Response prediction of an insect's olfactory receptor neuron by using structural parameters of odorant and Self-Organizing Map

摘要

Our group has for years been studying odor approximation to express a variety of odors using a small number of odor components. The insect's olfaction is more appropriate for systematic survey than mammal's one because of its database's availability. Thus, we propose prediction method of Drosophila's ORN(Olfactory Receptor Neuron) response using structural parameters of an odorant molecule and SOM (Self-Organizing Map) mapping. We obtained ∼5000 structural parameters of odorant molecules by using Dragon (Talete s.r.l.). Then, SOM maps structural parameters of an odorant molecule onto an OR response to corresponding odor. As a result, the ORN response prediction was roughly performed. For improving accuracy, we selected important parameters based on the condition number. Then, the simulation suggests the prediction can be performed only using a few tens of parameters.