The expression profiling approach – maximising the value of your microarray data

Dr Giles Yeo

Summer School 11-14 July 2006
The Møller Centre, Storeys Way, Cambridge, UK

The emergence of a new molecular tool in 1995, the DNA microarray, has led to a revolution in the way scientists can now study gene expression and regulation. Microarray technology allows us to assess the expression of large numbers of genes simultaneously, allowing us to correlate physiology with gene expression. All of the major array manufacturers now offer ‘Whole Genome Arrays’, where up to 65,000 different transcripts are represented on individual chips, giving us a wealth of gene expression data for each individual sample. However, as the collective wisdom regarding the use of microarray technology increases, it is clear that the unprecedented amount of data produced from each individual experiment is as much a burden as it is useful. How does one minimize the noise inherent in any microarray experiment, thus allowing the valuable nuggets of information to be sifted out?

After a brief overview of current technology, I will discuss the three key steps needed to maximise the value of your microarray data:

i.) Experimental design – e.g. how many replicates is enough? Which are the most appropriate time-points to use? Intelligent experimental design that balances required statistical robustness with invariable financial constraints is arguably the most critical step in obtaining useful microarray data.
ii.) Data analysis – a step overlooked by very many users of this technology! I will discuss basic ‘filtering’ strategies, e.g. using fold-change or confidence limits, as well as more unbiased ‘pathway enrichment’ analyses.
iii.) Orthogonal validation – I have now got my list of genes…..what next? How should I consider validating the array data? How do I determine if the changes in gene expression are actually physiologically significant?

In addition, I will outline uses of microarray technology that go beyond gene expression profiling. For example, array technology can now be used for whole genome typing of single nucleotide polymorphisms (SNPs), homozygosity mapping of consanguineous pedigrees, comparative genome hybridization (CGH), as well as a readout for chromatin immunoprecipitation (so called ChIP on chip).
Although I will not be able to make all of you Microarray experts in 40 minutes, I hope to make you aware of the limitations of this technology, but yet impress upon you its exciting possibilities if utilized correctly.

The opinions expressed in this paper are those of the speaker and do not necessarily reflect the views of the Society

Revised: 24-Aug-2006