RDKit community,
I signed up for this mailing and I'm trying it out for the first time
because I'll be discussing our work comparing the CDK toolkit to commercial
descriptors and models for QSAR and ADMET predictions tomorrow at the ACS
CINF *(March 25, 2012 at 2:50PM in Room 26A). *
PAPER TITLE: “Open source QSAR models with IP-sensitive SAR data: POCs in
secure online environments for collaborative drug discovery showing CDK
with SMARTS keys are statistically equivalent (κ = .43)…and open call for
all industry and academic projects”
DIVISION: COMP: Division of Computers in Chemistry
SESSION: Collaborative Drug Discovery for Neglected Diseases
LOCATION: San Diego Convention Center, Room 26A
I figured this would be a good way to get the word out and to connect with
others presenting at the ACS or who can't make it but want to connect via
email, web, gotomeeting, etc
-Barry
P.S. Here is the abstract with the reference to the full paper, bottom
line, there was and is nothing magical about commercial descriptors over
open source descriptors, suggesting with a little effort a convenient
mechanism could be created for collaboratively sharing models and methods
(disclaimer: at CDD (www.collaborativedrug.com) we've focused on doing
this for the underlying SAR data capture, sharing and publishing). Happy
to collaborate with anyone in the open or commercial modes....
Building upon existing selective and secure chemical registration and SAR
data sharing capabilities is a new research initiative for selective and
secure model sharing. Initial published results (see: “Using open source
computational tools for predicting human metabolic stability and additional
ADME/Tox properties.” Gupta R, Gifford EM, Liston T, Waller CL, Hohman M,
Bunin BA, Ekins S. Drug Metab Dispos, 38: 2083-2090, 2010) have
demonstrated that free, open source descriptors and models using CDK with
SMARTS keys are statistically equivalent (k = .43) to expensive commercial
descriptors and models for ADME/Tox prediction with some of the largest
training sets used to date (for 193,000 compounds). Initial efforts are
focused on ADME/Tox and Tuberculosis QSAR models, although the approach is
extensible to any SAR/QSAR data sets. A call is now open for projects using
distributed data sets and corresponding models to accelerate collaborative
drug discovery.
--
Barry A. Bunin, PhD
Collaborative Drug Discovery (CDD), Inc.,
CEO & President
Tel: (650) 204-3084
Fax: (650) 522-9498
Email: [email protected]
Web: www.collaborativedrug.com
LinkedIn: http://www.linkedin.com/in/collaborativediscovery
Blog: Collaboration as the Key to Turning Around the Drug Discovery
Business<https://www.collaborativedrug.com/buzz/2012/01/20/collaboration-as-the-key-to-turning-around-the-drug-discovery-business-part-7-collaboration-leverage/>
<https://www.collaborativedrug.com/buzz/2012/01/20/collaboration-as-the-key-to-turning-around-the-drug-discovery-business-part-7-collaboration-leverage/>
--
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
