Article ID: 123 - Last Modified: December 4, 2010
Which is faster: shape screening or pharmacophore screening? And does having excluded volumes affect the search speed?
A range of speeds is observed with shape-based searching. The typical speed with similarity filtering turned on is 500-700 conformers per second
(depending on the hardware of course), which usually translates to about 10 compounds per second. The very slowest searches are probably a factor
of 2 to 3 slower than this.
The range of speeds is greater with pharmacophore-based searching. In cases where the pharmacophore model is overly generic (e.g., AHH), or because partial matching is being used, a database search can be pretty slow. In fact, the variation in speed can be orders of magnitude. An extremely selective pharmacophore model that matches less than 1% of compounds may proceed at an average rate of hundreds of compounds per second. By contrast, if one builds a 10-point pharmacophore and requests matching on 5 out of 10 sites, it may take seconds to search each compound. A typical search is probably 25-50 compounds per second.
Using excluded volumes has a minimal impact on speed.
Keywords: Phase, pharmacophore, hypothesis, partial matching, shape,
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles