Finding Signal Peptides in Human Protein Sequences Using Recurrent Neural Networks

摘要

A new approach called Sigfind for the prediction of signal peptides in human protein sequences is introduced. The method is based on the bidirectional recurrent neural network architecture. The modifications to this architecture and a better learning algorithm result in a very accurate identification of signal peptides (99.5% correct in fivefold cross-validation). The Sigfind system is available on the WWW for predictions (http://www.stepc.gr/synaptic/sigfind.html).