We develop full-stack bioanalysis tools with the goal of transforming complex biological experiments into scalable, automated workflows. Our work is enabled by the integration of advances in chemistry, materials, microfabrication, microfluidics, robotics, and AI. Current research projects include the following:
Robotic Automation for 3D Cell Culture
Drug discovery relies heavily on biological models to evaluate effectiveness before testing in patients. Traditional models based on 2D cell culture, where cells grow as flat layers, often poorly predict drug efficacy, contributing to high failure rates in drug development. 3D cell culture addresses these limitations by growing cells in three-dimensional multi-cellular assemblies, such as spheroids or organoids, that more closely replicate cell-cell interactions, mechanical cues, and biochemical gradients. Despite these advantages, 3D cell culture remains technically challenging, labour-intensive, difficult to reproduce, and difficult to scale-up. We are developing new experimental methods and software pipelines to automate and scale-up 3D cell culture. By combining engineered culture substrates, robotics, high-throughput microscopy, and AI-powered data analysis, we aim to dramatically improve throughput and reproducibility to enable scale-up drug testing using physiologically relevant 3D tissue models. Students will work at the interface of design, computer vision, and biomedical technology to solve scalability problems in modern drug discovery.
AI-powered Drug Screening
Advances in artificial intelligence are transforming how biological information can be extracted from microscopy images. Modern AI tools can extract subtle cellular features and phenotypic patterns far beyond what is possible using human cognition. Recently, we developed a method, called “regularized imaging” to train AI models to identify cellular features that are generalizable across cell types and preparation methods. We are using this framework to develop next-generation drug screening workflows that avoid the use of fluorescence staining and perform time-resolved analysis to provide more informative drug response profiles. We are extending this work to 3D cell culture, where conventional imaging quickly becomes a bottleneck. Our AI-powered approach aims to dramatically reduce imaging time to increase screening throughput while lowering cost. Students involved in this project will tackle challenges in microscopy, computer vision, and AI model development.
Related publications:
Microfluidics for Red Blood Cell Biomechanics
The deformability of red blood cells (RBCs) is a potential biomechanical property that can serve as a biomarker for certain diseases, as well as for assessing the quality of donated blood used in transfusions. We have a long-standing research program focused on developing microfluidic devices to assess RBC deformability with high sensitivity and throughput. The next phase of our research aims to translate these technologies towards clinically deployable platforms suitable for routine analysis in clinical laboratories and blood banking settings. Students involved in this project will work in design, microfluidics, and biomedical assay development to gain experience at the intersection of engineering, diagnostics, and translational biomedical research.
Related publications:
Red Blood Cells with Reduced Deformability are Selectively Cleared from Circulation in a Mouse Model
Assessing Red Blood Cell Deformability using Deep Learning
Deformability based Sorting of Stored Red Blood Cells Reveals Donor-dependent Aging Curves
Image Cytometry and Microscopy-based Cell Separation
The ability to separate specific cells to analyze them in isolation is a fundamentally important capability in bioanalysis. For the past ~50 years, flow cytometry has been the dominant cell separation technology. While powerful, flow cytometry has several important limitations, including (1) shear stress that can damage sensitive cells; (2) significant cell loss during processing; (3) difficulty handling non-spherical cells, fragile cells, and cell clusters; and (4) an inability to separate cells based on morphology or sub-cellular structure. Many of these limitations can be addressed using microscopy-based approaches, but existing microscopy-guided methods are very low throughput, limiting their practicality for many downstream analytical workflows. We are developing microscopy-based cell separation technologies to expand the toolbox for cell separation and downstream analysis. Students involved in this project will work on design, optics, biomaterials, microfabrication, and application development.
Related publications: