MOA - A Lightweight Bioinformatics Workflow Framework aimed at Reproducible Research

Rapid expansion in complexity and volume of biological data analysis has led to heightened demand for reproducible research. Web-based systems such Galaxy and Gene pattern provide powerful tools to non-specialist users but to date no framework is aimed at supporting bioinformaticians working at the Unix command line in both development and production settings. To address this gap we have developed MOA (https://github.com/mfiers/Moa) , a lightweight framework which supports reproducible command-line bioinformatics, and provides simple web-based access to results by end-users. MOA implements commonly held principals of reproducible research and ‘best practice’ by organizing jobs as directories, and pipelines as directory trees. In addition to templates for common genomics tools, a key feature is provision of templates for ad hoc code to support flexible and reproducible use of shell or any other tools in pipelines. Since MOA is built on generic components it may be readily adapted and extended, and can provide an effective complement to support prototyping and data pre-processing in web-based systems like Galaxy. We will introduce MOA command-line and web interfaces and demonstrate simple usage in NGS data analysis.