IMAGE-BASED SCREENING OF HIGH-PERFORMING CLONES USING PHOTOACTIVATED CELL SORTING VIA DUAL PHOTOPOYLMERIZED MICROWELL ARRAYS

摘要

We demonstrate an image-based dual-photoactivated cell-sorting method to screen high-performing Chinese Ovary Hamster (CHO) cells. Cellular organelles including mitochondria, lysosome and nuclei were stained and imaged at high resolution. Each cell was characterized by a unique signature comprised of >100 quantitatively extracted phenotypic features from the images. We show that low- and high-performing cells can be successfully distinguished by applying a machine learning-based classifier to the high dimensional cell signature matrix. Follow the cell classification, we presented a single cell sorting technology to retrieve the desired cells by dual-photopolymerization.