Democratising high performance computing for bioinformatics through serverless cloud computing: A case study on CRISPR-Cas9 guide RNA design with Crackling Cloud

Organisations are challenged when meeting the computational requirements of large-scale bioinformatics analyses using their own resources. Cloud computing has democratised large-scale resources, and to reduce the barriers of working with large-scale compute, leading cloud vendors offer serverless computing, a low-maintenance and low-cost model that provides ample resources for highly scalable software applications. While serverless computing has broad use, its adoption in bioinformatics remains poor. Here, we demonstrate the most extensive use of high-performance serverless computing for bioinformatics by applying the available technologies to CRISPR-Cas9 guide RNA (gRNA) design. Our adaptation of the established gRNA design tool, named Crackling, implements a novel, cloud-native and serverless-based, high-performance computing environment using technologies made available by Amazon Web Services (AWS). The architecture, compatible with technologies from all leading cloud vendors, and the AWS implementation, contributes to an effort of reducing the barrier to large computational capacity in bioinformatics and for CRISPR-Cas9 gRNA design. Crackling Cloud can be deployed to any AWS account, and is freely available on GitHub under the BSD 3-clause license: https://github.com/bmds-lab/Crackling-AWSAuthor summary: Cloud computing platforms have changed how researchers access and use computing resources, yet its full potential in bioinformatics remains largely unrealised. To help overcome the barriers researchers face when needing high-performance computing, we propose using publicly available cloud platforms to run bioinformatics tools. We focus on serverless systems that operate only when needed, scale automatically, and reduce costs. This is an ideal combination for researchers who run large-scale analyses infrequently. In this work, we developed Crackling Cloud as a serverless system for gene editing research. It includes both a reusable cloud architecture and a working implementation that can be deployed to any Amazon Web Services account, allowing researchers to run experiments without specialised hardware or technical expertise. We tested Crackling Cloud on datasets of varying size and showed that it performs well as the complexity of the task increases. The system is open-source, making it easy to adopt and share. We believe this approach can help democratise access to bioinformatics tools and promote broader use of cloud technologies in scientific research.