You are visitor:

• Draw your structure using ISISDRAW or any other structure drawing software able to export MOLfiles
• Export the structure as MOLfile
• Prepare an email to "wr@alpha.univie.ac.at" holding exactly ONE MOLfile; the MOLfile can be either send as Mail-body or as "text-attachment"
• In any case the MOLfile must be plain text; don't use base64-encoding or uuencode.
• Optionally select a keyword and/or sample identifier

In order to support project- and sample-management any additional information can be entered into the subject-field behind the required processing method. This text will be used as subject for the reply.
 

Examples:

Your subject-line might look as below:
 
 

Subject: nnet3d  Project: Health-care / Sample-#4810

Subject: hose2d; sample: Not Fiagra / Stereoisomer-#1 / Project: High Life

 

You will receive two emails; the first one will hold an ASCII-protocol of the spectrum estimation procedure, the second one will contain a Postscript-file showing the usual CSEARCH-datasheet as given above.
 

• Solvents used: All HOSE-code calculations are performed using all solvents available with the reference data collection; all neural network calculations are done for CDCl₃
• Stereochemistry: The parameter ON/OFF in the headerline shows the utilization of stereochemical information
• Database: 75,000 C-13 NMR spectra
• Shifttable: The shifttable on the right-hand side gives you the following information

One line within the shifttable consists of the following information:
C-aa:   bbb.b   M   c   Min   Max   Hits
 

aa	Number of carbon
bbb.b	Predicted chemical shiftvalue
M	Signal multiplicity
c	Bond level used for calculation (only for HOSE-code)
Min	Lowest expected shiftvalue
Max	Highest expected shiftvalue
Hits	Number of hits contributing to this result

 

• Use stereochemical information whenever it is possible - stereochemistry is entered by using up/down-bonds on each stereocenter
• sp³-carbons should be handled at least at the 3-bond level to give reliable results
• sp²- and sp-carbons should be handled at least at the 4-bond level to give reliable results
• The number of hits should be at least 3
• Preview your Postscript file always with an appropriate previewer which supports colors and/or print the Postscript-file on a color-printer; any information which is in doubt will be given in red color and additionally marked by greyscale-printing. This is a serious warning automatically done by the CSEARCH-software in order to give you some additional information about the reliability of the predicted spectra.
• Always use HOSE-code and neural network technology and compare the results.
• When doing more than 10 spectrum estimations it would be nice to use a delay of 5 minutes between the requests in order to allow other people to do a single calculation in between - the server-procedure is running with a repetition rate of 4 minutes in order to avoid any system overload.