Student Theses and Dissertations

Date of Award


Document Type


RU Laboratory

Shaham Laboratory


C. elegans, directional behavior, off-food trajectories, random walk processes, daf-19 mutants


Like any mobile organism, C. elegans relies on sensory cues to find food. In the absence of such cues, animals might display defined search patterns or other stereotyped behavior. The motion of C. elegans has previously been characterized as a sinusoid whose direction can be modulated by gradual steering or by sharp turns, reversals and omega bends. However, such a fine-grained behavioral description does not by itself predict the longrange features of the animals’ pattern of movement. Using large (24 cm x 24 cm) Petri dishes, we characterized the movement pattern of C. elegans in the absence of stimuli. To collect trajectories over such a large surface, we devised an imaging setup employing an array of consumer flatbed scanners. We have confirmed quantitatively the results obtained with the scanner-array setup with a camera imaging setup, in a more stringently homogeneous environment. Wild-type worms display striking behavior in the absence of food. The majority (~60%) of the animals’ paths displays persistence in the direction of motion over length scales that are 50-100 times the body-length of C. elegans. The overall direction of movement differs from animal to animal, suggesting that the directed motion we observe might not be interpreted as a taxis to an external cue in the experimental environment. Interestingly, animals appear to exhibit directionality at large scales despite nondirectional motion at smaller scales. We quantified the extent of local directional persistence by computing the autocorrelation function of the velocities. Unexpectedly, correlations in the direction of motion decay over time scales that are much faster than the scales over which directional persistence appears to be maintained. We sought to establish quantitatively that the worm motion is, in fact, biased. To determine whether a null, random walk-like model of locomotion could account for directional behavior, we generated synthetic trajectories drawing from the same angle and step distributions of individual trajectories, and quantified the probabilities of obtaining larger net displacements than the experimental. Such a model fails to reproduce the experimental results. Moreover, the mean square displacements computed for the data display non-diffusive behavior, further demonstrating that the observed directional persistence cannot be explained by a simple random-walk model. To corroborate the hypothesis of biased movement in a model-independent fashion, we employed a geometrical characterization of the trajectories. Isotropic, unbiased walks result in paths that display a random distribution of turning angles between consecutive segments. In contrast, parsing of the worm’s trajectories yields different results depending on the segmentation scale adopted. In fact, increasing the segment size results in increasingly narrow turning angle distributions, centered around the zero. This suggests the emergence of directional coherence at long time scales. In order to investigate whether directional persistence is attained by a sensory mechanism, we analyzed the paths displayed by animals with impaired sensory function. Animals mutant for che-2, which display disrupted ciliary morphology and pleiotropic behavioral defects, exhibited non-directional behavior. Surprisingly however, daf-19 mutants, which lack sensory cilia altogether, displayed residual directionality, albeit at a lower penetrance (~20%) than the wild-type. This result suggests that directionality might implicate sensory modalities that do not require ciliary function, such as AFD-mediated thermosensation or URX-mediated oxygen sensation. Alternatively, the behavior of daf-19 mutants might imply that neural activity, but not sensory inputs, are required to achieve directed motion. Mutations in osm-9, a TRPV channel implicated in several avoidance behaviors in the worm, did not result in an observable phenotype. In contrast, mutations in tax-2/tax-4, a cGMP-gated channel required to transduce a number of sensory stimuli, resulted in loss of directionality. However, specific mutations targeting the signal transduction pathways for thermotaxis, olfaction, phototaxis, and aerotaxis, upstream of TAX-4, did not disrupt directional behavior. To get further insight into the nature of the stimulus directing the animals’ behavior, if any, we performed rescue experiments of TAX-4 function in specific subsets of neurons. In agreement with the results obtained by genetic lesions in the signal transduction pathways for thermotaxis and odortaxis, no rescue of directional behavior was observed when expressing TAX-4 in the thermosensory neuron AFD, or in the olfactory neurons AWB and AWC. Partial rescue of wild-type behavior was obtained by expression of TAX-4 in a set of five cells, which comprised the oxygen-sensing AQR, PQR and URX neurons as well as the ASJ and ASK sensory neurons, which transduce chemical stimuli and responses to dauer pheromone. To address the concern that the animals’ motion might be directed to a chemosensory cue within the plate, we investigated the correlation between path directions displayed by animals that were assayed on a same plate. We did not observe a detectable correlation between path headings, indicating that the worm is not chemotaxing to a plate-specific cue. In conclusion, our results indicate that the motion of C. elegans cannot be assimilated to a random walk, and that directional persistence arises at long times despite local nondirectional behavior. In addition, although we have not conclusively ruled out a sensorybased explanation, the genetic and phenomenological evidence gathered foreshadows the intriguing possibility that C. elegans might be achieving directional motion by relying solely on self-based information.


A thesis presented to the faculty of The Rockefeller University in partial fulfillment of the requirements for the degree of Doctor of Philosophy.

Permanent URL

Included in

Life Sciences Commons