As with any new sequencing technology, exploring the nuances of the data lays an essential foundation for developing platform-specific bioinformatics methods. The goal of this study was to provide a comprehensive evaluation of the types of errors and biases introduced during PGM library preparation and sequencing. Previous studies which used the manufacturers aligner, ‘tmap’, on default settings implicitly removed reads with two or more indels prior to data analysis [2], [3]. As these filtered reads can only be distinguished with the benefit of a reference genome, such studies are more relevant for re-mapping applications. By using an indel-tolerant mapping approach, we were able to map almost 100% of the reads, which largely explains why the global base error-rates calculated in this study are higher than previously reported.
