Jonathan Smith, Francis Crick Institute
At the Francis Crick Institute we have recently begun investigating the feasibility of using AWS for on-demand scale out of genomics workloads. To explore this, we converted a portion of an existing genomics pipeline for execution in the cloud. We were specifically looking to refactor existing pipeline sections as opposed to tailored cloud pipelines and strived to ensure any changes to the pipeline were effective locally as well as in the cloud. Using Nextflow we managed to leverage AWS’ elastic compute service to provide on-demand computation that automatically scales as needed. Throughout this exercise we documented any difficulties observed and barriers to entry encountered, as well as comparing the cost to our in-house cluster. This presentation will document our findings and touch on our emerging future strategy for the role of cloud in research computing at the Crick.