David Koes edited subsection_Shape_Indexing_and_Similarity__.tex  over 8 years ago

Commit id: e2e2a1de97aa893a8764921aebdb147835f52929

deletions | additions      

       

where a larger score indicates a greater degree of similarity.  We consider both similarities with a single query ligand and similarities with shape constraints. The similarity of a shape, $A$, with the minimum, $MIN$, and maximum, $MAX$, shape constraints is computed by averaging their shape Tanimotos:  \begin{equation*}  \begin{aligned}  \delta(A,MIN,MAX) &  = \frac{\delta(A,MIN) + \delta(A,MAX)}{2} \\  &  = \frac{1}{2}\left(\frac{A \cap MIN}{A \cup MIN} + \frac{A \cap MAX}{A \cup MAX }\right) \end{aligned}  \end{equation*}  The average ranges from zero to one, where a value of one is only achieved when all three shapes are identical.