UC Riverside

Chemosensory processing encodes environmental information, relaying it to neural systems that regulate key behavioral responses. This broad definition implies the study of chemosensory processing is relevant across model organisms, leading to multiple practical applications. Of interest is chemosensory processing in agricultural pests and insect vectors, since volatile organic compounds and tastants determine behavior toward humans and agriculture. Some work has been done to uncover key pathways mediating behavioral attraction and aversion in the fruit fly, Drosophila melanogaster, as well as mosquito vectors. However, the limited number of pathways that can be experimentally manipulated suggests computational methods offer a complementary method. Machine learning has been applied to successfully predict ligands of insect chemosensory receptors. But these tools have not yet been applied across sensory encoding, identification of important neural pathways for attraction or aversion, and the discovery of receptor ligands and chemical repellents. Such a comprehensive analysis pipeline is the aim of this work. Although emphasis is on insect repellent discovery, human as well as broader ecological toxicity remain highly relevant. This demands accurate in silico toxicity estimation in addition to cosmetic properties such as odor perceptual qualities that are a key consideration in designing topical formulations. Modeling of toxicological endpoints and human perceptual encoding by odorant receptors and the physicochemical features of odorants, are therefore discussed independently in detail, and later included into the repellent discovery pipeline. Ultimately, the discovery pipeline has helped identify numerous insect repellents that have desirable properties such as flavors and fragrances, has provided key insights into theories of chemosensory processing, and has been adapted to drug repurposing and discovery for COVID-19, with several top predicted compounds subsequently confirmed in vitro assays by others.