Andrea Wendel is a student at Potsdam University in upstate New York and is one of Martin Walker’s chemistry majors. Martin is on our advisory group for ChemSpider. Andrea was kind enough to answer the question “How do you use ChemSpider” and I felt it was of value to share the short report with the readers. We’d love to hear from other users about how you use ChemSpider and your feedback that could be shared with blog readers. Thanks

How I use ChemSpider

I use ChemSpider on a regular basis when I need additional or supplemental information about molecules or reactions. Homework or lab write-ups require the most need for this website. I use the main box in the center of the page to search for my information. I mainly use systematic names, trade names or formulas to search for my data. Most results appear in a very short time, which is very helpful. I rarely use the structure search option, although if my type of research was different I might find it more useful. For many of my lab write-ups and papers, the information I search for includes common physical properties such as boiling point, melting point, and density. The molecular weight and safety comments are also items I find extremely useful. The icon at the end of the information that leads you to the webpage it came from is another useful option, as the original source is very important to know. ChemSpider has many features that are useful to an undergraduate student.

By: Andrea Wendel

Stumble it!

2 Responses to “How Andrea Wendel Uses ChemSpider”

  1. Wolfgang Robien says:

    Dear Andrea:

    I fully agree with your statements regarding common physical properties, etc. – the most important sentence in your feedback is “The icon at the end of the information that leads you to the webpage it came from is another useful option, as the original source is very important to know.” Yes, the original source is extremely important to know !

    Lets do a few examples:

    Do a carbon prediction for the following 3 structures: #21934 , #26477 and #11249

    http://www.chemspider.com/Chemical-Structure.21934.html (2-methylundecane)
    http://www.chemspider.com/Chemical-Structure.26477.html (2,2-dimethyldecane)
    http://www.chemspider.com/Chemical-Structure.11249.html 4-(2-phenylpropan-2-yl)phenol

    1) All methylgroups in these 3 compounds will be predicted at 20.3 ppm
    2) The methylgroup in an n-alkylchain, the methylgroups in (formally) an isopropyl group and the methylgroups in a (formally) tert.-butylgroup resonate always at 20.3ppm
    3) The methylgroups in the bisphenol-A type compound resonates also at 20.3ppm

    These unbelievable values have been calculated by NMRSHIFTDB as implemented into Chemspider on April 22nd, 2010 – this feature is now available for more than 1 month ! (Screendumps available on request from wolfgang.robien(at)univie.ac.at)

    My questions:

    1) Nobody out there in the community who
    a) recognizes this nonsense and
    b) gets up the nerve to clearly state that (except me) !
    2) Is software provided by Stefan Kuhn and Christoph Steinbeck guaranteed to be untested ?
    3) Is ‘open data, open source’ the legitimation to bother the community with definitely wrong predictions far away from reality ?
    4) A few months ago proton prediction was online and vanished within a few days after bad feedback – why is NMRSHIFTDB still online ?

    Conclusion:

    1) Grant-Paul increment rules from 1964 produce MUCH BETTER predictions than a software-package
    written approx. 40 years later !
    2) For e.g. undecane reasonable values are given, but in this case it’s simply a visualization of the database values and NO prediction is done (BTW: Q: How to distinguish between simple visualization of a database-spectrum and a prediction based on NMRSHIFTDB ? A: When clicking the box ‘NMRSHIFTDB’ and the spectrum appears after a few seconds then the identity is in the database and only the display must be done, when you can easily take a sip of coffee and do a phone call when waiting for the result, it’s a prediction ;-) ) )
    3) Tony – Sorry ! I like Chemspider too as an excellent aggregation of data, according to my impression these predictions severly decline the confidence into all the other data presented on CHEMSPIDER !

    Wolfgang

  2. Antony Williams says:

    Stefan Kuhn is working on the fix for some of the errors. He already has a beta version up with the fix but its not integrated into ChemSpider yet. For now we have disabled the integration.

    We have on our list to distinguish between a DB hit versus a prediction. We’ll likely get it fixed before we switch it back on.

    Thanks for the feedback as usual

Leave a Reply