Caenorhabditis elegans is a small (1 mm) free-living nematode worm that offers unique advantages for investigating fundamental problems in biology. The developmental and anatomical characterization of this animal is unparalleled in the metazoan world. The complete sequence of cell divisions that occur as the fertilized egg develops into the 959-celled adult are known [1, 2]. Furthermore, the description of neuronal connectivity in C. elegans is exceptionally detailed. Serial section electron microscopy has identified the pattern of synaptic connections made by each of the 302 neurons of the animal (including 5000 chemical synapses, 600 gap junctions, and 2000 neuromuscular junctions), so that the full "wiring diagram" of the animal is known [3, 4]. Indeed, the neural circuit of C. elegans shows the characteristics of a "small-world" network [5]. Despite such a simple nervous system, animals display a rich repertoire of behaviors including elaborate responses to chemical, mechanical and thermal cues and specific locomotory patterns [6, 7]. C. elegans is the first metazoan for which the genome was sequenced to completion [8]. The wealth of information on the biology of the organism, coupled with the broad range of genetic and molecular techniques applicable in C. elegans, allows in-depth studies of how genes specify and control neuron function to generate behavior [7, 9]. To this end, computational tools that facilitate the detailed analysis of nematode locomotion are highly desirable.
C. elegans sinusoidal locomotion ensues from alternate contraction and relaxation of dorsal and ventral body wall muscles, which generates a canonical sinusoidal pattern of movement [3, 9]. The arrangement of the body wall muscles and their synaptic inputs restricts locomotion to dorsal and ventral turns of the body. While, in principle, movement of the worm body is restricted in a two-dimensional space and it resembles a sinusoidal wave, many factors could affect its behavior. Numerous mutations disrupt normal sinusoidal locomotion in C. elegans, resulting in animals with movement defects ranging from total paralysis, to severe uncoordination, to subtle and almost imperceptible irregularities in movement [10–12]. As a result, the rate and direction of movement and the shape of its trajectory may change dramatically leading to more complex patterns. Some other factors affecting locomotion are the processes of feeding, egg-laying and mating, environmental stimuli, animal age and treatment with chemical substances [13–15].
While in some cases, behavioral alterations pertaining to animal movement are pronounced and may easily be described qualitatively, frequently such alterations are rather subtle or even imperceptible by simple observation. Thus, to obtain a better insight into a variety of behavioral effects and elucidate the mechanism governing the underlying processes, a more systematic analysis is required through approaches that provide precise quantitative information.
Various automated systems have been described aiming to track single or multiple worms and to study quantitatively their locomotion and behavior [13, 16–20, 23–26]. These systems offer the capacity to calculate global direct measurable parameters such as position of the animals and movement paths or indirect parameters such as speed, change in direction, shape, wavelength, and amplitude. In principle, analysis is restricted to the location of the head and the tail while the rest of the body is not investigated thoroughly [13, 16–20, 23–26]. Some of the systems are designed to observe and analyze locomotion of multiple animals at low magnification [13, 24–26]. However, because in these systems observation is conducted at low magnification, the detailed path of animal movement cannot be studied. Alternative computer tracking methods have been developed to overcome these constraints in which higher magnification is used. To maintain the animal within view, systems are equipped with a tracking program designed to control the movement of a motorized stage to keep the worm in sight [18]. The accuracy of the information obtained depends on the mechanics of the system and the integration with the microscope and camera. Other systems are used to examine more complicated behaviour, involving bends and reversals, however, only video sequences with worms in sight are analysed [20]. Most of the computer-driven systems [13, 16–20, 23–26] perform an automatic tracking and feature extraction without allowing the user to intervene and define regions of interest, set thresholds, accept or reject information, process data easily, or modify the computer algorithm and they usually assume a deep knowledge of programming languages. Additionally, the majority of the methods produce data that need to be interpreted independently and do not yield a complete picture of animal behaviour. To our knowledge, only one system has been developed [23] that provides a Graphical User Interface (GUI) to assist in a more comprehensive analysis. It is designed to control tracking and recording of the animal and subsequently illustrate the progress of the recognition process rather than offer the basis for a systematic quantitative analysis of the locomotionary behaviour.
In this paper, we present Nemo, a computational tool for obtaining quantitative information about nematode movement. This tool is designed to track deformable objects from a video sequence in high resolution. The algorithm we developed initially extracts morphological features and proceeds with segmentation of the animal body, retrieving information related to the position of the centre of mass of each body section separately. Segmentation allows movement details, body thickness and other information about any section of the worm to be easily acquired with minimum intervention from the user. A routine has also been integrated to compute the displacement of the image due to movement of the camera in order to keep the animal in sight. A particular advantage of our system is that it allows the user to choose regions of interest and compute specific locomotion features, related to these regions. This enables viewing image information for any part of the animal in the form of plots and histograms depicting the magnitude of particular movement parameters. Thus, features such as the wavelength, amplitude and direction change can be calculated in regions of interest.
In addition we have developed a GUI that automates the analysis and enables researchers to collect movement data accurately. While, the algorithm has originally been built for the study of simple behaviors such as the sinusoidal locomotion of the animal it can be readily generalized to process and describe more complex movement patterns.
Data acquisition
We followed standard procedures for C. elegans culture and maintenance [21]. The strain utilized in this study was the wild type Bristol isolate N2. Nematode rearing temperature was kept at 20°C. Animals were grown on agar plates with nematode growth medium (NGM), seeded with the bacterium Escherichia coli as a food source. For videotaping, 5–10 gravid adult worms were placed on fresh, seeded NGM plates and allowed to move freely for 30 min before observation. Animals were imaged via a Zeiss Stemi SV11 stereomicroscope (Carl Zeiss, Jena, Germany) with a Moticam 2000 CCD camera (Motic Instruments Inc., Richmond, Canada), at a resolution of 800 × 600 pixels and a frame rate of 40 fps. Video files of moving nematodes were generated using the software package accompanying the camera (Motic Images Plus 2.0; AVI, audio video interleave format).