Miten Jain from the University of California, Santa Cruz presented at the Nanopore Community Meeting 2021 on “Nanopore sequencing updates using Q20+ and R10.4.” While this may not be the latest technology, I think it is useful to learn about the history and how improvement were made. Jain spoke about the early version of ultra-long sequencing in 2020. They shared data from thirty human genomes with three flow cells per genome and an average read length N50 of 77 kb. The SQK-ULK 001 kit was explained and how a single library preparation can be used to prepare enough library for multiple flow cells. Also, Jain noted that nuclease “flushes” every day help improve throughput. With this kit, the N50 remained at 77 kb, but now the coverage per flow cells was around 30X per flow cell. Jain mentioned that the R9.4 chemistry and software improvements improved the accuracy. The R10.4 Q20 chemistry further improved accuracy to a median of 99%. This median accuracy is estimated for simplex reads; duplex reads would further improve accuracy. Jain shared data for a workflow for SNV calling using the Q20 R10.4 reagents. Jain summarized human WGS output using long, ultralong, and Q20 R10.4 reagents. The summary presented emphasized how chemistries are improving increasing accuracy and making complex assemblies more efficient. I wonder if R10.4.1 and subsequent versions will further improve both read accuracy and throughput.
