The importance of genotype identity, genetic heterogeneity, and bioinformatic handling for properly assessing genomic variation in transgenic plants

Background The advent of –omics technologies has enabled the resolution of fine molecular differences among individuals within a species. DNA sequence variations, such as single nucleotide polymorphisms or small deletions, can be tabulated for many kinds of genotype comparisons. However, experimental designs and analytical approaches are replete with ways to overestimate the level of variation present within a given sample. Analytical pipelines that do not apply proper thresholds nor assess reproducibility among samples are susceptible to calling false-positive variants. Furthermore, issues with sample genotype identity or failing to account for heterogeneity in reference genotypes may lead to misinterpretations of standing variants as polymorphisms derived de novo. Results A recent publication that featured the analysis of RNA-sequencing data in three transgenic soybean event series appeared to overestimate the number of sequence variants identified in plants that were exposed to a tissue culture based transformation process. We reanalyzed these data with a stringent set of criteria and demonstrate three different factors that lead to variant overestimation, including issues related to the genetic identity of the background genotype, unaccounted genetic heterogeneity in the reference genome, and insufficient bioinformatics filtering. Conclusions This study serves as a cautionary tale to users of genomic and transcriptomic data that wish to assess the molecular variation attributable to tissue culture and transformation processes. Moreover, accounting for the factors that lead to sequence variant overestimation is equally applicable to samples derived from other germplasm sources, including chemical or irradiation mutagenesis and genome engineering (e.g., CRISPR) processes.