Model organisms have been essential for improving our knowledge about human diseases. The genomes of rodents and humans are highly similar but similarity alone is no guarantee that results obtained from an animal model can be successfully transferred to the human situation. At the molecular level cross-species differences manifest themselves in many ways which can prevent rodent models from serving as effective means to study complex human (patho)logical processes. Hence, it is crucial to better understand similarities and differences between rodents and humans by systematically comparing them at the molecular level. In order to validate the similarity of model organisms to their human counterpart, a promising approach is the identification of functionally conserved and diverged sub-networks of genes or proteins. Taking advantage of the gene expression data that has been accumulated over the past 15 years, gene-gene co-expression networks of human and a whole range of model organisms can be compared for various conditions to elucidate conserved and diverged gene co-expression sub-networks. To compare co-expression networks between human and mouse we used a network alignment algorithm (NATALIE) for their global alignment. In vertebrates, the liver plays a vital role in wide range of functions, specifically substrate metabolism, detoxification of chemicals and metabolization of drugs. Impairment in any of the molecular functions of the liver can result in a varied range of disorders such as hepatitis, cirrhosis, and liver tumours. The liver is also associated with metabolic syndrome, a well-known multi-factorial disease encompassing obesity and diabetes. We compared gene expression in healthy liver from human and mouse by aligning their gene-gene co-expression networks. We identified conserved sub-networks (modules), and investigated their differences (Nandal, in prep). In agreement with previous findings, we show that the conserved sub-networks between mouse and human derived using our approach have overall similar biological functions, but that, within these modules some genes do show poorly conserved co-expression profiles suggesting functional divergence. In a second application we use network alignment to compare human and mouse B lymphocytes (Nandal, in prep).
Tamoxifen resistance in breast cancer
The EpiPredict project aims to elucidate the epigenetic regulation of endocrine therapy resistance in breast cancer. We aim to develop novel methods for the analysis of multi-omics data (e.g., gene expression, ChIP-seq) to increase our understanding of Tamoxifen resistance and to new predict targets for treatment.