Search

Automatic mitosis detection from video is an essential step in analyzing proliferative behaviour of cells. In existing studies, a conventional object detector such as Unet is combined with a link prediction algorithm to find correspondences between parent and daughter cells. However, they do not take into account the biological constraint that a cell in a frame can correspond to up to two cells in the next frame. Our model called GNN-DOL enables mitosis detection by complementing a graph neural network (GNN) with a differentiable optimization layer (DOL) that implements the constraint. In time-lapse microscopy sequences cultured under four different conditions, we observed that the layer substantially improved detection performance in comparison with GNN-based link prediction. Our results illustrate the importance of incorporating biological knowledge explicitly into deep learning models., (© 2023. Springer Nature Limited.)

Breast cancer (BC) cell lines are useful experimental models to understand cancer biology. Yet, their relevance to modelling cancer remains unclear. To better understand the tumour-modelling efficacy of cell lines, we performed RNA-seq analyses on a combined dataset of 2D and 3D cultures of tumourigenic MCF7 and non-tumourigenic MCF10A. To our knowledge, this was the first RNA-seq dataset comprising of 2D and 3D cultures of MCF7 and MCF10A within the same experiment, which facilitates the elucidation of differences between MCF7 and MCF10A across culture types. We compared the genes and gene sets distinguishing MCF7 from MCF10A against separate RNA-seq analyses of clinical luminal A (LumA) and normal samples from the TCGA-BRCA dataset. Among the 1031 cancer-related genes distinguishing LumA from normal samples, only 5.1% and 15.7% of these genes also distinguished MCF7 from MCF10A in 2D and 3D cultures respectively, suggesting that different genes drive cancer-related differences in cell lines compared to clinical BC. Unlike LumA tumours which showed increased nuclear division-related gene expression compared to normal tissue, nuclear division-related gene expression in MCF7 was similar to MCF10A. Moreover, although LumA tumours had similar cell adhesion-related gene expression compared to normal tissues, MCF7 showed reduced cell adhesion-related gene expression compared to MCF10A. These findings suggest that MCF7 and MCF10A cell lines were limited in their ability to model cancer-related processes in clinical LumA tumours., (© 2022. The Author(s).)