Application of Kohonen maps for solving the classification puzzle in AGC kinase protein sequences

Availability of enormous number of sequences in public domain databases warrants the need for effective tools for clustering and classification of such data. AGC protein kinase family is known to contain many enzymes involved in important cellular processes. In the present study, 21 important physicochemical parameters were calculated for 115 sequences of AGC kinase family belonging to mouse and human. Kohonen maps, also known as Self Organizing Maps (SOM) were employed for the identification of clusters of similar sequences, projection and visualization of high dimensional data spaces owing to their capability of preserving topological relationships between the features. This simplistic approach can provide a method not only for studying intricate interplay of features and minute differences even in the members of same protein family but also for recognition of certain unifying common features. Each cluster obtained using SOM in this study has a distinct characteristic that sets it apart from the other clusters.