We had previously announced the ability to perform NMR prediction via ChemSpider using the NMRDB.org services. Based on a number of comments from people testing the system, as well as some of my own tests, we have chosen, for the time-being at least, to remove the connection to this NMR prediction service from the website.

We do not believe that the issue is our integration. A prediction for the proton spectrum of benzene on the nmrdb.org website gives the H1 NMR spectrum below.

GOOD NMR prediction is not easy. I’m an NMR spectroscopist by training. I’ve either been using NMR prediction tools or helping to design and build them for almost 2 decades. Some publications of interest are listed below:

Toward More Reliable 13C and 1H Chemical Shift Prediction: A Systematic Comparison of Neural-Network and Least-Squares Regression Based Approaches J. Chem. Inf. Model.  Web Release Date: 05-Dec-2007; 10.1021/ci700256n

The Performance Validation of Neural Network Based 13C NMR Prediction Using a Publicly Available Data Source. J. Chem. Inf. Model., J. Chem. Inf. Model., 48 (3), 550 -555, 2008. 10.1021/ci700363r

Automated structure verification based on 1H NMR prediction. Magn. Reson. Chem., 44, 524 (2006)

Computer-Aided Determination of Relative Stereochemistry and 3D Models of Complex Organic Molecules from 2D NMR Spectra, Tetrahedron, 61, 9980-9989 (2005)

We definitely want to have an NMR prediction capability hooked up to ChemSpider. We would like that to be a free service if possible but are also open to hooking up for-fee services and the users of ChemSpider can choose whether to use them, for free or fee, or not. If anyone has an interest in hooking up services to ChemSpider feel free to connect with me.

We have contacted the nmrdb.org hosts about the issue also.

Stumble it!

Leave a Reply