Alinhamento múltiplo de proteí­nas via algoritmo genético baseado em tipos abstratos de dados.

Abstract

This works presents a new model to the multiple protein alignment problem using genetic algorithm based on abstract data types - GAADT. This model uses a structure called chromosome, which is a set of gene that in turn is composed of basic units called bases. Each chromosome is a possible alignment of the input sequences and the chromosome fitness is calculated according with the bases order in alignment. This model is different from other paradigms of multiple alignment by aligning the input sequences as a whole instead of a progressive pairwise alignment approach. Some characteristics of this model concern to the data structure, well-defined genetic operations and the convergence to a solution close to that found in other tools. The validation was performed comparing reference alignments of protein families subset with the results of the model.