Discovering gene co-regulatory relationships is a new yet important research problem in DNA microarray data analysis. The problem of gene specific co-regulation discovery is to, for a particular gene of interest, called the target gene, identify its strongly co-regulated genes from the database and the experimental condition subsets where such strong gene co-regulations are observed. The study on this problem can contribute to a better understanding and characterization of the target gene. The existing technique, mainly using genetic algorithm (GA) to discover co-regulation conditional subsets, is slow due to its expensive fitness evaluation and long solution encoding scheme. In this paper, we propose a novel technique to improve the performance of gene specific co-regulation discovery using a bit freezing approach. Through freezing converged bits in the solution encoding strings, this innovative approach can contribute to fast crossover and mutation operations, achieve an early stop of the GA and facilitate the construction of kNN Search Table Plus (kNN-ST+) that leads to more accurate approximation of fitness function. Experimental results with a real-life gene microarray data set demonstrate the improved efficiency of our technique compared with the existing method. |