geneXplainR: An R interface for the geneXplain platform

title: "geneXplainR: An R interface for the geneXplain platform" tags: - "R" - "Bioinformatics" - "Systems biology" - "geneXplain" - "Platform" - "Upstream analysis" authors: - name: "Philip Stegmaier" orcid: 0000-0002-1881-8746 affiliation: 1 - name: "Alexander Kel" affiliation: 1 - name: "Edgar Wingender" affiliation: 1,2 affiliations: - name: "geneXplain GmbH, Am Exer 10b, D-38304 Wolfenbuettel, Germany" index: 1 - name: "Institute of Bioinformatics, University Medical Center Göttingen, D-37077 Göttingen, Germany" index: 2 date: "12 July 2017" bibliography: paper.bib

The geneXplain platform [@genexplain] is an online toolbox and workflow management system for a broad range of bioinformatic and systems biology applications. The platform is well-known for its upstream analysis [@upstreamAnalysis], that has been developed to identify causal signalling molecules on the basis of experimental data like expression measurements. Methods integrated into the system include

molecular network analysis such as pathway enrichment, identification of network clusters, common signaling regulators or effectors,
analysis of transcription factor binding sites like prediction of binding sites using positional weight matrices, testing for enrichment of binding sites in regulatory sequences, or identification of composite modules (combinations of binding sites), as well as motif finding
methods to test for enrichment of functional groups or categories, e.g. from the Gene Ontology [@GO], using the Fisher test or Gene Set Enrichment Analysis (GSEA) [@GSEA]
Flux Balance Analysis [@FBA] to analyze metabolic networks
methods for processing and statistical analysis of high-throughput data, e.g. Limma [@limma] or DESeq2 [@deseq]
as well as simulation of computational models, e.g. as collected in the BioModels database [@biomodels].

An important feature of the platform is the possibility to define and execute workflows that can implement sequential and parallel multi-step analysis processes. Workflows can be created and edited using a graphical editor. They are an effective tool to define complex analysis pipelines and to document, reuse and to reproduce analysis procedures. Figure 1 shows the graphical user interface of the platform with an example workflow for Flux Balance Analysis.

A workflow in the geneXplain platform Figure 1. Graphical user interface of the geneXplain platform showing the Flux Balance Analysis workflow.

We have developed geneXplainR, an R [@R] interface for the geneXplain platform, that makes it possible to define analysis pipelines in the R language using tools, workflows and other resources integrated in the platform. The package is based on and extends the rbiouml package [@rbiouml]. The geneXplainR adds basic functionality not covered by rbiouml such as creation of projects, folders or deletion of items from the workspace as well as functions that provide direct access to certain tools or workflows. Another purpose of geneXplainR is to offer a suite of examples scripts in the example branch that help users to get started with the software. We have also developed a similar project denoted as the genexplain-api [@genexJava] that addresses Java developers and shall be described elsewhere. With geneXplainR, developers can easily take advantage of other bioinformatics software and resources available in R, e.g. through the popular Bioconductor project [@bioconductor].

The development of geneXplainR has been supported by the MyPathSem, a collaborative project funded by the German Federal Ministry of Education and Research (BMBF) in the funding program “i:DSem – Integrative Datensemantik in der Systemmedizin”, as well as by MIMOmics, a collaborative project funded by the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 305280, research area FP7-HEALTH-2012-INNOVATION-1, topic HEALTH.2012.2.1.1-3: Statistical methods for collection and analysis of –omics data.

Issue reports and support requests are welcome either by writing an e-mail to info@genexplain.com or through the GitHub issue system (https://github.com/genexplain/geneXplainR/issues).