August 09th, 2019
Topic: BioCompute Objects as a Framework for Uniting Advances in Workflow Portability and Reproducibility
Presenter: Dr.Dennis A. Dean, II, PhD, Seven Bridges
Abstract:
The Cancer Genomics Cloud (CGC) and Seven Bridges utilize Common Workflow Language (CWL) to enable the portability of tools and workflows. We propose integrating CWL workflow development on the CGC with the FDA’s efforts to develop the BioCompute Objects as a mechanism for unifying efforts towards workflow portability and efficient communication of complex workflows. Unifying these efforts could have implications across the Cancer Research Data Commons and across local infrastructure. Whereas the CWL emphasizes maintaining analysis details that support execution, the BioCompute Object defines a framework for collecting the information describing appropriate use and limitations of complex NGS analysis. Unifying CWL and BioCompute Objects could provide a powerful framework to facilitate immediate reuse of complex NGS analysis tools that could take months to implement independently. In this talk, we will discuss our work integrating BCO generation through an R Shiny app with two supporting libraries (TidyCWL and BcoTool) on the Cancer Genomic Cloud. We will demonstrate that strong BCO support for CWL provides an opportunity to link other workflow tools (ex. Rabix Composer and Rabix Benten) in addition to advancing the FDA and CGC visions for standardization.