The result of this investigation has been reviewed in 'Trends in Analytical Chemistry' and is available here:
I hope that this paper promotes the integration of spectral similarity searches and spectrum prediction engines during the peer-reviewing process. The largest NMR-database available is the combined CSEARCH+SPECINFO collection named NMRPredict (available from MODGRAPH, see http://www.modgraph.co.uk/ ) holding some 450,000 spectra at the moment with the expectation of massive upgrades in the near future. Details about the ongoing data-extraction process can be found on:
The integration of the CSEARCH-based prediction engines into data processing programs is also possible - the available DLL fits into MESTRENOVA ( http://www.mestrec.com/ ) and into Bruker's TopSpin-program ( http://www.bruker-biospin.com/topspin.html ).
Stay tuned - more to come ! Despite my 'impressive office', which can visited on http://nmrpredict.orc.univie.ac.at/csearchlite/Home_of_CSEARCH.html ;-))