Automated tracing of the axes of beating Chlamydomonas cilia

Ganesh Srinivasan, Suphatra Adulrattananuwat, Nisa S. Foster, Jureepan Saranak, and Kenneth W. Foster

Physics Department, Syracuse University, Syracuse NY, 12344-1130 USA

To understand the network of pathways that control ciliary behavior there is a significant advantage to analyzing long records of responses to stimulation. Using relatively crude imaging of held cells with beating cilia, it has been possible to delineate the response pathways of the trans and cis cilia responsible for steering in phototaxis. With EMCCD imaging of the cilia at high spatial and temporal resolution, it is now possible to be more specific with respect to what parameters of beating are regulated. Whereas a human can relatively easily trace cilia in images, with slightly noisy images it is a very hard problem for a computer. Successful computer tracing of curves depends on providing specific information to the computer and enhanced images. Enhancement involves image filtering, spectral processing, and thresholding. The most important information to provide is the order of the points along each cilium. Hence, using a series of images we estimate the pivot point in the held cell body for each cilium at its base and provide an initial estimate of the traces for the first image of a series. The actual curve is found by optimizing the fit of a set of points along each cilium using an anchored modified fuzzy c-means algorithm. With a spline fit of these points and transformation to the domain of ciliary coordinates we extract the biologically relevant parameters for the ciliary motion. Subsequently, initial estimates for the following images are then automatically generated based on the fitted proceeding image and our knowledge of the potential patterns of ciliary beating.

e-mail address of presenting author: kwfoster@syr.edu