Measuring and analyzing tissue specificity of human genes and protein complexes

It is essential for human systems biology and medicine to understand the tissue specificity of expressed genes and their products, which are involved in important cellular processes and diseases. Over the last years, many studies were based on the freely available Novartis Gene Atlas data to investigate the tissue specificity of human gene expression and its biological impact on protein expression and protein interaction networks [1,2]. The Gene Atlas data consists of comprehensive gene expression datasets for a wide variety of tissues and cell lines [3]. However, these data were already published in 2004, and the microarrays employed to obtain the data were of low probe density and specifically designed to measure genes that were assumed to exist at that time. This raises the question whether these relatively old datasets should still be regarded as reliable source for tissue specificity of human genes. A more recently developed microarray is the Affymetrix Exon Tiling Array, which has been developed to measure exon expression rather than gene expression [4]. Its probe density per gene is much larger than the microarray technology used to generate the Gene Atlas. Furthermore, the advent of next-generation sequencing machines allows further technological advances in accurate transcriptome measurements [5].In the following, we explore three tissue-dependant gene expression datasets produced by microarray technologies and high-throughput sequencing of RNA. We first study the detection sensitivities of the technologies and compare the measured gene expression datasets. Furthermore, we investigate protein interactions to identify tissue-specific and housekeeping interactions. Last, we utilize expression data for the detection and comparison of tissue-specific protein complexes and analyze to what extent functional implications on tissue specificity depend on the applied expression detection method.All analyses are based on the Ensembl database, version 52 (genome build