Title:Phylotastic: An Experiment in Creating, Manipulating, and Evolving Phylogenetic Biology Workflows Using Logic Programming

Abstract:Evolutionary Biologists have long struggled with the challenge of developing analysis workflows in a flexible manner, thus facilitating the reuse of phylogenetic knowledge. An evolutionary biology workflow can be viewed as a plan which composes web services that can retrieve, manipulate, and produce phylogenetic trees. The Phylotastic project was launched two years ago as a collaboration between evolutionary biologists and computer scientists, with the goal of developing an open architecture to facilitate the creation of such analysis workflows. While composition of web services is a problem that has been extensively explored in the literature, including within the logic programming domain, the incarnation of the problem in Phylotastic provides a number of additional challenges. Along with the need to integrate preferences and formal ontologies in the description of the desired workflow, evolutionary biologists tend to construct workflows in an incremental manner, by successively refining the workflow, by indicating desired changes (e.g., exclusion of certain services, modifications of the desired output). This leads to the need of successive iterations of incremental replanning, to develop a new workflow that integrates the requested changes while minimizing the changes to the original workflow. This paper illustrates how Phylotastic has addressed the challenges of creating and refining phylogenetic analysis workflows using logic programming technology and how such solutions have been used within the general framework of the Phylotastic project. Under consideration in Theory and Practice of Logic Programming (TPLP).