An integrative methodology based on protein-protein interaction networks for identification and functional annotation of disease-relevant genes applied to channelopathies

Figures

• 

  Fig. 1. Semi-automatic workflow for the identification and functional annotation of the most relevant genes in a pathology. Stage 1) Phenopedia [30] is a disease-centered view of genetic association studies summarized in the online Human Genome Epidemiology (HuGE) encyclopedia. It provides a list of genes involved in the disease of interest. It is in this stage that a complex disease, a set of diseases or a certain disease can be chosen to be studied. Stage 2) STRING [22] is a database of known and predicted protein-protein interactions that allows the discovery of relationships across disease genotype and thus the creation of the PPI network. Cytoscape [23] is an open source software platform for the visualization and analysis of complex networks that measures each gene and identifies network nodes. Stage 3) DAVID database [24] works as a semi-automatic functional annotation tool of the genes obtained after Stage 2

• 

  Fig. 2. Protein-protein interaction (PPI) network of channelopathies analysed in Stage 2. Each network node represents the protein produced by each single, protein-coding gene locus from the gene dataset of channelopathies. The representation is a circular layout based on the betweenness attribute with undirected edges (other layouts shown in Additional file 10). The node size marks the level of degree and therefore neighbourhood (the larger nodes represent proteins with a higher number of interactions). The node colour shows the level of betweenness and therefore the level of centrality (the warmer the colour of the protein, the shorter path between the two which indicates how important the node is within the wider context of the entire network) [31]. HLA proteins are discarded due to their disconnection from the principal component. The image was generated by Cytoscape [23]

• 

  Fig. 3. Nine genes as the most relevant in Stage 2 using different centrality statistics. Venn diagram representing the intersections calculated through the use of other statistics for the proposed centrality measures such as closeness, Eigenvector and radiality. The same set of genes identified with degree and betweenness only still turned out to be the most relevant in channelopathies: SCN1A, SCN2A, SCN4A, SCN4B, SCN5A, SCN9A, KCNH2, KCNQ2 and ANK3. Venn diagram was obtained using a free available tool provided by Ghent University [32]. All the statistics information is included in Additional file 2

• 

  Fig. 4. Systematic review procedure for the genes of interest through OMIM, Gene and MedGen databases

• 

  Fig. 5. Genotype-phenotype association network for cardiovascular diseases category. Association networks created for cardiovascular diseases category from the evidences obtained by a) last stage of the workflow, b) systematic review and c) exhaustive review. The “MeSH”-based categories include each pathophysiological evidence shared by two or more genes. Other heart diseases include heart arrest, cardiomyopathies, myocardial ischemia or cardiomegaly

• 

  Fig. 6. Genotype-phenotype association network for nervous system diseases category. Association networks created for nervous system diseases category from the evidences obtained by a) last stage of the workflow, b) systematic review and c) exhaustive review. The “MeSH”-based categories comprise each pathophysiological evidence shared by two or more genes

• 

  Fig. 7. Genotype-phenotype association network for mental and other diseases categories. Association networks created for mental and other diseases categories from the evidences obtained by a) last stage of the workflow, b) systematic review and c) exhaustive review. The “MeSH”-based categories comprise each pathophysiological evidence shared by two or more genes. Other mental disorders include bipolar disorder, Alzheimer disease, autism, depression or schizophrenia