therapeutic and diagnostic research. In addition to identifying relevant biomarkers, the top gene candidates were cross-referenced with existing drug databases, and it was discovered that several are targeted by FDA-approved drugs used to treat metabolic disorders. This indicates that this model not only pinpoints meaningful genetic drivers but also has the potential to refine existing treatments for obesity and T2D by targeting multiple shared pathways. The project not only confirmed the feasibility of using multivariate techniques for gene prioritization but also highlighted the importance of model transparency and parameter tuning.
The STS paper is driven by the question: How do advancements in genetic biomarker research influence healthcare equity, particularly in addressing metabolic disorders like T2D? This question is significant because while personalized medicine promises better outcomes, it may also deepen disparities if access to these innovations remains uneven. I approached the question using multiple methods focused around the Actor-Network Theory, incorporating a literature review on genetic research impacts, a case study analysis of METSIM and similar genomic initiatives, and a policy analysis to assess gaps in equity-focused regulations.
Through these methods, it was found that while studies like METSIM provide invaluable data, their lack of ethnic diversity and limited clinical integration risk reinforcing systemic inequities. For example, many genetic models are based on European ancestry populations, leading to reduced accuracy in other demographic groups. Furthermore, the cost of genetic testing and insufficient provider training can hinder equitable access to precision treatments. The paper concludes that to make genetic innovation both impactful and inclusive, future research must incorporate diverse populations, and policy must address structural barriers to access and implementation.

School of Engineering and Applied Science

Bachelor of Science in Biomedical Engineering

Technical Advisor: Mete Civelek

STS Advisor: Pedro Augusto Francisco

Technical Team Members: Dhanush Banka