From owner-chemistry@ccl.net Fri Jul 11 13:14:01 2008
From: "Rajarshi Guha rguha]=[indiana.edu" <owner-chemistry##server.ccl.net>
To: CCL
Subject: CCL: Estimating applicability of fingerprint model
Message-Id: <-37329-080711131050-13382-9D/I3oFj3niYCMV93Z1Luw##server.ccl.net>
X-Original-From: Rajarshi Guha <rguha_._indiana.edu>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Date: Fri, 11 Jul 2008 12:22:50 -0400
Mime-Version: 1.0 (Apple Message framework v753.1)


Sent to CCL by: Rajarshi Guha [rguha- -indiana.edu]
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On Jul 11, 2008, at 11:33 AM, Iain Wallace iain.m.wallace * gmail.com  
wrote:
> Hi all,
>
> I have built a model to classify compounds into two classes using  
> the Pipeline pilot Bayesian fingerprint classifier (ECFP_4  
> Fingerprints).   I was wondering if anyone has any experience on  
> how to estimate how well the model I have built will transfer to  
> other libraries? I know that I should only apply the model to  
> compounds drawn from a similar distribution, but I have no idea how  
> to what steps I should take to ensure that this criteria is met.

A relatively simple approach is to use the method described in JCAMD,  
2008, 22, 367-384 (http://dx.doi.org/10.1007/s10822-008-9192-9)

- -------------------------------------------------------------------
Rajarshi Guha  <rguha(_)indiana.edu>
GPG Fingerprint: D070 5427 CC5B 7938 929C  DD13 66A1 922C 51E7 9E84
- -------------------------------------------------------------------
Brain fried -- Core dumped


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (Darwin)

iEYEARECAAYFAkh3iNoACgkQZqGSLFHnnoSpXACeI62eKO/AP3HSGuogUC+50xYb
jrUAn2nM0c0rDcCV2vsTDFdsjGmO3Iz5
=RzlN
-----END PGP SIGNATURE-----