The aim of our study group is developing the software Cell Illustrator, by using hybrid functional Petri net as its basic architecture, which can be used in modeling and simulation of life development. To use this tool to estimate the debacle points of biological pathways, it is necessary to stop the considerable reaction paths by many sequential manual procedures so that it needs a huge amount of time. In order to solve this problem, we have been developing a system to modificate the HFPN into RNN. However, it still produces a low prediction rate. It was because we did not take into account the three kinds of arc of HFPN when modificating it into RNN. In this paper we report the examination we made after taking into consideration the three arcs which were introduced in the model to reform the RNN.
hybrid functional Petri net
recurrent neural networks
biological pathways