Journals Proceedings

Abstract

Recently development of automated and integrated pipelines for phylogeny and evolutionary analyses has received a lot of attention. In this paper, we describe our approach of pipeline development. Our design focuses on integration, not only in the integrated result (i.e., the pipelines), but also in the integrated approach in the design process, so that we can exploit all kinds of resources available to us. In addition, we want integrated pipelines not only work effectively, but also efficiently. In this paper, we describe our two-step design: A basic approach and an optimized approach which takes advantage of parallelism offered by supercomputers. We present the major phases of pipeline development, describe methods used in optimization (including the sequence-split-logic), show the experimental result, and compare it with non-optimized implementation.