Researchers at TOBB University, Turkey, recently revealed that social network analysis (SNA) can be successfully employed in the identification of disease biomarkers, namely cancer biomarkers such as lymphoma. The study is entitled “Employing social network analysis for disease biomarker detection” and was published in the International Journal of Data Mining and Bioinformatics.
The detection of disease biomarkers, especially cancer biomarkers, is an important task in clinical research and diagnostics that has drawn significant attention in the field of in silico genomic experiments: experiments based on computer simulations.
The popularization of online social networks has allowed the rapid development of resources that can provide a better understanding of the connections between members of the network, including their interactions, activity, hubs and nodes. These resources can be applied in relationships between different entities, from Facebook users, to bees in a colony or genes and respective proteins in the body.
In the study, researchers describe a new strategy based on SNA techniques for the identification and improved understanding of cancer biomarkers in the body based on patient genomic microarray data. Similar to social interactions, the authors believe that one can envision genes as actors or members in a social network, where similarities between them can be seen as connections.
Since the human genome comprises around 20,000 genes, genomic databases derived from it can be massive; in this sense, this approach can dramatically reduce the number of features that need to be analyzed to identify a useful disease biomarker. Once these biomarkers have been identified, they can be evaluated in screening programs in individuals who are at risk for a certain disease.
The team showed a proof of principle for their approach with three types of cancer: lymphoma, leukemia and colon cancer, where the results of their experiments with the selected biomarkers were found to be promising. “We showed how our approach is capable of effectively detecting cancer biomarkers out of high-dimensional genomic data,” reported the research team according to a news release. “We combined clustering and classification into the developed framework to help in detecting the links between the various genes within the model and to validate the outcome, respectively.”
The team’s next goal is to optimize their approach and exploit it in terms of possible interactions between protein-gene, protein-protein, disease-protein and disease-drug, with the aim of improving diagnostics and design personalized therapies for the individual patients based on the results of their personal biological network analysis.