A modified fuzzy co-clustering (MFCC) approach for microarray data analysis

Biologically a gene or a sample could participate in multiple biological pathways, and only few genes are concurrently involved in a cellular process under some specific experimental conditions. Hence, identification of a subset of genes showing similar regulations under subsets of condition in microarray data has become an important research issue. Many investigators develop bi-clustering methods to attack this problem. In this study, we adopt fuzzy co-clustering concept and design a procedure to iteratively extract bi-clusters with co-expressed gene patterns (here the entire proposed process is called a modified fuzzy co-clustering (MFCC) approach). We have applied synthetic data and compared our MFCC's performance with four well-known state-of-the-art methods. Here we have not only shown that our MFCC approach can successfully extract each designed bi-clusters in the synthetic data sets, but also have demonstrated the better performance by our MFCC approach.