A Hierarchical Approach for Clustering and Pattern Matching of Gene Expression Data

Clustering data is a well-known and challenging problem that has been widely studied in data base analysis. This paper shows how it made possible in genetic engineering to observe simultaneously the expression levels of huge genes during important genetic processes and identifies similar pattern genes and maximum matching regulation genes. This paper presents simple and static method to find regulation of gene from statistical gene expression data, comparison of regulation of genes and then form clusters with genes having similar regulation. Also extracts sub clusters from big cluster and perform hierarchical analysis. Finding coherent patterns or means from sub-clusters and by comparing sub cluster genes with coherent pattern it refines clusters and form fine clusters. This is a simple gradient base technique of finding regulation, hence it easily removes noisy genes. This technique relates bottom up approach. Also this process relates Density-Based clustering approach, hence can be done using density estimation. This scheme can extract clusters efficiently with reduced number of comparisons. Extracting the patterns from large genetic data enhances identification of similar characteristic genes and variety of hidden information about genes. Grouping similar data in large dimension spaces is to hidden pattern or meaningful subgroups have many applications in biological and other fields.