We analyze massive genomic datasets, like the UK Biobank 500K whole-genome sequencing data, to identify disease-associated coding variants using advanced statistical methods. Variants are mapped to transcript-specific amino acid sequences to enable subsequent analyses.
Using AI-predicted protein structures, we assess the structural and functional consequences of genetic variants. This analysis reveals alterations in protein conformation and interaction interfaces, providing a structural basis for understanding disease mechanisms and guiding structure-based drug design.
We integrate advanced large language models, such as ChatGPT and DeepSeek, to deepen our understanding of protein mechanisms and functions. These models enhance the interpretation of complex genomic data, providing clear and insightful explanations of variant impacts.