Ke Zhai

Profile

AI-enabled vulnerability assessment of fungal proteins using multi-omics and multi-relational knowledge graphs

This project focuses on developing a computational approach to evaluate protein vulnerability and function across fungal species. By integrating high-dimensional proteomic and transcriptomic data into a unified graph framework, it captures the physical and functional interactions that define cellular states. The aim is to learn a more complete representation of biological systems, identifying important protein properties within their systemic context rather than in isolation.

To achieve this, the project maps diverse multi-omics layers and biological relationships into a heterogeneous Knowledge Graph. By combining this graph architecture with Multi-modal foundation model techniques, the framework is designed to extend insights from model organisms to industrial data. Predictions from the model are intended to be supported by Syngenta’s wet-lab validation, ensuring that the computational results remain grounded in experimental evidence. This integrated approach aims to provide a scalable strategy for studying complex protein properties and linking computational predictions with practical biotechnology applications.

Research