Power Boosts for Cluster Tests

摘要

Gene cluster significance tests that are based on the number of genes in a cluster in two genomes, and how compactly they are distributed, but not their order, may be made more powerful by the addition of a test component that focuses solely on the similarity of the ordering of the common genes in the clusters in the two genomes. Here we suggest four such tests, compare them, and investigate one of them, the maximum adjacency disruption criterion, in some detail, analytically and through simulation.