To ensure our therapeutic strategy is scientifically rigorous and practically viable, we implemented a comprehensive multi-scale modeling workflow. This computational framework addresses core challenges at every stage of our project: from revolutionizing the biomolecule research, design, and data analysis workflow with AMPilot, to revealing the fundamental material properties of our hydrogel (L-HBC) carrier via Molecular Dynamics Simulation, and finally, predicting the clinical efficacy of the entire system using a Reaction-Diffusion Model.
We built AMPilot, a multi-agent AI system designed to streamline our entire dry-lab workflow through three integrated, specialized agents. Its Research Agent accelerates the discovery of optimal antimicrobial peptides from vast databases. At its core, our innovative AMP Rerank Agent pioneers a new, explainable paradigm for synthetic biology; moving beyond traditional 'black-box' models, it intelligently prioritizes candidate sequences to directly reduce wet-lab costs and bridge the gap between computational design and experimental success. Finally, its Data Analysis Agent automates the complex analysis of our wet-lab data, transforming raw experimental results into actionable insights.
We utilized Molecular Dynamics (MD) Simulation to investigate the atomic-scale behaviors that govern our entire therapeutic system. These simulations revealed the precise molecular mechanisms behind two critical phenomena: first, our hydrogel's (L-HBC) smart, temperature-driven gelation, and second, its powerful and stable affinity for our chosen therapeutic antimicrobial peptide, Pexiganan. This provides a fundamental understanding of how the material carrier both forms the necessary structure and effectively anchors the therapeutic agent, allowing them to function as a cohesive and effective unit.
We developed a Reaction-Diffusion (PDE) Model to simulate our therapy's performance within a "digital twin" of a diabetic wound. This model specifically predicts the spatiotemporal delivery of the therapeutic proteins (IL-4 and VEGF) produced by our engineered yeast. By quantifying the system's critical "time-to-effect," this model provides invaluable foresight into the therapy's clinical feasibility and directly informs the design of future animal experiments.