Evaluation of experimental design and computational parameter choices affecting analyses of ChIP-seq and RNA-seq data...

One of the great advantages of next generation sequencing is the ability to generate large genomic datasets for virtually all species, including non-model organisms. It should be possible, in turn, to apply advanced computational approaches to these datasets to develop models of biological processes. In a practical sense, working with non-model organisms presents unique challenges. In this paper we discuss some of these challenges for ChIP-seq and RNA-seq experiments using the undomesticated tree species of the genus Populus.

We describe specific challenges associated with experimental design in Populus, including selection of optimal genotypes for different technical approaches and development of antibodies against Populus transcription factors. Execution of the experimental design included the generation and analysis of Chromatin immunoprecipitation-sequencing (ChIP-seq) data for RNA polymerase II and transcription factors involved in wood formation. We discuss criteria for analyzing the resulting datasets, determination of appropriate control sequencing libraries, evaluation of sequencing coverage needs, and optimization of parameters. We also describe the evaluation of ChIP-seq data from Populus, and discuss the comparison between ChIP-seq and RNA-seq data and biological interpretations of these comparisons.

These and other "lessons learned" highlight the challenges but also the potential insights to be gained from extending next generation sequencing-supported network analyses to undomesticated non-model species.