Workflows with model selection: A multilocus approach to phylogenetic analysis

Abstract

The workflow model of description and execution of complex tasks can be of great use to design and parallelize scientific experiments, though it remains a scarcely studied area in its application to phylogenetic analysis. In order to remedy this situation, we study and identify sources of parallel tasks in the main reconstruction stages as well as in other indispensable problems on which it depends: model selection and sequence alignment. Finally, we present a general-purpose implementation for use in cluster environments and examine the performance of our method through application to very large sets of whole mitochondrial genomes, by which problems of biological interest can be solved with new-found efficiency and accuracy. © 2011 Springer-Verlag Berlin Heidelberg.