Utilizing Genes Functional Classification in Microarray Data Analysis: a Hybrid Clustering Approach

We present an integrated clinico-genomics environment. The proposed reference architecture provides for the seamless integration of clinical and genomic information, and aims towards the future genetic-medicine environment. Intelligent processing operations (i.e., data mining) are in the heart of this environment. In this context, we also present a novel graph-theoretic hybrid clustering approach that utilizes information about the functional classification of genes in order to achieve more knowledgeable, and by though, more naturally interpretable clustering arrangement of the genes. The clustering approach was tested on an indicative real-world datasets with satisfactory and interpretable results.