The paper considers the problem of enhancing the semantic capabilities of information systems, which improve the quality of the solution of the search problem on demand. The information in the system is assumed to be presented in the form of an information graph. The paper considers queries specifying a sample of the structure of a graph that must be received as a response, or exemplar queries. The response is supposed to be received to the query using the semantic filtering, based on the ranking of the exemplar results similar to the query graph. The problem of manipulating similar information graph structures is considered, this problem is necessary for semantic filtering. The paper offers to solve the problem basing on the use of methods of applicative computing systems and transformational semantics. The paper discusses the architecture and means of supporting the search for and manipulation of similar structures of information graphs.