Composition of bioinformatics model federations using communication aspects

摘要

Scientists in bioinformatics research utilize multiple software tools and models for their analysis of genomic data. This involves a mixture of stand-alone and custom software. Stand-alone application tasks include computational data analysis and data visualization. The genomic data links the software together; the output from one may be suitable as input for another. Scientists often create custom tools and scripts to perform additional computational and processing tasks, such as filtering, before passing data from one tool to the next. This interoperating collection of tools and models becomes a "federation" of collaborative software. We propose an approach using aspect-oriented programming and Linda-style tuple space coordination language to facilitate connecting the individual components of the federation together; we assemble the separate components into the larger complex system, ideally without altering the original components. This paper describes the concepts behind our approach of using AOP and tuple spaces to intercept, filter and transform data as well as manage tool execution and coordination.