The Evaluation and Predictive Properties of the “MemSys3” Algorithm

Abstract

Our initial interest in the quantitative MemSys3 software (the Cambridge software) was its application to spectroscopy and chromatography. Of particular concern was the application to high resolution and in vivo NMR spectroscopy. To this end the results output by the algorithm on known data have been evaluated as part of a major validation exercise. Results obtained have enabled the generation of empirical equations which predict both the ability of the algorithm to deconvolve overlapping peaks and the magnitude of output errors assigned to peak areas. These findings are considered to be a particularly valuable aid in experiment design because it is now possible to predict the outcome of applying the algorithm by measuring defined parameters from the starting data.