Community Research and Development Information Service - CORDIS

FP7

MechanoSensoryTrees Result In Brief

Project ID: 293604
Funded under: FP7-PEOPLE
Country: Israel

Sensory development variation from different movements

A nematode worm, Caenorhabditis elegans, has a sensory neuron system similar to humans. Using new computer vision and modelling techniques, swimming and crawling worms have been used to classify the genetics and behaviour behind different types of movement and development.
Sensory development variation from different movements
Traditionally, classification of different types of motility phenotypes of C. elegans has been based on the study of mutations and manual annotation. When an image is partitioned into multiple segments, the stack of resulting images can then be used for 3D reconstruction using algorithms.

The EU-funded MECHANOSENSORYTREES (Role of mechanosensory touch-based cues on arborization of neuronal dendritic trees) project has developed new protocols and systems to analyse videos of moving worms automatically. Investigating how touch-based sensory input determines development of sensory nerve endings, dendrites, is highly applicable in many areas including behavioural genetics and neuroscience.

The resulting MECHANOSENSORYTREES automatic segmentation system combines intensity, texture and timing to ultimately provide refined segmentations using a Markov random field model. The researchers also developed a statistical computer vision method for segmenting dendritic tree structures using laser-scanning confocal microscopy. Results showed that the system outperforms traditional methods of segmentation and provides a dendritic tree outline comparable with other sophisticated systems.

A fully-automated approach to characterise C. elegans phenotypes without the need for defining nematode-specific features was also developed. Based on the computer vision scale-invariant feature transform, this technique represents a range of different crawling movements by the nematode compared with mutants in histogram form. The algorithm successfully detected different strains and showed clusters with the expected pattern based on genetic relationships.

Research results have featured in several peer-reviewed journals including the Biophysical Journal as well as in presentations at conferences. This also included participation in international symposiums.

Defects in neuronal development and mechanosensory function can contribute to Down's syndrome and autism. Evidence is accumulating that environmental cues, as well as the genetics of neural differentiation, determines the definitive morphology of dendritic trees. A better understanding of sensory development could lead to production of targeted treatment strategies.

Related information

Keywords

Caenorhabditis elegans, modelling, automatic segmentation, dendritic tree, targeted treatment
Record Number: 188298 / Last updated on: 2016-08-18
Domain: Biology, Medicine