Low-latency label-free image-activated cell sorting using fast deep learning and AI inferencing.

Classification and sorting of cells using image-activated cell sorting (IACS) systems can bring significant insight to biomedical sciences. Incorporating deep learning algorithms into IACS enables cell classification and isolation based on complex and human-vision uninterpretable morphological features within a heterogeneous cell population. However, the limited capabilities and complicated implementation of deep learning-assisted IACS systems reported to date hinder the adoption of the systems for a wide range of biomedical research. Here, we present image-activated cell sorting by applying fast deep learning algorithms to conduct cell sorting without labeling. The overall sorting latency, including signal processing and AI inferencing, is less than 3 ms, and the training time for the deep learning model is less than 30 min with a training dataset of 20,000 images. Both values set the record for IACS with sorting by AI inference. . We demonstrated our system performance through a 2-part polystyrene beads sorting experiment with 96.6% sorting purity, and a 3-part human leukocytes sorting experiment with 89.05% sorting purity for monocytes, 92.00% sorting purity for lymphocytes, and 98.24% sorting purity for granulocytes. The above performance was achieved with simple hardware containing only 1 FPGA, 1 PC and GPU, as a result of an optimized custom CNN UNet and efficient use of computing power. The system provides a compact, sterile, low-cost, label-free, and low-latency cell sorting solution based on real-time AI inferencing and fast training of the deep learning model.

Microfluidic cell sorter sample preparation for genomic assays.

Single-cell RNA-Sequencing has led to many novel discoveries such as the detection of rare cell populations, microbial populations, and cancer mutations. The quality of single-cell transcriptomics relies heavily on sample preparation and cell sorting techniques that best preserve RNA quality while removing dead cells or debris prior to cDNA generation and library preparation. Magnetic bead cell enrichment is a simple process of cleaning up a sample but can only separate on a single-criterion. Droplet-based cell sorters, on the other hand, allows for higher purity of sorted cells gated on several fluorescent and scatter properties. The downside of traditional droplet-based sorters is their operational complexity, accessibility, and potential stress on cells due to their high-pressure pumps. The WOLF® Cell Sorter, and WOLF G2®, developed by NanoCellect Biomedical, are novel microfluidic-based cell sorters that use gentle sorting technology compatible with several RNA-sequencing platforms. The experiments highlighted here demonstrate how microfluidic sorting can be successfully used to remove debris and unwanted cells prior to genomic sample preparation resulting in more data per cell and improved library complexity.