Deconvolution and identification algorithms for use on spectroscopic data

摘要

A new algorithm is taught for identifying compounds from spectroscopic or mass spectra data, wherein the improved order of operations of the present invention are defined as 1) background noise removal, 2) deconvolution by smoothing peaks, finding peaks and grouping peaks into unknown compounds, 3) preparing correlation values for combinations of unknown compound and target compound pairs, 4) sorting the combinations of unknown compound and target compound pairs by their correlation values, 5) removing complete ions from the mass spectra data using a peak, a retention time, and a retention window, and 6) matching unknown compounds to target compounds such that no target compound appears twice.

主权项

1. A method of extracting spectra from spectroscopic data obtained by a GC/MS analyzer, the spectroscopic data representing a plurality of different compounds, said method comprising the steps of:
1) receiving spectroscopic data from a GC/MS analyzer; 2) removing any signals having a wavelength that is significantly longer than a typical chromatographic peak from the spectroscopic data by processing the spectrographic data in the GC/MS analyzer before the spectrographic data is processed further; and 3) performing deconvolution on the spectroscopic data using the GC/MS analyzer to thereby find peaks and then group peaks into unknown compounds by first smoothing all peaks using a smoothing filter, and then using the three highest points of each peak of a mass as data points, fitting a parabola to the data points in order to find an intensity of the peak and a time of the maximum point of each peak.