Welcome to the 2011 QSAR Challenge!

Questions, answers, discussions related to SIAM SDM'11 Contest

Welcome to the 2011 QSAR Challenge!

Postby robertf » Mon Nov 15, 2010 11:11 pm

Thank you for your interest and participation in SIAM ICDM'2011 QSAR Challenge. On the surface it seems quite easy: a single endpoint - a biological property of chemical molecules, and a binary classification problem. But don't be fooled, almost all standard classification algorithms "break their teeth" on this data set. Submitted solutions will be evaluated on model's final predictivity (i.e., on how well it will predict the biological property of new molecules), not on it's training performance, and not on it's preliminary test results. Designing good measures of predictivity will thus be a part of this competition. Most likely, a lot creative thinking will be required to get a successful, balanced outcome. The data is real and represents real challenges faced by Cheminformatics scientists - mining noisy biological data for useful information leading to discovery of new medical drugs. Overall, this challenge should be quite interesting.

Good luck!
Organizing Committee
robertf
 
Posts: 32
Joined: Wed Oct 13, 2010 7:27 pm

Re: Welcome to the 2011 QSAR Challenge!

Postby jamesxli » Wed Nov 17, 2010 8:12 pm

Looks like a very interesting problem. Where can I find information about the data format for solutions?
jamesxli
 
Posts: 19
Joined: Wed Dec 09, 2009 6:55 pm

Re: Welcome to the 2011 QSAR Challenge!

Postby robertf » Wed Nov 17, 2010 9:31 pm

"Solutions should contain a list of labels ("S" or "N"), one per line. There should be 163 labels for Preliminary Test examples followed by an empty line and another 209 labels for Final Test examples."

The last 209 Final Test examples will be ignored in Preliminary submissions until the contest deadline. But surely they must be present in your final submission!

More details in http://tunedit.org/challenge/QSAR?m=task

Enjoy!
robertf
 
Posts: 32
Joined: Wed Oct 13, 2010 7:27 pm

Re: Welcome to the 2011 QSAR Challenge!

Postby jamesxli » Tue Nov 23, 2010 7:20 pm

Is there a sample solution file available?
I have submitted several times with different formats, but always got error evaluation.
What line delimit is required (\r\n or just \n) ? Must the N and S put into quota? etc.
jamesxli
 
Posts: 19
Joined: Wed Dec 09, 2009 6:55 pm

Re: Welcome to the 2011 QSAR Challenge!

Postby robertf » Tue Nov 23, 2010 7:32 pm

Dear James,

1) Please remove all double quotes
2) After your first 163 solutions add an empty line
3) After that add your 209 solutions (in the same format) for 209 final compounds

Best regards,
Robert
robertf
 
Posts: 32
Joined: Wed Oct 13, 2010 7:27 pm

Requesting about Molecular Structure

Postby jricardo » Thu Nov 25, 2010 11:25 pm

Dear Robert,
the tittle of the challenge is "SIAM SDM'11 Contest: Prediction of Biological Properties of Molecules from Chemical Structure". however only the molecular descriptors (ADMET Prediction) of three data set are given. My researcher group has been working on the definition of new molecular descriptors. I really appreciate if you send me the chemicals structures in some informatics file (*.smi, *.sdf, or some on).

thanks you in advance
Best Regards

UCLV Team.
jricardo
 
Posts: 1
Joined: Thu Nov 18, 2010 5:42 am

Re: Welcome to the 2011 QSAR Challenge!

Postby domcastro » Fri Nov 26, 2010 5:03 pm

Hi

Did a .sdf file become available? or the Pubchem CIDs?

EDIT: The competition says "structure" but I think it should say "properties"

thanks
Mandy
domcastro
 
Posts: 10
Joined: Fri Nov 26, 2010 5:00 pm

Re: Welcome to the 2011 QSAR Challenge!

Postby robertf » Fri Nov 26, 2010 10:02 pm

Dear Teams,
Sorry, NO extra data or files will be released; at least not before the end of this contest. And no .sdf files are available. This competition must be kept fair to everyone. I hope you understand.

With best regards,
Robert
robertf
 
Posts: 32
Joined: Wed Oct 13, 2010 7:27 pm


Return to SIAM SDM'11 Contest: Prediction of Biological Properties of Molecules from Chemical Structure

Who is online

Users browsing this forum: No registered users and 1 guest

cron