Development of a fast and accurate method of 13 C NMR chemical shift prediction

abstract Article history:Received 20 April 2008Received in revised form 25 December 2008Accepted 29 January 2009Available online 11 February 2009Keywords:Computer-assisted structure elucidationNMR chemical shift calculationPLS regression In this article we describe a fast and accurate method of 13 C NMR chemical shift prediction. The high speed ofchemical shift calculation described is achieved using a simple structure description scheme based onindividual atoms rather than functional groups. The systematic choice of an appropriate encoding schemeand the usage of partial least squares regression on a large training set has resulted in a robust and fastalgorithm. The approach provides accuracy comparable with other well known approaches but demonstratesaccelerated calculation speeds of up to a thousand times faster.© 2009 Elsevier B.V. All rights reserved. 1. IntroductionIn various fields of chemistry such as the investigation of naturalproducts or the design of new compounds, scientists often need toeither determine de novo the structure of an unknown or newcompoundor toverifya hypotheticalchemicalstructure.Thisprocess,known as Structure Elucidation, is based on the analysis of availablespectral data. Nuclear Magnetic Resonance (NMR) spectroscopy iscertainly one of the main analytical methods applied to thesechallenges and is a powerful technique for acquiring highly informa-tive spectra associated with a structure.Nuclear magnetic resonance (NMR) is a physical phenomenonbased upon the quantum mechanical magnetic properties of an atom'snucleus. Magnetic nuclei, like