You can鈥檛 just send 46 million gigabytes of data over the internet
REMEMBER when you could only fit 10 songs on a cassette tape? Or store a couple of books on a floppy disc? It is getting harder to remember a time when you couldn鈥檛 fit whole TV series on to a gadget smaller than a matchbox.
That鈥檚 just as well. It is becoming increasingly quick, cheap and easy for scientists to collect vast amounts of data, whether it鈥檚 about people鈥檚 shopping habits, internet usage, or even their genetic make-up. The challenge now is to store, share and make sense of that information. A new breed of cross-disciplinary researcher is meeting that challenge, and answering some of the toughest questions in science.
Advertisement
Around the world, geneticists and molecular biologists are busy sequencing proteins and genomes. Many of their results are sent to the , based in Hinxton, UK, which stores the information in huge databases, readily accessible to all.
Over the past five years or so, much of the genetic data sent to the EBI was on viruses, bacteria and animals. 鈥淣ow, the majority of sequencing is actually devoted to human beings,鈥� says , director of the EBI.
Take the , for example. This international collaboration will make the full genomes of 2500 anonymous individuals freely available for research. In December 2012, the prime minister David Cameron announced plans for an even bigger set 鈥� the , which should be completed by 2017. The project鈥檚 goals include using the genetic data to help diagnose and treat illnesses such as cancer within the next four years.
The EBI has the capacity to store 46 petabytes of information. That鈥檚 46 million gigabytes, or enough to fill about 10 million standard DVDs. 鈥淭he size of the sequence data is becoming a challenge, technically, in terms of transferring it,鈥� says Thornton. 鈥淵ou can鈥檛 just shift this data over the internet.鈥�
What鈥檚 more, the databases are growing rapidly. 鈥淟ast year, our volume of data doubled,鈥� says Thornton.
Share and share alike
, an intergovernmental organisation covering 17 European countries, is trying to find a way for scientists around the world to add to and share all that data. 鈥淎t the moment, the EBI is the centrepiece, where all the big archives in Europe are kept,鈥� says , director of Elixir.
鈥淲hen biology data becomes big, it is not produced at a single site, like at the Large Hadron Collider,鈥� Blomberg says. 鈥淚t鈥檚 not just coming from one large hole beneath Geneva 鈥� it is coming from every university, and there are thousands of them in Europe. So we need to build an infrastructure that scales with that diversity and distribution.鈥�
Blomberg and his colleagues use two methods. The first involves renting out a fibre-optic cable that usually takes internet traffic between countries. 鈥淲e can rent part of the network for a specific task,鈥� Blomberg says. That might be sending a particularly large number of genome sequences from the UK to Finland, for example.
The other involves creating temporary 鈥渄igital embassies鈥� at the EBI. If a research group wants to compare some secure information to the reference data at the EBI, it can do so by essentially creating a secure data cloud within the EBI. This allows researchers to access smaller amounts of information in a secure way, without having to download data.
This kind of collaboration is helping to answer all kinds of questions. Olivo Miotto and his colleagues at the Centre for Genomics and Global Health are collecting and mining another set of genetic data for clues about the spread of drug-resistant malaria. 鈥淲e have brought together a network of collaborations to create which causes malaria,鈥� he says.
The centre, funded by the UK Medical Research Council, is based in Oxford, and partners with MalariaGEN, a network of clinical researchers working all over the world. Miotto, who is in Bangkok, Thailand, is particularly concerned about the spread of drug-resistant malaria across South-East Asia. 鈥淲e鈥檝e seen a rise in resistance to artemisinin, the front-line antimalarial therapy worldwide,鈥� he says. 鈥淭here is a major, justified concern that artemisinin-resistant malaria will spread from South-East Asia to Africa, which would cost many lives.鈥�
In an attempt to track the spread of resistance, Miotto鈥檚 team collaborate with others in the region. Colleagues in health clinics take blood from people with malaria, isolate the parasites in the blood, and send the samples to Miotto鈥檚 team. The team then sequence the parasite genomes and compares them to records at the Wellcome Trust Sanger Institute in Hinxton, UK. 鈥�,鈥� Miotto says.
Once his team share this information with doctors in countries affected by malaria, they look at how patients are responding to treatments. They can compare the genomes of parasites found in people whose treatment was successful with those taken from people who aren鈥檛 responding as well. This helps to determine which genes in the parasite might be conferring resistance. The team hope to be able to use this information to drive health policy. For example, if they spot a drug-resistant pathogen in a region previously unaffected by the problem, they could alert public health authorities, who might be able to intervene to prevent its spread.
The approach, known as genetic epidemiology, has only taken off in the last few years, Miotto says. 鈥淚t鈥檚 very much an emerging field, and is being applied to a variety of diseases,鈥� he says. 鈥淚t is an exciting area that is here to stay.鈥�
Of course, big data isn鈥檛 just about banks of computers. 鈥淭he people who work at the EBI are the jewels of the institution,鈥� says Thornton. The work attracts a certain kind of person, says Miotto, a breed of researchers who are very multidisciplinary and willing to focus on the broader impacts of their work in biology, computing and statistics.
Miotto came to genetic epidemiology from a career in software engineering, while Blomberg initially studied biochemistry. 鈥淚 had to sit down with a thick mathematics book, because there wasn鈥檛 enough maths in my undergraduate degree,鈥� Blomberg recalls. Miotto鈥檚 team write some of the software used to analyse the data. But biology underpins what they do. 鈥淎 lot of this is very specific to malaria parasites,鈥� says Miotto. 鈥淚t requires a lot of knowledge about malaria as well as statistics.鈥�
This blend of expertise is vital for the EBI and Elixir, too. 鈥淲e鈥檝e got people coming from physical sciences, computer sciences and life sciences,鈥� says Blomberg.
The scientists making sense of big data can play an important role in global health, says Miotto. 鈥淏ig data gets condensed into smaller sized, but higher value, data, until it gets to a point where it is a decision,鈥� he says. This might mean changing the dose of a drug, for example. 鈥淚f you鈥檙e working in diseases that affect poor countries, there鈥檚 no way that the local government health authority will have the skills or resources to analyse terabytes of data,鈥� he says. 鈥淵ou have to work on this translation to make it actionable 鈥� condensing terabytes into a single sentence.鈥�
Blomberg says it is very rewarding work. 鈥淭his is great fun. It is where science is happening right now 鈥� you can really start to understand the genetic drivers behind both rare and complex diseases in a way that was unthinkable four or five years ago.鈥�
鈥淵ou can start to understand diseases in a way that was unthinkable five years ago鈥�
This article appeared in print under the headline 鈥淭oo much information鈥�
