Re: QSAR - Modeling biological data

Suzanne, I agree with Professor Franke on the following: mass correction
should be your primary concern. Thus, you need to adjust the 2^1 ... 2^7
mg/ml values [it looks like dilutions?] to molar concentrations for each
compound. For > 128 mg/ml you should try (1) ignoring these rows or (2)
using 2-3 orders of magnitude higher values, in this case 2^9...2^10 .
Of course, you have to be *careful* since this is an artificial value,
and in fact it could be close to 256 or 4096! Perhaps it is best if you
can factor in some experimental error as well - a range of values that
you can select from - before you discard/validate this experimental
descriptor. Ultimately, it must help you derive a consistent model.

Good luck,

Tudor

Tudor I. Oprea, MD PhD
Director, Office of Biocomputing
Professor of Biochemistry and Molecular Biology
MSC08 4560
University of New Mexico School of Medicine
Albuquerque, NM 87131, USA
tel (505) 272 3694
fax (505) 272 8738
email toprea%salud.unm.edu
http://poblano.unm.edu/biocomp/ (academic)
http://www.sunsetmolecular.com (private)

>>> drugdes*o*snafu.de 02/07/03 08:46AM >>>
NB: Unless you reset the To: line, your reply goes to the entire list

---
On 6 Feb 2003 at 9:09, Suzanne Sirois wrote:
> NB: Unless you reset the To: line, your reply goes to the entire
list
> ---
> 
> Hello,
> I would like to develop a QSAR model using calculated descriptors as
well as experimental values. The problem with the experimental values is
the range. They are from 2 to 128 mg/ml. But some could not be
determined and they are said to be > 128 mg/ml. How can I take into
account in my training 
set of data that are greater than.
> 
> Thank you with your help
> 
> 
> Suzanne Sirois, Ph.D
> Cheminformatics, Computational Chemistry
> 
> _______________________________________________
> qsar_society mailing list
> qsar_society^_^accelrys.com 
> http://ftp2.accelrys.com/mailman/listinfo/qsar_society 
Hello,
if some of your experimental descriptors have only values of the type >
128 mg/ml 
you have the following options:
- Eliminate the corresponding compounds from your training set. If
calculations                  
show that this particular descriptor is of no importance for the
biological activity to 
be considered, the eliminated compounds can, of course, again be
included.
- Do not use this descriptor.
- It might also be possible to derive some model for this descriptor
which would then 
allow to estimate the missing values. 
I do not know what your measured values are. Since, however, drugs act
on a molar 
and not on a mass basis a correction with the molecular weight might be
necessary.
Best regards
Rainer Franke
-- 
Prof. R. Franke / Dipl.-Chem. A. Gruska
Consulting in Drug Design GbR    phone: +49-33397-62396
Gartenstrasse 14                          fax: +49-33397-62154
D-16352 Basdorf bei Berlin          e-mail: drugdes*snafu.de 
home page: www.consulting-drug-design.com 
_______________________________________________
qsar_society mailing list
qsar_society|accelrys.com 
http://ftp2.accelrys.com/mailman/listinfo/qsar_society