Ideally a program to calculate a hydrophobic parameter from structure should be:

1. comprehensive -- no (or very few) 'missing fragments'
2. fast, to handle databases of millions of compounds
3. informative -- gives more than a number from a black box

Characteristics (1) and (2) are vital for similarity/diversity studies on the large 'real' and 'virtual' databases resulting from combinatorial chemistry. When (3) is desired, more time is required as well as a close interaction with molecular mechanics calculations.

Presentation