Automated characterization of cell shape changes during amoeboid motility by skeletonization

BMC Systems Biology

Table 1 Values of preset parameters in algorithms¹.

	Variable	Definition	Value used
Branch pruning	p _threshold	Length threshold to identify a short outer branch	1/10 of the length of the cell body²
	r	Ratio threshold to decide whether to combine two branches or to delete the shorter one	1.5
	PrDist _threshold	Distance threshold to delete a branch far away from the boundary	1/6 of the length of the cell body
Backward tracking	T	The maximal length of time apart by which two activities can be considered as from the same pseudopod³	50 s
	α	Weight on the first order difference in computing the cost function in tracking	0.5
	β	Weight on the spatial distance in computing the distance score	0.5
	R	Cell radius	5 μm
	Dist _threshold	Threshold on distance score to group two activities into the same pseudopod	1/10 of the perimeter of the cell boundary

¹ These values are chosen to minimize the discrepancy between the automated method and manual judgments. For the branch pruning algorithm, representative cells with different sizes and shapes are selected, and their pseudopods during movement are identified by experienced cell biologists. For backward tracking algorithm, manual tracking results from representative movies are independently acquired. Different parameter values are then tested in algorithms, and the best combinations are selected as given.
²The "length of the cell body" is computed as the average body length among all the cells from the same experiment. Thus, it is a constant when analyzing movies from the same experiment, but may vary for different experiments.
³This parameter needs to be set empirically according to specific frame rate used during imaging; the value here was chosen for movies sampled at 10 s/frame. For all other parameters, the values given here are applied for all our movies, which include different cell types, frame rate, and magnification of imaging.

ISSN: 1752-0509