2007 年 127 巻 10 号 p. 1712-1718
Various clustering methods are applied to characterizing gene expression data from DNA microarrays. However, to elucidate functional dependencies of genes analysis of their associations is important. The REVerse Engineering ALgorithm (REVEAL) was developed for analyzing the functional dependencies. Although the algorithm has been tested using binary models of genetic networks, it remains unclear how the method or similar technology will operate with systems of continuous variables. In this study, first, the REVEAL was examined using noisy, continuous data and the results suggested that its application to such data required considerable refinement of the algorithm. Then a new implementation method of REVEAL was proposed. This implementation method was tested through numerical experiments. The results of the simulations demonstrated the potential of the proposed method for extracting gene associations.