Microarrays are now an established tool for biological research and have a wide range of applications. In this thesis I investigate the BeadArray microarray technology developed by Illumina. The design of this technology is unique and gives rise to many computational and statistical challenges. However, I show how knowledge from other microarray technologies can be used to our advantage. I describe the beadarray software package, which is now used by researchers around the world. The development of this software was motivated by the fact that Illumina's software (BeadStudio) gives a summarised view of Illumina data and does not gives users any control over several processing steps that were found to be crucial for other microarray technologies. A main feature of beadarray is the ability to access raw data. The advantages of such data include the ability to perform more detailed quality assessment and greater control over the analysis at all stages. The analysis of a control experiment shows that the processing steps used in BeadStudio can be improved. In particular, utilising variances calculated from the raw data can increase the ability to detect genes which have di erent expression levels between samples, a common goal for microarray studies. The data from the control experiment are made available for other researchers to use and validate their own analysis methods. One issue discovered during the analysis of the control experiment was that only half of the intended genes could be reliably measured due to problems in the design of the probes targetting particular genes. By considering a large set of publicly available Illumina arrays, I show how such unreliable measurements can a ect the analysis of Illumina data. I also show how potential problems can be identi ed in advance of an experiment and incorporated into an analysis pipeline.