Feature
Scientist analysing data
Creative Commons attribution information
©Eugenio Marongiu via Getty Images
All rights reserved

From AI to vaccines – how the cloud enables research

As research focuses more and more on data analysis, how is the impact of cloud computing affecting the landscape? 

The world of research is complex and includes a multitude of subjects and approaches. But one thing that is beginning to unite them all is data. 

“Paraphrasing James Hetherington, director of e-infrastructure at UKRI, there is no science without data science,”

says Andy Powell, cloud CTO at Jisc. 

“And with the increasing amounts of data that researchers – both in technical fields and beyond – need to process, cloud computing is becoming rapidly more important.” 

Service implications 

Cloud technology enables the use of relatively new services such as artificial intelligence (AI), machine learning, text analysis and image recognition. 

There are an increasing number of ready-made platforms available through cloud providers, so that less in-house coding is needed. This is particularly advantageous for researchers who may not have much coding experience and perhaps work outside of technical fields, and who therefore may not have access to the tools required to build their own platforms.

Microsoft, Google and Amazon Web Services (AWS) have done a lot of work in developing platform-as-a-service (PaaS) and software-as-a-service (SaaS) tools, to reduce the workload for the individual.  Cloud technologies have also created the ability for ‘on-demand computing’.

“Cloud is affecting how research is done,”

says Andy.

“For example, instead of waiting for time on an institution’s high-performance computing (HPC) facility, a researcher can spin up compute nodes in the cloud, using a provider such as Amazon Web Services or Microsoft’s Azure, and work from there, shutting the nodes down once they are no longer needed.

This provides a greater flexibility in work schedules, allows for more rapid experimentation and can improve efficiency.” 

Open research 

Organisations outside of Big Tech have been working to develop cloud resources for research, too. 

The European Open Science Cloud (EOSC) is a project first proposed in 2016 as part of the European Cloud Initiative to build a ‘competitive data and knowledge economy in Europe’. It offers 1.7 million European researchers and 70 million professionals across science, technology, the humanities and social sciences an open service for storage, data management, analysis and reuse of research data.

The EOSC works across borders and disciplines by federating existing data infrastructures dispersed across the EU.  

The EOSC is also currently running a Galaxy portal for coronavirus (COVID-19) research. The portal contains data on coronavirae, including samples from SARS-CoV-2 (COVID-19), which is updated daily from public international databanks.

The cloud also includes some key tools for identification of mutations, phylogenetic analysis, sample processing and visualisation, meaning researchers can run experiments without having to install tools. The EOSC setup also means that tools for sharing data, workflows and results are easily accessible without downloads. The EOSC-Synergy collaboration supports data processing. 

The EOSC’s executive board co-chair, Cathrin Stöver, says:

“EOSC is an ideal tool to respond to public emergencies such as the COVID-19 virus by,
1: opening up scientific data on the virus,
2: sharing live on-the-ground data on the spread of the virus, and
3: accessing and combining this data not only to eradicate the virus but learn how to deal with similar viruses and future outbreaks.” 

Life-saving vaccines 

Cloud technologies are also being utilised in the development of a COVID-19 vaccine. 

Imre Berger, professor of biochemistry at the University of Bristol, and his team has developed a new class of synthetic vaccines, called ADDomer. This consists of a protein which is engineered to form a large particle structure resembling a virus. In order to be able to create this protein, the team needed to know the structure of the ADDomer at near atomic resolution. This is where the technology played a central role.

The team were able to determine the structure of the protein through cryo-electron microscopy and then use cloud computing to compute the structure from the microscope data.  

Using cloud also allows an increased level of collaboration in projects such as these, says Adam Finn, professor of paediatrics at Bristol Children’s Vaccine Centre

“That's what's exciting about this; distinct people coming together to solve a problem, not actually only for vaccines but for medicine in general and actually work out how we can help each other to understand the best way of doing things.”