Filopodia detection and tracking. (A) Four inscribed circles (a, b, c, and d) with centers on a filopodium skeleton (green). The radius R and distance L from tip to center along the skeleton are labeled for the largest circle (d). (A′) The shortest distance to the boundary R as a function of the distance from the tip L. Vertical lines (a, b, c, and d) correspond to the circles in A. (A″) An example of a cell outline (red) and corresponding MAT (green). The red box corresponds to the view in A, B, and C. (B and C) Filopodia skeletons (blue, 1, 2, and 3) satisfying a critical radius of 2 or 4 pixels, respectively. (B′ and C′) R(L) curves for medial paths in B and C, respectively. Points where the line for Rcr intersects with R(L) curves define filopodia bases. (B″ and C″) Resulting filopodia and cell body mapping of the same cell using two different critical radii Rcr: 2.5 (B″) and 7.5 (C″). (D) An example of protrusion and retraction rate analysis for the cell in Video 2. (left) The length of five of the tracked filopodia as a function of time (blue) and corresponding smoothed curves (red). Only filopodia that went through the full cycle of protruding and retracting are chosen for analysis. (right) Maximal and minimal velocities of filopodia (cyan dots) and corresponding mean values (red lines) calculated using smooth curves in the left graph. (E) The parameter D is a measure of the distance between detected filopodium at time t (blue) and t + 1 (red). If the change in position of two filopodia in consecutive frames is less than Dcr, they are considered the same filopodium. In this example, tracking only filopodia tips would fail to correctly track filopodia in time. However, FiloTrack correctly associates the tips 1′ with 1 and 2′ with 2 (green lines) because D is computed using the skeletons, not just tips.
