Cloud-SPAN: Specialised analyses for environmental 'omics with Cloud-based High Performance Computing

Environmental Biotechnology (EB) addresses global challenges using engineered microbial systems for environmental protection, bio-remediation and resource recovery. It is a critical and expanding area for the UK and underpins some of the world's most important industries. This is acknowledged by the funding invested in the creation of BBSRC's Networks in Industrial Biotechnology and Bioenergy (NIBBs).
A deep mechanistic understanding of the complex microbial communities involved in the biological cycling of global resources is essential to meet global challenges such as Net Zero, waste management and increased demand. The complexity of these microbiomes can be orders of magnitude larger than those found in the human gut, requiring different approaches to experimental design and analysis with High Performance Computing (HPC). However, EB is an interdisciplinary field that attracts researchers from a broad range of disciplines including Mathematics, Engineering, Biology, Social Sciences, Management, Physics and Chemistry and big data 'omics analyses on HPC systems are often not core skills for such researchers.
This proposal aims to develop and deliver highly accessible resources that will upskill these interdisciplinary researchers so that they are able to generate and analyze big data relating to EB using Cloud HPC. Although infrastructure and resources exist for microbial 'omics (e.g. JGI's IMG/M, MG-RAST, Galaxy, CLIMB, EBI) there is a lack of systematic training tightly linked to the EB domain and documentation is often focussed on technical proficiency rather than contextualised with a strong understanding of experimental design. We will provide foundational training and develop and deliver new advanced modules covering the specialised skills required to generate and analyse 'omics data using Cloud HPC resources. These will include experimental design and statistical modules to ensure researchers can generate data appropriate to investigate their research question. Modules will deploy cloud-based containerised instances provided by Google Education and Amazon Web Services (AWS) for exemplar workflows free to the learner. They will form a complete training resource with fully articulated prerequisites and learning objectives that can be used for in-person or online tutor-led workshops or self-paced learning. Our proposal offers structured Learning Paths from the statistical skills required for robust experimental design through to the reproducible execution and interpretation of 'omics analyses with HPC to cater to researchers with differing levels of previous experience and which allow self-assessment of training needs. We also provide Diversity Scholarships to enable members of underrepresented groups to participate in online or in-person training.
The collaboration between the University of York and the Software Sustainability Institute (SSI) brings together excellence in data science pedagogy and environmental 'omics research with the SSI's UK-leading expertise in research computing and community building. This will ensure the training developed genuinely complements, and aligns with, existing materials to enhance national provision.
Sustainability will be fostered by making the resources Findable, Accessible, Interoperable and Reusable (FAIR), providing cross-platform images for deployment and by developing and proactively engaging with a Community of Practice and providing Code Retreats for the supported practice of methods to participants' own data. In addition, "Cloud Administration Guides" will be developed for institutional HPC Teams to run specialised modules with their own resources. These Guides will be supported 1-to-2 day training by Cloud-SPAN systems administrators.
The project will be promoted by our partners, the SSI, Google Education, AWSand the N8 Centre of Excellence in Computationally Intensive Research and through delivery of conference talks and seminars.

Environmental Biotechnology (EB) is an interdisciplinary field involving advanced molecular and applied microbiologists, environmental chemists and engineers that addresses global challenges using engineered microbial systems. This proposal aims to trainer these researchers to generate, analyse and mine big data relating to EB microbiomes which are larger than those found in the human gut and require different approaches to both measurement and analysis in order to manage reagent costs and effectively leverage available HPC resources.
Easily accessible and scalable HPC-based training is required to provide researchers with the skills and self-confidence to manipulate and analyse big data generated from 'omics technologies and generate biological insights from these highly interconnected systems. This area of bioinformatics involves a steep learning curve which can be confounded by the need to install packages with multiple dependencies onto different HPC architectures based on what is available at a researcher's home institution, even before the user can engage with writing scripts to manage workflows, manipulate or visualise data, or manage job schedulers. We will deploy Cloud-based containerised instances which are (1) accessible to anyone anywhere as long as they have an adequate internet connection, (2) have a very low hardware entry requirement and (3) allow for easily scalable and replicable installations of software that will not become deprecated as quickly as might occur on a local server. The cloud providers we are working with run grant schemes that provide significant resources to researchers that will support deployment of production instances of the images we will generate. We expect that our resources will be easily deployed and used by groups who do not necessarily have devoted bioinformaticians or expertise in HPC, providing a cost-effective route to useful analyses for researchers in a strategically key area of expansion in the biosciences.