We have created a ML-based algorithm, CeSnAP, paired with a rapid snapshot acquisition technique that serves as a simple and versatile workflow to quantify C. elegans behavioral phenotypes . CeSnAP can be used for smart analysis of videos or snapshots upon training of a convolutional neural network (C-NN). The program maintains low data overhead, eliminates the need for user supervision, and can be utilized by investigators with no computational background, greatly accelerating efforts to perform high-throughput screens.
Using CeSnAP, we performed high-throughput curling analysis of a total of 17,000 worms in order to identify drugs that ameliorate PD-like motor dysfunction in C. elegans. Our high-throughput automated curling assay can record, process, and analyze experimental data 40 times faster than the manual thrashing assay, and has allowed for high-throughput drug screening  and testing of disease mechanisms [2, 3]. The source code along with demo examples are available here.
 Sohrabi, S., Mor, D. E., Kaletsky, R., Keyes, W., & Murphy, C. T. (2020). High-throughput behavioral screen in C. elegans reveals novel Parkinson disease drug candidates. bioRxiv.
 Mor, D. E., Sohrabi, S., Kaletsky, R., Keyes, W., Tartici, A., Kalia, V., ... & Murphy, C. T. (2020). Metformin rescues Parkinson’s disease phenotypes caused by hyperactive mitochondria. Proceedings of the National Academy of Sciences.
 Yao & Kaletsky, Keyes, W., Mor, D. E., Wong, A. K., Sohrabi, S., Murphy, C. T., & Troyanskaya, O. G. (2018). An integrative tissue-network approach to identify and test human disease genes. Nature Biotechnology, 36(11), 1091.