New supercomputer for life science

Data analysis
Industrial biotechnology and pharmaceutical research will both benefit from a new supercomputer to be located at a new supercomputer centre at Risø Campus.

A new supercomputer currently being established at Risø Campus will be among the 100 largest in the world. It will form part of the European research infrastructure ELIXIR, helping researchers to store and analyse the rapidly growing volumes of biological data.

Every six months, the volume of digital data on living organisms doubles. In everything from bacteria and yeast fungus to humans, the smallest components of living organisms—cells, genes and proteins—are described, their laboratory living conditions are recorded and hospital patient record data, including lifestyle knowledge, are gathered.

All these data on millions of species stored on powerful supercomputers can teach us a great deal—e.g. how diseases are linked and coupled to particular genes or how to change bacteria so they act as the perfect biotechnological cell factory.

Generally speaking, the huge data volumes are extremely heterogeneous and—in the case of personal data or data intended for commercial use—extremely sensitive, so the computer must be optimized to deal with this diversity and data transfer should therefore be kept to a minimum. Hence the emergence of large, so-called high-performance computing centres (HPC) connected to networks such as the European ELIXIR, which Denmark joined this spring. Such networks make it possible to utilize each other’s data to analyse new and larger contexts.

Professor Søren Brunak, DTU Systems Biology, has acted as Chairman of the Board of Directors during ELIXIR’s initial phase and is also responsible for the acquisition of the HPC-centre currently being installed at Risø Campus.

Potent and eco-friendly
Among the 100 largest in the world, the new supercomputer can store 7.5 petabytes of data and is optimized to process the rapidly growing volumes of data. A petabyte is 1015 bytes, and the sum total of all US research library data is 2 petabytes, which gives some idea of the supercomputer’s capacity.

Over the summer, a space has been cleared in the northernmost corner of Risø Campus behind the former DMU Building. Here, 3,500 m2 were set aside for new infrastructure such as this life science supercomputer, which will occupy 800 m2 and consume a megawatt of power.

The computer will arrive at the beginning of October in two prefabricated containers equipped with the latest chip-cooling technology.
“The CPUs are cooled using direct water technology. The water is then channelled into the district heating system via heat pumps so that the heat can be utilized. Even with the first server modules, Risø Campus will be almost fully heated using surplus heat for four months of the year, and things will only improve as the site expands,” says Senior Executive Officer Peter Løngreen, DTU Systems Biology, who is coordinating the Danish bioinformatics group’s work across all Danish universities.

National facility
The HPC server will, of course, be an invaluable tool for everyone involved with life science data at DTU, but other DTU research groups will also have the opportunity to book calculation sessions on the supercomputer. Similarly, all major life science research groups in Denmark will be among the users, as will several of DTU Systems Biology’s international partners, including Beijing Genomics Institute.

Article from DTUavisen No. 7, September 2014.