Basecalling – How Good is Good Enough?

With the release of the R10 flow cells in 2019 and the Dorado basecaller in 2023, Oxford Nanopore Technology® long-read sequencing has displayed a significant improvement in accuracy. Since then, there have been studies investigating the basecalling runtime, output read quality, methylation detection accuracy, and nanopore-only assembly quality. However, these studies have been limited to small quantities of organisms and have not explored the general use cases that may apply to many laboratories. How much better performance does dorado offer over guppy at the various accuracy modes in relation to the time spent basecalling for as many diverse organism types as possible? While at the time of writing this abstract, the analysis is ongoing, the study described here will evaluate 283 different Oxford Nanopore Technologies® GridION® sequencing samples across bacterial, fungal, and viral species in terms of basecalling speed and average read quality for both guppy and dorado with each using the fast, hac, and sup accuracy modes. Further, both nanopore-only and Illumina®/ONT® hybrid assemblies from the resultant 6 collections of data will be evaluated for completion with respect to taxonomic domain, GC content, assembly N50, genome size, replicon count, small variants, and gene content. To date, this represents the first time such a broad and multi-method study has been performed; further underscored that all sequencing was performed by the same user, instrument, chemistry, flow cell, laboratory, and physical basecalling hardware. This study aims to guide sequencing laboratories on the effects of the options available.