Computational analysis of human protein interaction networks

Abstract

Large amounts of human protein interaction data have been produced by experiments and prediction methods. However, the experimental coverage of the human interactome is still low in contrast to predicted data. To gain insight into the value of publicly available human protein network data, we compared predicted datasets, high-throughput results from yeast two-hybrid screens, and literature-curated protein-protein interactions. This evaluation is not only important for further methodological improvements, but also for increasing the confidence in functional hypotheses derived from predictions. Therefore, we assessed the quality and the potential bias of the different datasets using functional similarity based on the Gene Ontology, structural iPfam domain-domain interactions, likelihood ratios, and topological network parameters. This analysis revealed major differences between predicted datasets, but some of them also scored at least as high as the experimental ones regarding multiple quality measures. Therefore, since only small pair wise overlap between most datasets is observed, they may be combined to enlarge the available human interactome data. For this purpose, we additionally studied the influence of protein length on data quality and the number of disease proteins covered by each dataset. We could further demonstrate that protein interactions predicted by more than one method achieve an elevated reliability.

Computing topological parameters of biological networks

“…In recent years, high-throughput experiments have produced large networks of interacting molecules, which are represented as nodes linked by edges in complex graphs (Albrecht et al, 2005;Ramı´rez et al, 2007;Zhu et al, 2007). In this context, the characterization of biological networks by means of graphtopological properties has become very popular for gaining insight into the global network structure (Albert, 2005;Almaas, 2007;Barabasi and Oltvai, 2004;Dong and Horvath, 2007;Zhu et al, 2007).…”

Novel search method for the discovery of functional relationships

“…BioSim achieved the best performance by consistently ranking functionally related proteins among the top two out of over 18 000 human gene products. BioSim in contrast to other scoring methods might be particular useful for applications based on functional similarity when consistent scores are especially desirable, for example, for the quality assessment of protein–protein interactions (Ramírez et al , 2007) and for the clustering of genes or proteins by function (Huang et al , 2007). We also showed how BioSim can be applied to discover potential disease genes.…”

Research prioritization through prediction of future impact on biomedical science: a position paper on inference-analytics

“…Although PPIs form the basis of many biological phenomena, 90% of the estimated number of PPIs are currently unknown [7-9]. Extensive research is being carried out both with high-throughput biotechnology and computational methods to discover PPIs [10-14].…”