Microarray technology makes it possible to simultaneously measure the expression of thousands of genes. Gene expression data can be analysed in many different ways to produce putative knowledge on for example co-regulated genes, differentially expressed genes and how genes interact with each other. One way to derive gene interactions is to use rule induction algorithms such as association rule discovery algorithms or decision trees. The application of such algorithms to gene expression data sets typically generates a large set of rules serving as hypotheses of how genes interact. It is necessary to apply different measures to assess the interestingness of the rule hypotheses. There are well known domain independent objective measures, but there is a lack of domain specific interestingness measures tailored for microarray gene expression data. Without domain specific interestingness measures it is impossible to know if the hypotheses are interesting from a biological perspective, without resorting to time consuming manual evaluation of every single rule. The aim and contribution of this work is to develop a method for assessing the interestingness of rules induced from microarray gene expression data using a combination of objective and domain specific measures.