CRC Wound: the embedded bioinformatician
A wound is a complex ecosystem, seething with a diverse population of the patient’s cells, and a complex community of microbes contending to colonise the wound. There is constant communication between them via a molecular intranet.
The Wound Management Innovation Cooperative Research Centre in Brisbane was established in 2010 to explore these processes, and to apply its findings to develop novel treatments to accelerate healing, reduce scarring, and prevent infection.
Every patient is genetically unique, and the microbes vying to colonise their wound will vary according to how and where the wound occurred – it might have been in an operating theatre, a car accident, a farm accident, or the consequence of necrosis associated with chronic diabetes or a lifetime smoking habit.
Over time, the wound’s microbial flora may vary with time, through succession processes. Fibroblasts and other specialised tissue-repair cells go about their business as the innate and adaptive immune systems’ armies of natural killer cells, macrophages, B cells and T-cells swarm in to mount a coordinated defence of the breach against microbial invaders. The high-speed DNA sequencers and various microarray technologies required to study the dynamics of these processes over time, generate enormous amounts of data.
The Co-Program Leader of the Enabling Technologies Program in the Wound CRC, Professor Zee Upton, now Assistant Dean of Science at Queensland University of Technology, says it was clear from the outset that the CRC was going to need high-order capabilities in bioinformatics to make sense of the huge volumes of DNA, SNP and proteomics data flowing from its research.
In previous projects the bioinformaticians were not physically co-located with the biological scientist generating the data. “That didn’t work as well as we hoped,” Professor Upton said. “We found we needed a bioinformatician with specific expertise in handling wound-healing data, so QFAB placed one of its researchers in the CRC for two days a week.
“Since then, QFAB has assigned a succession of its research students and staff to work with our biologists and protein chemists, helping to design experiments and to interpret the data. As we were doing our research QFAB was building up a pool of expertise in wound-healing bioinformatics.”
And that’s the model QFAB is offering to potential customers for its expertise and facilities. QFAB Chief Executive Officer, Jeremy Barker says the success of biological, medical and pharmaceutical research will depend increasingly on the expertise of bioinformaticians who not only understand the research, but can help biologists to design experiments and analyse the data to answer the questions posed.
As the cost of purchasing and operating high-speed DNA sequencers has fallen, more research groups have been successful in obtaining ARC infrastructure grants to buy these machines.”Some set out to use the new equipment without really understanding how it can be used,” he said.
“They may have bought the machine after reading a paper in the scientific literature that used the same device. It might describe the methods used to produce the data, and the conclusions, but reveals very little about the assumptions underlying the methodology, which tend to be project-specific.”
“So they bring us this huge data set, saying they want do to this type of analysis, hoping to get a certain result, without fully understanding what they’ve got from the experiment, or how to extract what they want from the dataset. “That’s where we can help. First, we can explain what their data will allow them do, or how it might allow them to do more than they expected. “Secondly, we can help by showing them how to do what they really wanted to do, and help them design experiments to produce the type of data they need to answer particular questions. Early engagement saves the costs involved in repeating experiments.
Barker says QFAB’s expertise in experimental design and bioinformatics analysis can be applied at any scale, from a simple microarray comparison of differential gene expression, to a full-blown systems biology analysis integrating genomic, proteomic, metabolomic and clinical data. “We engage with a prospective client, discuss what they want, then submit it to our team for a think-tank session on how it might be done with the expertise and resources that we have in-house, and at the client’s end. “We then go back to the client with a proposal detailing where and how we believe we can help.”
Barker says QFAB has served a variety of clients over the past seven years including CSIRO, Australia’s universities, medical research institutes, state primary industry, fisheries and forestry departments, and private biotech and pharmaceutical companies. Barker says bioinformatics is moving “at a mile a minute”, so keeping abreast of the changes requires a commitment to ongoing learning as part of the job and an investment in the professional development of QFAB’s team. Barker says he has seen an example of a research institute’s embedded bioinformatician being rapidly overwhelmed by data. “They have no reserve time to learn on the job. That is the advantage of having a larger team to draw on.
“Around 60 per cent of QFAB’s recruits are PhDs. Everyone else is at least postgraduate trained, and we have a core of experienced bioinformaticians in senior positions to bring the necessary rigour to project planning.”
“You need the depth of experience provided by staff who have worked on other projects before QFAB recruited them.”
“A bioinformatician joining the team may have had experience working on a particular class of cancers, or on viral infections of mangoes, so they’re good with that type of data and have that specialist knowledge as well as the more general bioinformatics skills. “Collectively we can call upon a broad range of expertise within our teams, and ask them what they think about a particular question.”
“We apply the team approach to every project, but we like to broaden the discussion on the projects we have on the boil at any point in time, because someone from another team may have an insight into a particularly tricky question.”
Barker says clients can be certain that QFAB will have the best available expertise to do the job – “We stay up to date with the technologies out there, and the specialised methodologies that go with them, so we will have a pretty good idea which technology is most appropriate for each client’s project.”
“The technology changes rapidly, and all bioinformaticians think they can produce a better algorithm than the one described in the latest journal.”
“We learn and adapt as need dictates. You don’t always have to reinvent the wheel; often, there is already an algorithm that can do the job. The trick is knowing that the algorithm exists, and how to apply it.” As for data-crunching power, Barker says the facility has just updated its computer cluster. It has seven nodes built around a 64-core processor, each with 256 gigabits of RAM, and terabyte storage capacities.
“That’s enough for most jobs,” he said. “But if it isn’t, we have access to a supercomputer facility at the University of Queensland that runs a terabyte of RAM, which can handle whole genome comparisons.”
“With these facilities, our broad expertise, and our data-integration abilities, we believe we offer a service beyond the in-house bioinformatics capabilities of any single institution or company in Australia and the Asia Pacific region.”