Program

Poster Presentations

Day 1, May 15（Tue.）　　Poster

1P-13　　PDF

Food functionality evaluation using extended-weight-updating self-organizing map

(Miyazaki Univ.)
^oKiyoko Nagahama, Mio Hirayama, Katsuhisa Kurogi, Kunihito Yamamori, Nozomu Eto, Kazuo Nishiyama, Masahito Suiko, Yoichi Sakakibara

In our previous study, we have developed a food functionality evaluation system that allows for simultaneous estimation of the multiple food functionalities from expression data of intracellular marker proteins using artificial neural network (ANN). This study attempted to simultaneously estimate the effect and action mechanism of food factor using extended-weight-updating self-organizing map (SOM) instead of ANN. To estimate human T-cell lymphotropic virus type 1-infected cell growth suppression activity, SOM was constructed using expression data of marker proteins as the feature vector, and activity value as the target vector. The SOM model could estimate the growth suppression activity of food factor with reasonable accuracy. Moreover, based on bibliographic information, food factor possessing the activity were clustered around compounds with a similar mechanism of action. These results suggest that the effect and action mechanism of food factor can be estimated simultaneously from the expression data of marker proteins using SOM.