Single-nuclei transcriptomics (snRNA-seq) is a breakthrough technology that maps gene activity (providing high-resolution gene expression data) with incredible detail, helping drive the future of precise, personalized medical treatment. This technology is especially transformative for analyzing specific cell types or frozen biobank tissues where it is difficult to keep whole cells intact. However, the success of this process depends heavily on one critical first step: accurately counting the cell nuclei. Currently, this is often done manually—a slow process prone to human error—which frequently leads to costly improper sample loading, wasted reagents, lower data quality and failed sequencing runs.

In this project, Dr. Colin Collins of the University of British Columbia and SnapCyte Solutions Inc. have partnered to develop an AI-powered tool that automatically counts intact nuclei from biomedical microscopy images.The team will build and validate a deep learning model to distinguish nuclei from cellular debris across various types of imaging and tissue samples—including fresh, frozen and Formalin-Fixed Paraffin-Embedded (FFPE) clinical samples. Once ready, the model will be directly integrated into SnapCyte’s software suite. To ensure accuracy, the Vancouver Prostate Centre’s Laboratory for Advanced Genome Analysis will provide diverse real-world samples to test the AI tool's performance against manual counts, commercial counters and sequencing-derived quality metrics.

The result will make snRNA-seq more accessible and efficient, especially for labs working with limited or hard-to-obtain samples. By lowering costs and improving reliability, this project will allow for a phased expansion to other academic and sequencing labs across Canada and foster more inclusive genomics research. Ultimately, this project will solidify BC’s leadership in AI-enabled life sciences, contribute to more equitable and impactful healthcare innovations and help translate genomics into health and economic benefits.