On the gene team mining problem

Let Σ be a set of n genes. A chromosome G can be represented as a permutation of Σ. A subset D of Σ is a δ-set of G if two consecutive genes in G∩D has distance at most δ. For a set G of m chromosomes, a set D is a δ-team of G if D is a δ-set of every chromosome of G. Given a gene set Σ, a chromosome set G, and an integer δ, the gene team finding problem is to find all possible δ-teams of G. Given a gene set Σ, a chromosome set G of m chromosomes, an integer k ≤ m, and an integer δ, the gene team mining problem is to find all possible δ-teams for any possible chromosome set G´ such that G´ ⊆ G and |G´| ≥ k. In this paper, we study the gene team mining problem. It is known that the Apriori technique is used wildly in data mining. However, the gene team mining problem has no Apriori property that all nonempty subsets of a δ-team (δ-set) must also be a δ-team (δ-set). Thus, many techniques used in data mining cannot be applied for this gene team mining problem. In this paper, we propose a concept of pseudo-support. By using this concept, an Apriori-like algorithm can be obtained to solve the gene team mining problem.