I continued watching a longer session from Oxford Nanopore Technologies (ONT) summer programming tonight. “The era of complete genomes, at any scale” was the title of a session Cora Vaher gave. They talked about how methylation can be detected without additional library preparation. Haplotype-specific methylation could be very useful in learning about genetic diseases. Vaher shared a slide with the current lineup of ONT devices, noting that they can all be placed on a benchtop. Vaher ended by explaining how ONT sequencing makes a wealth of information more accessible. They then introduced Greg Elgar from Genomics England, who spoke about WGS on the PromethION. Elgar established the Nanopore workflows with the Genomics England program. The program has three benchtop PromethION devices! The Genomics England program was set up in 2012. Human genomes from people were sequenced. Elgar said that despite some issues and troubleshooting, going with ONT was “absolutely the way to go.” The workflow has three PromethION 48 instruments and uses a 10 Gbs link to a cluster with five NVIDIA DGX A100 processors for base calling. They have now implemented an Amazon Web Services system and perform most processes on the cloud. Elgar highlighted how updates in Guppy and chemistry improved the quality of the sequencing, resulting in increases in the number of genomes. For the program, one flow cell produces 30X+ human genome coverage. Consistent depth quality is an issue related to sample quality and flow cell variability. Elgar shared that some cancer samples have been particularly challenging. The Cancer 2.0 program has its own lead and aims to take nanopore sequencing into the clinical space. The objective is to improve diagnosis with an end-to-end pipeline with nanopore long-read data. Elgar explained that because they are a government agency, they need to constantly review procedures and generate evidence that the system in place is efficient and reproducible. The team has been performing variant analysis. The bioinformatics team has reviewed and benchmarked the R10 performance of Severus SV calling. CNV calling has been more complicated than anticipated, and Elgar explained that they are collaborating with several groups. HATCHet2, so far, has been working. Elgar explained that the group focuses on rare diseases and has sequenced ~1,500 genomes with long-read sequencing!
