Hemaxi Narotamo
PhD Student in Biomedical Engineering
SIPg/ISR
Summary
Endothelial cell (EC) polarity plays a crucial role in blood vessel formation and remodeling. Vascular biologists study EC polarity to understand both physiological and pathological processes associated with blood vessels. EC polarity is typically quantified as a vector between the centroids of the nucleus and the corresponding Golgi complex. Currently, these vectors are computed in 2D using 2D projections of 3D images. This is a time-consuming process and not practical for large scale analysis. To address these limitations, we have developed a deep learning-based method that jointly detects and pairs nuclei and Golgi in 3D. Our proposed method outperforms classical approaches for vector prediction on a dataset of mouse retinal microscopy images, which was acquired at the Vascular Biology and Disease Laboratory from Instituto Gulbenkian de Ciência.
Endothelial Cell (EC) Axial Polarity: a vector from the nucleus centroid to the Golgi centroid.
Example of a 2D Projection of a 3D Microscopy Image (on the left) and a 3D Sub-Volume (on the right): blood vessels in blue, nuclei in green and Golgi in red.
Note: Further details regarding the proposed methodology will be provided upon publication. The implementation code will be made publicly accessible at https://github.com/HemaxiN/3DCellPol.
