Evolving Turing machines for biosequence recognition and analysis

This article presents a genetic programming system for biosequence recognition and analysis. In our model, a population of Turing machines evolves the capability of biosequence recognition using genetic algorithms. We use HIV biosequences as the working example. Experimental results indicate that evolved Turing machines are capable of recognizing HIV biosequences in a collection of training sets. In addition, we demostrate that the evolved Turing machines can be used to approximate the multiple sequence alignment problem.