Clustering of chemical structures on the basis of two-dimensional similarity measures