Machine learning and AI methodology developed by Sbider
CelFdrive
CelFDrive is a semi-automated pipeline to improve the speed of image acquisition in Lattice LightSheet microscopy. The figure shows maximum projected images showing phases of mitosis at 100X magnification, 488 nm (white) and 640 nm (magenta).
AI assisted image acquisition
Automated microscopy methods are an emerging field which has been facilitated by developments within both microscopy and deep learning. Lattice LightSheet microscopy is a powerful imaging method that offers high spatial and temporal resolution images, whilst minimising photobleaching. Advancements in deep learning and computer vision have facilitated networks that can classify and localise objects in close to real-time.
Graph neural network based classification
Sbider researchers have used graph neural networks to predict the location of macropinocytic cups, cell surface deformations involved in cell drinking.
Classification of cell surface features
The figure shows a Dictyostelium cell visualised by lattice lightsheet microscopy. Top row: segmented cell surfaces with PIP3 and cortical Actin levels projected onto the cell surface. The network can include geometric features such as curvature and predicts (bottom right) the centre of cups (green) and cup boundaries (blue).
Predicting tumour clocks
We use a machine learning approach called TimeTeller to estimate clock function from clock gene expression in biopsies of patient tumours.
Circadian systems biology
The cell-autonomous circadian clock is a multi-genic system modulating up to 20-50% of the transcriptome in human cells. Understanding normal clock function is essential for detecting and combatting circadian disruption and understand its impact in chronic diseases such as cancer.