Tracking the Swimming Motions of C. elegans Worms with Applications in Aging Studies

Abstract

Abstract. Quantitative analysis of the swimming motions of C. elegans worms are of critical importance for many gene-related studies on aging. However no automated methods are currently in use. We present a novel training-based method that automatically tracks and segments multiple swimming worms, in challenging imaging conditions. The position of each worm is predicted by comparing its latest motion with a set of previous observations, and then adjusted laterally and longitudinally to fit the image. After segmentation, a variety of measures can be used to assess the evolution of swimming patterns over time, allowing a quantitative comparison of worm populations over their lifetime. The complete software is being evaluated for mass processing in biology laboratories.

Vision-based Kinematic Modeling of a Worm's Posture

“…The match cost of a part with 12 different possible orientation angles (15°, 30°, 45°, … , 180°) at every possible integer position was computed by convolving the binary worm image with a convolution kernel composed of a match rectangle with different orientation angles. Restif and Metaxas [9] modeled a worm as an open polygon represented by a central body line, and the line is composed of a number of segments and vertices. At each vertex, the local base vectors were defined as the unit vectors parallel and perpendicular to the segment between two adjacent vertices as illustrated in Fig.…”

“…In order to achieve high modeling accuracy, the number of segments should be large, but this will increase the complexity of the process. For example, in [9], 100 vertices were sampled to model a small (1 mm long) worm. A further difficulty is that the length of a worm varies while its soft body moves, and rigid segments of the model need to be added or deleted.…”

“…3 for a robot arm having three degrees of freedom, its kinematics can be described using only two of the four parameters, a and q, because the other two parameters are always zero. In the existing methods [8,9] described in Section II.1, a worm is modeled as an open-loop articulated chain with rigid segments. This is the same case as that of a robot manipulator that has been modeled as an open chain of multiple solid links [10].…”

“…The arc angle f is computed by f =s/r=ks (9) where k is the curvature. We can represent any link segment in the model by specifying only two parameters, one for length and the other for angle: s and f for a curved link, respectively, or a and q for a straight link, respectively.…”

“…For nematode worms, a common approach is to represent the body by an open chain of rigid segments [8,9]. However, for a soft, flexible, and long animal such as a worm, the number of segments can be quite large, and this will cause some problems in processing time, algorithm design, and model complexity.…”