The authors suggest two approaches to constructing hypotheses about the interactions of proteins on the basis of the results of tracking the changes in their levels during the studied biological process. The first approach is based on the use of machine mining of biological texts, which are freely available on the Internet. The second is related to the execution of database queries on binary protein-protein interactions . To specify the experimental context in obtained protein networks, filtering of the protein entities on the basis of needed ontology from database Gene Ontology was used.

Keywords: text mining, data base, protein-protein interaction, protein networks