RASCAL (Rapid Similarity Calculation)
RASCAL is based on clique detection, using both existing and novel techniques to increase efficiency, but has two characteristics that make it particularly suitable for the measurement of chemical similarity
Detection of the maximum common edge subgraph (MCES), rather than the normal MCS, using a line-graph representation
Use of a bounding procedure to eliminate many target structure-to-database structure comparisons without the need for clique detection (cf screening in substructure searching)