Use of AI in Basic Research: Diagnosis & Prediction of Liver Disease Progression Ondemand-Use of AI in Basic Research: Diagnosis & Progression (Part 3)

Video Transcription

Video Summary

In this final webinar of a series on AI in basic research, Dr. Wei Ouyang of KTH Royal Institute of Technology presents his pioneering work on integrating generative AI with cell biology, focusing on microscopy, imaging, and whole-cell modeling. He aims to build an AI-powered virtual cell simulator to enable in silico experiments, drug screening, and ultimately reduce reliance on animal models. The approach involves training neural network models to generate synthetic microscopy images and videos that replicate real cellular phenotypes, using diffusion transformers and extensive live-cell microscopy datasets.<br /><br />Dr. Ouyang discusses challenges including the need for large, diverse datasets to accurately simulate various cellular conditions. To address this, his lab is developing scalable cloud-based bioimage analysis platforms and collaborative annotation tools, thus facilitating data sharing and model training. He also introduces the Bioimage Model Zoo, a repository for AI models in bioimage analysis, and a novel Bioimage Chatbot that simplifies complex image analysis using conversational AI.<br /><br />A key highlight is the REEF (imaging farm), an automated, AI-guided microscope farm combining robotic sample handling and smart imaging to conduct live-cell experiments autonomously and iteratively. The system enables real-time data acquisition and analysis, closing the loop between experiment design, execution, and interpretation. Dr. Ouyang demonstrates interactive AI agents capable of controlling instruments remotely, analyzing images, searching literature databases like PubMed, drafting experimental protocols, and refining analyses through iterative human-AI collaboration.<br /><br />Dr. Ouyang envisions a future where AI models not only augment but potentially transform biological research by compressing vast experimental data into virtual cell simulations. He emphasizes the complementary role of AI and human expertise, with AI accelerating discovery while researchers provide contextual understanding. His work offers practical tools and conceptual frameworks empowering basic researchers to harness AI in their studies, making advanced bioimage analysis accessible without extensive programming knowledge.

Keywords

AI in basic research

generative AI

cell biology

microscopy imaging

whole-cell modeling

virtual cell simulator

synthetic microscopy images

diffusion transformers

cloud-based bioimage analysis

Bioimage Model Zoo

REEF imaging farm