Determining the Secondary Structure of Elapid Toxins using Multi-Layer Perceptrons and Kohonen Networks

Abstract

In this paper, a two-stage neural network consisting of a feed-forward neural network and a Kohonen self-organizing map, has been used to predict secondary structure. We have applied our methods to determine the structure of 245 proteins containing neurotoxins, cytotoxins, cardiotoxins and three-finger toxins, derived from venoms of Elapid snakes. In doing so, the system achieved a Q3 score of 70%, which is quite remarkable.