Unsupervised Clustering

Heteromotility features can be used to identify heterogeneous motility states in a cell population by unsupervised clustering. Tools to perform unsupervised clustering are available in the analysis/ directory of the Github page. Here, we demonstrate two methods of clustering cells based on heteromotility feature outputs in R.

To begin, we generate heteromotility feature data from simulated random walks and Levy fliers.

heteromotility demo/ --tracksX demo/rw_x.csv --tracksY demo/rw_y.csv --output_suffix rw
heteromotility demo/ --tracksX demo/pf_x.csv --tracksY demo/pf_y.csv --output_suffix pf

Please see the unsupervised clustering R notebook for detailed analysis steps.

Unsupervised Clustering Notebook