A neural-network approach for biclustering of gene expression data based on the plaid model

Biclustering techniques, for simultaneous row-column clustering, are widely used in the analysis of the gene expression data. Many different biclustering techniques have been proposed, such as the iterative signature algorithm (ISA) (Bergmann et al., 2003), global biclustering (Wolf et al., 2006), evolutionary fuzzy biclustering (Mitra et al., 2007), etc. Among these techniques, the plaid model is often used for multivariate data analysis. However, difficulties exist because there are mixed binary and continuous variables in this model for which the traditionally used optimization algorithms suitable for continuous variables cannot be employed in the realization of the biclustering process. In this paper, a novel neural-network approach is proposed to tackle such a mixed binary and continuous optimization problem in the plaid model. Experiment results show that the accuracy of the biclustering can be significantly improved with the proposed algorithm.