Guiding the rudder of bio-IT research

Researchers at the Walter and Eliza Hall Institute of Medical Research (WEHI) and the University of Melbourne are working on a project they hope will result in clear experimental guidelines and methods for mathematical analysis for microarrays.

Dr Gordon Smyth, of the WEHI Genetics and Bioinformatics Division, has received an NHMRC grant for the project, with collaborators Dr Hamish Scott, also of WEHI and Dr Robert Di Nicolantonio and Dr Leanne Delbridge, of the University's Department of Physiology.

"The purpose of the grant is really methodological," explaines Smyth. "We'll be running experiments to test methods on the arrays."

The researchers plan to investigate a number of methods with the aim of improving the analysis of the arrays. Among these are studies to look at whether RNA samples from animal studies should be pooled prior to creating arrays, or whether better data is achieved by keeping them separate.

In another series of tests, the use of SAGE (serial analysis of gene expression) will be used to calibrate the microarray results to get a more accurate quantitative result. Techniques for assessing both the quality of the arrays and the results will also be developed.

"The emphasis is on low-level analysis for microarrays. This is the first level of analysis that has to be done by everyone," says Smyth.

Smyth says that the ultimate aim of the project was to develop a suite of methodology that would help other microarray users, as part of the bioinformatics research ongoing in the Division under the direction of Prof Terry Speed.

But in addition, the project will also produce results of biological interest for the researchers involved, says Smyth. Scott is working on Down Syndrome and leukaemia, while Di Nicolantonio and Delbridge are working on hypertension and cardio-renal physiology.

Smyth says that the cross-disciplinary approach to developing improved methodology for microarray experiments and analysis was unusual as it combined his knowledge of statistics with the biology.

"A lot of microarray development is done by statisticians with no biological knowledge, or by biologists with no statistics. This is one of the first cross disciplinary approaches," he says.