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

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.