Erin Shappell

BioE PhD Proposal Presentation

Monday July 10th, 2023 – 10 AM – 12 PM

EBB 3029 (Zoom: https://gatech.zoom.us/j/98940303610)

Advisor:

Hang Lu, Ph.D.                    Chemical and Biomolecular Engineering, Georgia Tech

Committee:

Eva Dyer, Ph.D.                        Biomedical Engineering, Georgia Tech

Gordon Berman, Ph.D.            Biology, Emory University

Simon Sponberg, Ph.D.            Physics and Biological Sciences, Georgia Tech

Patrick McGrath, Ph.D.         Biology, Georgia Tech

To eat, or not to eat: A deep learning pipeline for quantifying the contribution of monoamines to feeding-related decisions in C. elegans

Hunger is the universal drive to eat. However, the decision to eat or not eat is dependent on more than just hunger signaling; organisms can sometimes experience hunger but choose not to eat. The decision to not eat despite feeling hungry can become chronic and pathological, leading to the development of an eating disorder. Despite extensive knowledge of the signals that control hunger, we do not fully understand the origins and mechanisms of pathological eating-related decision making in humans. The challenges associated with studying this problem in humans may be greatly reduced, without sacrificing biological relevance, through the study of a simpler organism such as C. elegans. C. elegans is a soil-dwelling microscopic nematode with many evolutionary connections to humans whose eating behavior has been widely studied. No group has successfully identified the source(s) of eating-related decision making in C. elegans, but the tools necessary to do so exist. This thesis will refine and combine existing technologies, including deep learning methods and a scalable microscopy system, to measure the effects of two evolutionarily-conserved monoamines on the timescale of eating- related decisions in C. elegans. To do this, eating behavior will first be measured using three methods: a newly proposed pipeline to track eating rate, Faster R-CNN to track worm motion, and DeepLabCut to track worm posture. These eating features will then be compiled into a map of eating “states” (i.e., distinct types of eating) using an existing pipeline called MotionMapper. Then, eating state maps will be calculated across worms that lack the monoamines of interest; these will be compared to identify how each monoamine controls the length, frequency, and timing of the transitions between different types of eating. We anticipate that the findings from this work will reveal a basic model of how monoamines control eating-related decisions.