A Hopfield Neural Network Based Algorithm for RNA Secondary Structure Prediction

In this paper a Hopfield neural network (HNN) based parallel algorithm is presented for predicting the secondary structure of ribonucleic acids (RNA). The HNN here is used to find the near-maximum independent set of an adjacent graph made of RNA base pairs and then compute the stable secondary structure of RNA. We modified the motion equation proposed in paper to reflect more biological essence of RNA secondary structure in which the ther mo dynamic parameters of base pair is used in our algorithm to control the variation rate of inhibitory and encouragement terms in the equation. Comparisons with the algorithm presented in paper and other two classical prediction methods (Zuker ´s and Nussinov ´s) show that our method is more sensitive and specific. In addition, our algorithm can be very efficient and be applied to sequences up to several thousands of base long with more degree of parallelism