AIMES deployment of HPC (High performance Computing) Cluster
AIMES are extremely proud to announce that we have recently deployed a new High Performance Computing (HPC) cluster to further enhance the capability provided to Trustworthy Research Environment (TRE’s) clients. Following the deployment of Graphical Processing Unit (GPU) capability earlier this year to support the ingesting and processing of healthcare imaging data, the HPC capability will support research using genomic and proteomic data.
Modern healthcare research increasingly involves combining insights from clinical, imaging and ‘omics data sets; and these are often large volumes of critical data. We are therefore seeing an increasing demand to create secure environments capable of supporting advanced processing and analytical needs.
AIMES CEO, Dr Dennis Kehoe, highlighted to need for TRE’s to provide high-performance processing capability “The ability to undertake ground-breaking medical research requires secure data environments in which the data and the analysis can be collocated. By continually investing in processing capability we are able to support the world-class UK health data research”.
Who will benefit from the HPC cluster?
A number of projects in both genomics and proteomics are earmarked to use the new facility and to demonstrate the benefits to researchers in having easy access to HPC capability within a TRE. Senior Business Development Manager at AIMES, Antony Shimmin, identified the use cases for which advanced processing is required; “Adding HPC capacity to our already successful Trustworthy Research Environment solution will enable much faster processing times for genomic, proteomic and other research datasets. With the use of parallel computing we can provide our customers with the ability to vastly reduce processing time to in some cases within minutes, when it would typically take weeks by human process. We are working with the University of Liverpool who are developing AI/ML techniques to enable enhanced precision in the measurement of paediatric brain tumours”
Heather Glover, AIMES Senior Technical Architect, commented: “AIMES are providing a significant uplift in the compute capacity within the Data Provisioning Zone of the AIMES TRE, with and addition of hundreds of compute cores, terabytes of RAM and hundreds of terabytes of raw storage, and this is just the beginning. We are building automated services that will allow researchers to transfer data sets securely in and out, as well as incremental write-outs and clear-down facilities. We will be providing professional services as required, to help customers who are not conversant with HPC coding, to help covert their code into something that can be processed on the HPC stack.”
To develop the tooling and processes needed by researchers to access the HPC, Dr Andrew Collins from the University of Liverpool is working with AIMES to on-board the initial use cases. Dr Collins explains, ‘One principal challenge in my field of research (Proteomics and Mass Spectrometry) is the volumes of data we are faced with. To put it into context, to identify a single protein variant requires the analysis of thousands of mass spectra. To scale this up across not just a single tissue, but a whole organism such as the human proteome (~75,000 known proteoforms), can potentially result in hundreds of millions of spectra to identify. The HPC platform will allow us to carry out large scale studies in a much more efficient manner, reducing the workload and time significantly and ultimately improving the results and outcome of our research.’