Topological Properties of Protein Interaction Networks

The study of biological networks is an important area of medicine, as it allows the development and modification of drugs and therapies, as well as the detection of diseases in the early stages of their inception. Despite the wide variety of developed methods for finding and predicting protein-protein interactions, the number of undetected interactions remains large. Expanding knowledge about known proteins and their functional modules can be critical to improving the efficiency of computational methods. We investigated the topological properties of protein interaction networks and the correlation of the presence of a link between proteins and the path length between them. According to the results obtained, a universal method for predicting interactions using paths of a certain length has not yet been developed, and the effectiveness of existing algorithms depends on the data set used. In addition, we conducted a percolation analysis and were able to identify evolutionarily ancient ribosomal cluster.