Microscopic screen and cell death pathways

In image-based studies like cell research, crucial information in cell objects must be accurately extracted for further analysis and linked to biological pathways. Instance segmentation is one technique used for this purpose. In our study, investigating cell death and regulatory networks through a genome-wide microscopic screen, we trained nuclei segmentation models on a small, well-annotated dataset paper, achieving F1-scores of 89% and 83%. We used the superior U-Net model to segment 5 million image frames, revealing correlations between extracted features and key genes linked to cell death pathways like apoptosis.