Microarray Data Mining: Issues and Prospects

Introduction

With the developing of new technologies and revolutionary changes in biomedicine and biotechnologies, there was an explosive growth of biological data during the last few years. Genome wide expression analysis with DNA microarray technology has become a fundamental tool in genomic research. Since microarray technology was introduced, scientists started to develop informatics tools for the analysis and information extraction from this kind of data. Due to the characteristics of microarray data (i.e. high levels of noise, high cardinality of genes, small samples size) data mining approaches has become the suitable tools to perform any kind of analysis on these data. Many techniques can be applied to analyze microarray data, which can be grouped in four categories: Classification, Feature Selection, Clustering and Association Rule Mining.

Classification is a procedure used to predict group membership for data instances. Given a training set of samples with a specific number of attributes (or features) and a class label (e.g., a phenotype characteristic), a model of classes is created. Then, the model is exploited to assign the appropriate class label to new data. Model quality is assessed by means of the classification accuracy measure, i.e., the number of correct label predictions over the total number of unlabeled data. The classification of microarray data can be useful to predict the outcome of some diseases or discover the genetic behavior of tumors.

Since genetic data are redundant and noisy, and some of them do not contain useful information for the problem, it is not suitable to apply the classification directly to the whole dataset. Feature Selection techniques are dimensional reduction methods usually applied before classification in order to reduce the number of considered features, by identifying and removing the redundant and useless ones. Moreover, feature selection algorithms applied to microarray data allow identifying genes which are highly correlated with the outcome of diseases. Another way to identify redundant genes is to group together sets of genes which show a similar behavior, and then select only a representative for the group. Furthermore, genes with similar expression pattern under various conditions or time course may imply co-regulations or relations in functional pathways, thus providing a way to understand functions of genes for which information has not been previously available.

Finally, relationships among genes or annotations and sample conditions can be detected also by exploiting the association rule mining techniques, which extract correlations among dataset attributes. This technique is also used to analyze time-series microarray data to discover gene regulatory networks.

In this review, application of data mining techniques on microarray data is focused, with the aim of making researchers aware of the benefits of such techniques when analyzing microarray data. The chapter is organized as follows. The first two sections provide a description of microarray data, to highlight the issues concerned with their analysis, and a brief discussion about the data cleaning approaches that can be exploited to prepare data before data mining. The following four sections provide a survey of classification, feature selection, clustering and association rule mining techniques based on their aims and characteristics. Finally, the last two sections describe new trends and provide some prospects of data mining application to microarray data.