Research

AI-Driven Nanomedicine Design Platform

Research descriptions:
We are developing an AI-driven platform to revolutionize the design of nanomedicines for cancer therapy. This platform integrates physicochemical properties of nanoparticles, experimental protein corona fingerprints, and multi-task machine learning models to predict key therapeutic outcomes, including tumor delivery efficiency, tumor volume reduction, and off-target organ accumulation. By learning from a curated database of
inorganic nanoparticle formulations, our model enables rapid in silico screening and optimization of nanoparticle candidates before in vivo testing. This data-guided approach enhances the precision of nanomedicine design, reduces reliance on animal models, and accelerates the translation of safe and effective therapeutics to the clinic.

Applying Machine Learning and Artificial Intelligence for Predicting ADME-tox Properties of Environmental Chemicals

Research descriptions:
Our lab applies machine learning and artificial intelligence (AI) to advance the prediction of ADME-tox (Absorption, Distribution, Metabolism, Excretion, and toxicity) properties of environmental chemicals, with a growing emphasis on micro- and nanoplastics. By
integrating quantitative structure–activity relationship (QSAR) modeling with physiologically based pharmacokinetic (PBPK) models, we aim to improve mechanistic understanding of how environmental contaminants interact with biological systems. Our
research supports data-driven risk assessment by predicting chemical behaviors such as tissue accumulation, biotransformation, and toxic effects. We are also developing AI-assisted PBPK models to address complex exposure scenarios, such as PFAS and co-exposure to microplastics. In parallel, we extend this framework to estimate tumor delivery efficiency of nanomedicines, creating a unified modeling pipeline that bridges environmental toxicology and nanotherapeutics.

AI-Powered Prediction of E-Cigarette Emissions and Toxicity

Research descriptions:
We are developing an AI-driven computational platform to address the growing public health concerns surrounding e-cigarette use. When e-liquids are heated during vaping, they can undergo thermal decomposition, producing new chemical entities with largely
unknown compositions and toxicities. Traditional experimental approaches to identify and evaluate the health effects of these compounds are time-intensive and impractical given the thousands of commercially available e-liquid formulations. To overcome this
barrier, our research integrates artificial intelligence and machine learning, particularly predictive modeling and reinforcement learning, to simulate the formation of thermal degradation products and assess their pulmonary toxicity in silico. This approach
enables rapid screening of e-liquid ingredients, identification of high-risk formulations, and recommendation of safer substitutes. Our goal is to support the development of lless harmful e-cigarette products and provide policymakers with data-driven insights to
inform public health regulations.

Development of Generative AI in Reconstructing the Structure of Unidentified Chemicals and Predicting their Combined Toxicity in Environmental metrics from Non-Target Analysis

Research descriptions:
By employing generative AI, the study aims to enhance the accuracy and efficiency of reconstruction of the molecular structures of unknown substances in environmental samples based on the detected signals from non-target analysis. Furthermore, the project will be extended to predicting the combined toxicity of these unidentified chemicals in environmental metrics. This innovative approach promises to improve our understanding of the environmental impact of unknown chemicals, offering valuable computational tools for effective risk assessment and management.