RE: QSAR - new question on chirality descriptors

Francisco,

>I should appreciate if you answer me the following question:
>In order to do a QSAR on chiral compounds (e.g. barbiturates) and to
>include chirality, which chiral descriptor seems more interesting:
>a) experimental (+)/(-) deviation of the polarization plane, or
>b) "geometric" (R)/(S) IUPAC (Cahn-Ingold-Prelog) structure?

Both descriptors have problems:
The specific rotation (a) must be experimentally determined
and (b) depends on the wavelength (sodium D line). By changing the
wavelength, the value and even the sign might change.
The CD spectrum is a much better description of the optical properties.

The absolute configuration (R/S) describes a local chiral element,
and not the optical properties of the molecule. Also, there is no
relationship between R/S and the sign (+/-) of the specific rotation,
even when a compound has only one chiral center. Changing a substituent
might change the specific rotation sign, without changing the
absolute configuration.

For chiral compounds, 3-D QSAR is the best approach, folowed by
docking with a good scoring function (of course, if the receptor is known
or if it can be modeled by homology modeing).

Best regards,
Ovidiu Ivanciuc
Received on 2004-09-09 - 14:01 GMT