Pulling information out of a genome has proved to be a challenging task, which requires complex statistical tools and powerful computers to run the analysis. And if results are to be delivered in a reasonable timeframe, you’d better ask for fast computers as well.
This sort of computing power is not available to all researchers interested in how animals inherit their physical traits. To counter this problem, Dr Sara Knott from the University of Edinburgh and her team developed GridQTL, a grid-based platform that provides fast and robust analysis to identify trait-related genome regions.
“You don’t need to understand the grid, or know about programming to use GridQTL.”
These are called quantitative trait loci (QTL). Knott explains: “QTL are regions of the genome that have an effect on a given physical trait.” A knowledge of the QTL involved in the expression of a trait is crucial for our understanding of variation between individuals and how traits are passed on from generation to generation, she adds.
But the interest in the study of QTL goes far beyond simple academic curiosity. “Knowledge of QTL has many applications in the pharmaceutical industry and in risk management as well,” says Knott. Understanding QTL helps to predict the risk related to diseases with an underlying genetics basis.
QTL are also important to the way we interact with domesticated plants and animals. Careful genome analysis helps breeders to select the right parents for a next generation with improved physical traits. With insights from QTL, it is possible to breed better animals and plants through husbandry practices alone.
The technique has been used to study cattle, sheep, obesity in pigs, or disease resistance in poultry. Recent applications have moved to other fields, from farmed fish to Australian crocodiles.
GridQTL builds on the experience of QTLExpress, a software developed by Knott’s team at Edinburgh as a tool to analyse datasets. QTLExpress proved to be a success and increasing usage meant that the software required more and reliable computing power.
The platform is used worldwide by about 300 active users and at least 30 papers have been published with data analysed with this software, since January 2009. GridQTL allows users to input their data into an intuitive web-based graphical interface implemented with the GridSphere portal project. The portal hosts JSR 168 compliant Java portlets especially designed for job submission, querying, file management, data manipulation and other services. The computing jobs are submitted to the UK’s National Grid Service (NGS) with the Globus Toolkit.
The analysis could take 30 hours to complete on single core computers. Grid computing means that you can have the job done in 12 minutes.
“It would be very difficult to run this kind of software without grid computing,” says Knott. “We have an uncertain user base and the grid provides the flexibility to sustain an adequate speed of analysis, regardless of online users,” she adds. “It would also be very difficult to maintain the hardware ourselves.”
Knott is confident that GridQTL is not an headache for the user. “You don’t need to understand the grid, or know about programming to use GridQTL,” she says. Scientists simply add their data and the software provides the statistical analysis ready to be interpreted.
Veterinary scientists from the University of Sydney used GridQTL to identify the first reptile QTL on a saltwater crocodile. The species is farmed extensively in Australia for its highly-prized leather used in luxury goods. The researchers found the QTL responsible for the number of scale rows, an important quality trait. Mapping and understanding this QTL will help crocodile farms to design better husbandry programmes and breed crocodiles with improved skin quality. (Miles et al 2009; doi:10.1111/j.1365-2052.2009.01978.x)
A team of British and Swiss scientists described for the first time the QTL involved in heaves, a hereditary respiratory disease that cripples horses around the world. The disease causes permanent wheezing and coughing and is thought to be an allergic reaction to moulds in hay and straw. With GridQTL, the scientists identified the genes most likely to control the horse’s immune response to these substances and hope to shed light on this problematic condition. (Swinburne et al 2009; doi:10.1007/s00335-009-9214-5)
A group of European researchers used GridQTL to detect QTL in sea bass, a North Atlantic fish commonly bred in aquaculture facilities. The team found two genome areas linked with the fish’s body weight. The QTL explain only up to 38% of the weight variation of the studied sea bass population, but it’s the first step towards identifying the genes involved in economic traits. (Massault et al 2009; doi:10.1111/j.1365-2052.2009.02010.x)
Salmonella-infected poultry is a serious health hazard, but giving prevention antibiotics to chicken is not a wholesome silver bullet. A solution for this problem is to breed chickens with improved natural resistance to nasty bacteria. Following this lead, a group of scientists from Britain used GridQTL and an innovative statistical approach to detect four areas in the chicken genome involved in susceptibility to Salmonella colonisation. The team hopes that the results will inform future breeding strategies designed to control the spread of disease in poultry. (Fife et al 2009; doi: 10.1111/j.1365-2052.2010.02090.x)
Animal and plant physical traits are controlled by DNA. Eye colour and skin freckles, for example, are regulated by single genes and follow the relatively simple dominant/ recessive pattern of Mendel's laws of heredity. But traits measured over a continuous range, such as height or propensity to a disease, depend more on areas of the genome called quantitative trait loci, or QTL.