Machine learning segmentation tool trained on synthetic data for tracking cytoskeleton polymerisation and depolymerisation
Elangovan, K.; Spelman, T.; Dupouy, G.; Singh, G.; Chaboute, M.-E.; Jonsson, H.
Show abstract
The cytoskeleton is important in controlling the growth and morphology of plant cells, so tracking its morphological changes is essential. Here, we develop a new machine learning based segmentation tool for microtubules (MTs), which can distinguish between polymerised and depolymerised fibres. To circumvent the low abundance of data, we trained on synthetic images of microtubules from a computational micro-tubule model, pre-processed to reproduce microscope effects and partial depolymerisation. We used this tool to investigate how the MT network in an Arabidopsis thaliana root hair cell repolymerises after depolymerisation under Oryzalin (OZ) drug treatments. Specifically, we show the network initially repolymerises from the shank region. This work demonstrates the viability of using synthetic data to train machine learning systems handling cytoskeletal image data.
Matching journals
The top 9 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Shape-to-graph Mapping Method for Efficient Characterization and Classification of Complex Geometries in Biological Images 94%
- DeLTA: Automated cell segmentation, tracking, and lineage reconstruction using deep learning 94%
- Multiple neural network approaches, including use of topological data analysis, enhances classification of human induced pluripotent stem cell colonies by treatment condition 94%
Similar papers in this journal
- Caveolae and scaffold detection from single molecule localization microscopy data using deep learning 94%
- Small hand-designed convolutional neural networks outperform transfer learning in automated cell shape detection in confluent tissues 94%
- Deep Learning Classification of Lipid Droplets in Quantitative Phase Images 94%
Similar papers in this journal
- Global alignment and local curvature of microtubules in mouse fibroblasts are robust against perturbations of vimentin and actin 93%
- The nuclear lamin network passively responds to both active or passive cell movement through confinements 92%
- Spatially heterogeneous dynamics of cells in a growing tumor spheroid: Comparison between Theory and Experiments 89%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.