Research

Cryo-Electron Tomography Analysis

The cell is the basic structural, functional, and biological unit of all known living organisms. It is a tiny but very complicated “living machine” that can do a lot of amazing things. However, so far we have very limited knowledge about its complicated molecular machinery due to lack of high resolution and systems level data of individual cells. Cellular Cryo-Electron Tomography is an emerging imaging technique that captures the 3D electron density distribution of cells at sub-molecular resolution and at close-to-native state. Our lab aims to use cutting-edge computation, mathematics and artificial intelligence techniques, particularly those related to computer vision, machine learning and big data, to build structural organization models using such imaging data. Such modeling would be useful for getting new insights into the machinery of cellular systems. Structural patterns in cellular tomograms

The figure above shows extracted structural patterns in cellular tomograms. For details, see our paper published at PNAS.

Embodied AI for Autonomous Experimentation

The integration of autonomous experimentation and high-throughput robotics is fundamentally transforming the scientific landscape by enabling precise, reproducible, and massive-scale investigations while drastically reducing resource consumption. To realize the full potential of these automated systems, we are actively exploring solutions for closed-loop Embodied AI and robotic platforms with real-time perception and autonomous decision-making capabilities. By coupling autonomous steering engines and active learning methodologies with the continuous generation of interactable 3D digital twins, we aim at establishing rigorous, predictive solutions where automated systems can visually observe physical phenomena, simulate potential interactions, and autonomously execute the optimal experimental trajectory.

The video above is a demo of our ongoing project on constructing a robot system for lab automation. In this demo, our MRSD team has adapted a robot arm to transport a plate between an automatic liquid handler and shaker. The liquid handler and shaker are from our MSAS program.