October 11, 2007, ACD/Labs NJ NMR Seminar 2007, New Brunswick, NJ

High-Throughput NMR Analysis: The End Game

Anthony Macherone

Abstract

In support of the pharmaceutical industry, our analytical lab includes liquid chromatography-mass spectrometry (LC-MS), tandem gas chromatography-mass spectrometry (GC-MS/MS), and nuclear magnetic resonance (NMR). The NMR has a 300 MHz field with a 4-channel multi-probe (¹H, ¹³C, ¹⁹F, ³¹P) and provides walk up support for a synthetic chemistry team, support for directed parallel synthesis, quality control analysis of a large commercial library composed of diverse reactive intermediates, capping groups, scaffolds, and fragments in addition to high throughput analysis of non-trace pure compounds or compounds purified by preparative LC-MS. More than 1000 compounds are directed to analysis by 1-D proton NMR weekly. To achieve the required throughput, the analytical "game" has been divided into three phases: pre-, middle-, and end-game.

Pre-game: Assign system suitability a priori
At its essence, a priori screening is a system of virtual software filters that allow personnel to assign a suitable analytical method for a compound in the absence of empirical data. Successful a priori screening "bins" compounds into appropriate analytical queues such as LC-MS, NMR, or GC-MS based upon each compounds inherent physicochemical properties. Through this tool one can screen a chemical library in silico and create sub-sets that are directed towards a specific method of chemical analysis.

Middle-game: Data acquisition
Compounds that have been binned into the NMR queue are prepared by the compound management team who plate each compound at a pre-determined mass in 96-well format under inert conditions. To that, 750 µl of an appropriate NMR solvent is added robotically under inert conditions. The ¹H isotope interferogram is collected averaging over 8 or 16 scans.

End-game: Data analysis and management
Upon completion of an automation run (100 samples) the data is processed via ACD/1D NMR Processor software v10.07. To this end an sdfile containing each compound's 2-D structural representation and an appropriate "key-field" that associates each molecule with it's corresponding FID is created and associated to a custom macro. The macro enables the ACD/Labs processing software to: attach the sd representation of the structure to the appropriate FID, Fourier transform, phase, baseline correct, integrate the spectrum, calculate a theoretical spectrum, fit the empirical data to the theoretical, provide "red-light/green-light" data assessment, and finally store the data in a searchable database. The time required to perform the verification on a set of 100 compounds is ten minutes or less—a significant improvement over sitting at your desk or at home in your pajamas interpreting NMR spectra. The quality of the verification ranges between 60-80% accuracy, which can be greatly improved through training sets. The resulting data set can be exported as an sdfile for archiving, and future data sets can be incorporated into the database. It is the auto-verification process facilitated by the ACD/Labs software that permits the expected throughput of greater than 1000 1-D NMR analysis each week with corresponding data management and archival.

Download the presentation in Adobe Acrobat format (574 Kb PDF file).