Michael-shannon/cellPLATO: CellPLATO 2024

This is the 2024 version of cellPLATO (cell Plasticity Analysis Tool) developed in collaboration with @tsloan1377 with help from @quantumjot . CellPLATO is a semi-automated morpho-kinetic analysis tool for creating classes of motion and morphology from cell tracking data.

It takes cell tracking data as an input, makes motility, morphology and clustering measurements for each cell at each timepoint, then outputs statistical plots of difference and plots of difference over time (inspired by those developed in R by @JoachimGoedhart ).

Then, dimensionality reduction and cluster analysis are used to separate cells into behavioural classes based on their morpho-kinetic measurement between single timepoints or in time windows, producing a fingerprint of behaviour types per biological condition.

Trajectories of behaviour are derived from clustering on sequence similarity, comparing strings of single timepoint behaviours.

For visualization, there is an emphasis on de-abstractification. Thus each trajectory is visualizable on the raw data, and the top contributing features are listed in order for each trajectory class.