Discovery of Genetic Interactions Across Diverse Oryza sativa Sequencing Data Sets: a Workflow-Based Approach

The interrogation of genome, transcriptome, and epigenome information is key to understanding genetic regulatory networks in crop species.   While applications such as RNA-seq, ChIP-seq, and bisulfite sequencing can yield valuable information about such networks, it can be challenging for a scientist, or even a biostatistician, to bring all of this information together in a meaningful way.  Many-to-one relationships across data types and variability in how data are distributed can make interpretation of the data difficult.  Here we present a workflow-based approach for analyzing and integrating data across these diverse data types.  Our approach has been designed not only to overcome these challenges, but also give researchers easier access to their data.  By “mapping” data types to a common context, diverse sequence data, as well as raw data from other platforms such as microarray and quantitative PCR, can be efficiently integrated into an analysis.  To demonstrate this, we created a workflow to analyze a previously published Oryza sativa data set containing mRNA, siRNA, and bisulfite sequencing reads (Chodavarapu et al, 2012).   This information was used to build a correlation network of genetic interactions, which is informed by information across all of the data types.  Despite broad variation between the individual strains, an integrated analysis utilizing all three data types revealed important differences between the groups.  The method and results presented here highlight the power of integrated analysis across data types, and the importance of algorithms that can efficiently place large volumes of diverse data types into a shared context.