Our findings call for empirically derived, run-specific error models in sequencing simulation. GemSIM meets this need with a set of python scripts that can be run on a standard desktop computer. By allowing analysis of run-specific error models created from the user's own data, GemSIM helps researchers to identify unique features of their data - understanding of which may be invaluable for downstream data analysis. Error-model analysis also facilitates quality control and identification of 'bad quality' sequencing runs, such as the Illumina v4 run described here. Finally, simulation of sequencing data based on empirically-derived error models allows researchers to choose the most appropriate sequencing platform for their project, assess the impacts of errors on downstream data analysis, and objectively interpret any findings.
