Högskolan i Skövde

Cancer is a broad term for a wide spectrum of diseases and which involves the alteration in expression levels of several hundreds of genes. As such, the study of the disease from a systems biology point of view becomes rational, as the properties of a system as a whole may be very different from the properties of its individual components. However, understanding a network at the systems level not only requires knowledge about the components of the network, but also the interactions between them.

Here, a systems biology view of the rat PHD finger protein 5A (Phf5a) gene was attempted; a gene previously identified as aberrantly expressed in estrogen dependent endometrial adenocarcinoma tumors from both rat and human. Phf5 ais a highly conserved cysteine rich (C4HC3) zinc finger and such proteins predominantly have a role in chromatin mediated transcriptional regulation. Moreover, PHF5A is a component of the macromolecular complex spliceosome that takes part in pre-mRNA splicing and spliceosome component coding genes have previously been shown to be implicated in various cancer types and suggested to potentially be novel antitumor drugs.

To derive a systems biology view, in this study, a weighted gene network was inferred from a list of genes having correlated expression profiles to Phf5a as nodes, and common transcription factors and microRNAs regulating these genes together with annotation about biological process ontology term(s) and pathway(s) as edge weights. In the inferred network a higher weight indicates more annotation shared between two genes and, hence, the network facilitates the identification of closely interacting genes with Phf5a. The results show that highly weighted edges connect Phf5a to other spliceosome components, but also to genes involved in the metabolism of proteins, proteasome and DNA replication, repair and recombination. The results also link Phf5a to the Myc/Rb/E2F pathway, one of the central pathways associated with cancer. The proposed method for inferring a weighted gene network can easily be applied to other genes and diseases.