Improved Parameterized Algorithms for Network Query Problems

Abstract

Abstract. Biological networks are often explored through extensive use of network queries. Partial Information Network Queries (PINQ) address the major challenge of analyzing such networks in the absence of certain topological data. In the PINQ problem, we are given a host graph H, modeling a network, and a pattern P, whose topology is only partially known. We seek a subgraph of H that resembles P. In this paper, we study a generalization of PINQ which allows near resemblance between H and P. We obtain an exact parameterized algorithm as well as an FPT-approximation scheme for a wide class of inputs, where the parameter is the number of nodes in P. Our algorithms substentially improve the best O * running times in solving PINQ, as well as the special cases of the Alignment Network Query problem and the Topology-Free Network Query problem.