Mantas Sereika, a Ph.D. student at Aalborg University in Denmark, spoke at the Nanopore Community Meeting 2021 about “Nanopore R10.4 enables near-perfect bacterial genomes.” This topic is of interest for several of the courses I teach. Sereika spoke about the improvements in accuracy of Nanopore sequencing. Raw read accuracy has increased but there are still challenges with homopolymers. Improvements described by Sereika included the R10.4 Q20 LSK112 kit. The modal raw read accuracy reported is round 99%… though homopolymer basecalling is still a challenge. Sereika sequenced the Zymo mock community and identified improvements and areas of indels. Polishing helped correct consensus accuracy. Sereika tested the performance of R10.4 and Q20 library prep by taking anaerobic digester sludge and sequencing on ONT, PacBio, and Illumina. This complex metagenomic sample was sequenced with these instruments/chemistries and allowed for comparison of calling of metagenome assembled genomes (MAGs). Longer homopolymers are rare, and R10.4 helped resolve them. MAG truncations issues could only be resolved with polishing with short reads with R9.4, but could be polished with Medaka and long reads with R10.4. Sereika concluded that the “R10.4 Q20 improves raw read accuracy and homopolymer calling” and “R10.4 Q20 can be used to acquire near-perfect genomes or MAGs (coverage >40x).” Sereika explained that duplex reads represent about 1% of the reads yet have a modal accuracy of 99%+. Now with 10.4.1 flow cells and new kits and chemistry replaced LSK112, I wonder how polishing has improved without the need for short reads?
