Abstract Search

ISEF | Projects Database | Finalist Abstract

| Print PDF

Gut Instinct: An AI-Driven Approach to Inflammatory Bowel Disease Prediction Using Microbial and Metabolite Data

Booth Id:
BMED064

Category:
Biomedical and Health Sciences

Year:
2025

Finalist Names:
Ganti, Aditi (School: La Cueva High School)

Abstract:
Today, up to 3.6 million Americans suffer from Inflammatory Bowel Disease (IBD) —a group of debilitating conditions, including Crohn’s disease and ulcerative colitis, that cause chronic inflammation in the gastrointestinal (GI) tract. The prevalence of IBD and lack of a cure underscore the need for a predictive tool that could provide insight into solutions for the disease. This project explores the interactions between basic demographic, metabolic, and microbiomic data and IBD in patients using multiple advanced AI models. Additionally, the usage of synthetic data generation introduces an aspect that could be universally applicable to the field of healthcare. The research mapped correlations between certain data labels and IBD, was able to predict IBD with accuracies up to 86%, and demonstrated the application of SMOTE and oversampling techniques for synthetic data generation. This research has the potential to enhance our understanding of IBD pathophysiology and contribute to the development of precision medicine approaches.

Awards Won: