Analysis of tissue-specific co-expression networks

Understanding context-specific behavior of biological systems plays a fundamental role toward analyzing molecular mechanisms contributing to genetic diseases and the development of efficient therapeutic strategies [1]. The majority of genes show significant differences in expression levels in different tissues that reflect variations in biological pathways characteristic for a particular tissue or cell types. Thus, the reconstruction and analysis of the gene co-expression networks provides an approach to understanding tissue-specific mechanisms in health and disease. In this study, we reconstructed and analytically assessed tissue-specific co-expression networks in order to identify and characterize molecular pathways specific for a particular tissue (e.g. heart, brain, kidneys) or shared between a number of tissues. The reconstructed tissue-specific networks will be further used for the development of network-based disease models and subsequent predictions of causative genetic factors. To accomplish this goal we have reconstructed gene networks applying a well-established method for modeling the gene expression correlations as a multivariate Gaussian distribution with an L2 norm penalty. n 2 log det S − tr SΣ − λ S !