Test your microbiome analysis pipeline with PLuMA

Study of the microbiome and metagenomics are actively developing areas of research. In the meantime, analyzing microbiome data has become increasingly complex, with multiple tools required to be run sequentially. Running these so-called “pipelines” of tools can be challenging because of the diversity of coding languages and compatibility issues.

 

To overcome this problem, Trevor Cickovski and Giri Narasimhan from the Bioinformatics Research Group of the Florida International University have developed PLuMA, a Plugin-Based Microbiome Analysis lightweight back end pipeline that supports multiple dynamically loaded plugin extensions. Here, they describe their tool and its main features.

 

Plugin-Based Microbiome Analysis

 

If you are an algorithm developer who wants to prototype, test and debug a new pipeline stage in your programming language of choice and run alongside existing stages with no overhead for interfacing, the Bioinformatics Research Group (BioRG) at Florida International University would like to introduce you to Plugin-Based Microbiome Analysis (PluMA).

 

Our goal with PluMA is that developers devote as much energy as possible to their new algorithm with minimal energy devoted elsewhere. We have therefore made the package lightweight, with an executable program that is under 200K. This program is then essentially “infinitely extensible”, loading pipeline stages on-the-fly at runtime as plugins, which can be downloaded from the PluMA plugin pool at http://biorg.cis.fiu.edu/pluma/plugins.

 

Users only need to install the plugins they need for their pipeline, keeping the disk and runtime memory footprints minimal.

 

PLuMA Omictools microbiome pipeline metagenomics
Conceptual design of PLuMA.

 

PluMA supports plugins in C++, CUDA, R, Python and Perl and a developer may construct their algorithm in any of these supported languages. They can also run their new plugin alongside other plugins from the pool without any knowledge of their underlying languages or implementation, through a simple interface that specifies the plugin name, an input file and an output file.  Some plugins in this pool also wrap existing tools (i.e. Mothur). We provide a Plugin Generator (PluGen) for this purpose.

 

 

The plugin pool currently contains more than sixty plugins and is growing. We encourage users to send links to their completed plugins to the BioRG so that we can incorporate them into the pool for the external community. We hope that the PluMA plugin pool ultimately becomes a portal from which community members can freely contribute and download plugins, which they can then seamlessly integrate into their respective pipelines.

 

PluMA is available at: http://biorg.cis.fiu.edu/pluma. For any questions, please contact lead developer Trevor Cickovski at [email protected].

Reference

 

Trevor Cickovski and Giri Narasimhan. (2018). Constructing Lightweight And Flexible Pipelines Using Plugin-Based Microbiome Analysis (PluMA). Bioinformatics