Development of an Agentic AI Framework for Microbiome Data Analysis

Microbiome test results are often complex and require specialized expertise to interpret. To address this challenge, an Agentic AI framework was developed to make the analysis process faster, more structured, and more accessible for general users, clinicians, and researchers. The system consists of two AI components working together. The first component, Mini AI, operates locally on the user's device and retrieves information from a curated knowledge base to provide initial explanations about detected microorganisms and the meaning of the test results. The second component, Big AI, supports deeper analysis when the information is complex or exceeds the capability of Mini AI. Users can upload their test results and submit questions through a web application, which generates coherent and context-appropriate responses to support microbiome interpretation. This project focuses on developing the Agentic AI framework to summarize and explain microbiome results from respiratory and sexually transmitted diseases. The proposed system is designed as an interpretation assistant rather than a diagnostic tool, pairing every result with traceable biomedical evidence, and establishes a foundation for future advancements in analytical accuracy, knowledge base expansion, and reasoning.