London Calling 2023: Ultra-rapid nanopore sequencing: computational pipeline 2.0

In 2021, the ultra-rapid nanopore sequencing pipeline set the world record for the fastest diagnosis using DNA sequencing in 7 hours 28 minutes. With further improvements to basecalling and nanopore chemistry, we further improved the pipeline in terms of accuracy, speed, and cost. While leveraging the improvement in basecalling accuracy with newer versions of Guppy, the pipeline now consists of autoscaling of the resource in the high-performance cloud computing infrastructure using Ray. Autoscaling of resources based on the rate of sequencing data generation and fine-grained control over resource-scheduling improves cost efficiency while maintaining the speed of basecalling and alignment. Additionally, a suite of tools associate compute performance across batches of sequencing data. The pipeline can now call small variants, structural variants, and copy number variants within 30 minutes of alignment. The variant curation stage is now automated using the GEM software from Fabric Genomics. As a result, we can generate a  curated list of variants within two hours of completing sequencing, for almost 180 GB of data.