Versatile Image Assisted Cell Sorting by Selective Trapping with Spatio-temporal Multiparameter Targeting

Current cell separation techniques often fall short in terms of flexibility, requiring intricate setups, significant starting cell quantities, and they also come with size constraints. This limits their applicability across varied objectives and experimental setups. Introducing 2D-SIGMAT, our new approach uses expansive imagery to analyze hundreds of cell visuals concurrently, rather than one at a time. This allows us to pinpoint individual cells within dense populations using real-time light-induced cell trapping, achieving high efficiency. This system, which emphasizes both time and space properties, reduces initial sample preparation, eliminates the need for specialized microfluidic tools, and works with small sample sizes. Being compatible with standard and fluorescent imaging, sophisticated image analysis, and advanced machine learning, the technique boasts a recovery rate of up to 98% when sorting cells via YOLOv5, processing up to 2000 cells per second. Additionally, our research proves that a regular microscope with a UV projector can be transformed into a multifaceted cell sorting tool, offering unique scan-and-select sorting and catering to an array of samples, from singular cells to groups of cells.<br>