283537 High Throughput Genetic Screens of C. Elegans with Microfluidics and Computer Vision

Wednesday, October 31, 2012: 8:48 AM
Somerset East (Westin )
Adriana San Miguel Delgadillo1, Matthew Crane2, Peri Kurshan3, Kang Shen3 and Hang Lu4, (1)Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA, (2)Interdisciplinary Program in Bioengineering, Georgia Institute of Technology, Atlanta, GA, (3)Stanford University, Stanford, CA, (4)Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA

                C. elegans, a soil dwelling nematode, is a highly studied multi-cellular organism that offers several experimental advantages, including a short life span, ease of culture and transparency, to name a few. Additionally, its full genome and neuronal wiring diagram are known. This, combined with current molecular biological techniques, make it an excellent model system for biological study, especially for neuroscience.  Genetic screens in C. elegans have led to understanding the function of many relevant genes, but the identification of mutants with significantly altered phenotypes, easily identifiable by simple visual inspection, has reached a saturation point. Aside from the difficulty of typical screens that require manual handling and inspection of a very large number of animals, the current challenge lies on the identification of mutants with very subtle phenotypes difficult to identify by eye. Performing screens based on fluorescent reporters of very small features, such as synapses, present an exceptionally difficult scenario.

                Here, we use computer vision algorithms to objectively quantify relevant information from synapses in C. elegans from a mutagenized animal population and thus, identify mutants with subtle phenotypes. Additionally, we incorporate microfluidic devices as a platform for automated worm imaging, handling and sorting. Computer algorithms allow the identification of worms which are in the correct orientation and are suitable for imaging, as well as those which have a slightly altered phenotype and are sorted as mutants. Integrating microfluidics and computer vision we have generated a platform for automated high-throughput imaging and sorting of thousands of worms based on quantitative synapse-related features. This method enables imaging, phenotyping and sorting about 100 times faster than manual handling. Not only does this method allow performing genetic screens in a simple, automated and fast manner, but it also provides a platform for discovery of mutants which would otherwise be overlooked in a typical manual screen.

Extended Abstract: File Not Uploaded