Liquid purification or separation – Processes – Chromatography
Reexamination Certificate
2001-08-02
2004-06-01
Therkorn, Ernest G. (Department: 1723)
Liquid purification or separation
Processes
Chromatography
C210S198200, C250S282000, C702S019000, C702S022000
Reexamination Certificate
active
06743364
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to the field of data processing and evaluation. In particular, the invention relates to the processing and evaluation of mass chromatographic and mass spectrometric data.
BACKGROUND
Developments both in mass spectrometric technology and in the combination of mass spectrometers (“MS”) with a broad variety of separation and micro-scale separation techniques, are quickly increasing the capacity of MS in terms of data production. Using modern instrumentation, the time required to obtain the above-mentioned data, such as chromatograms and mass spectra, is no longer the critical factor; rather, it is the time necessary for analyzing the data. In particular, a data set often comprises thousands of mass spectra measured over a mass-to-charge (“m/z”) range of two to three orders of magnitude. An extended study using such a data set can occupy days if a complete analysis is required. In a research environment in particular, this analysis typically must be carried out by highly qualified, and consequently expensive, personnel.
In this context, the use of efficient data processing and evaluation to improve speed in data handling is highly desirable. Depending on the application, information extraction can be approached from different points of view. In impurity studies by capillary electrophoresis/mass spectrometry (CE/MS) or liquid chromatography/mass spectrometry (LC/MS), for example, data processing and evaluation tools must be able to perform efficient peak detection of compounds present at very low levels. On the other hand, if screening and comparison of very similar complex mixtures is to be performed, such as in the rapidly expanding field of proteomics, data processing and evaluation tools must be able to correlate data on multiple complex mixtures.
One prior approach to processing data produced by a combination of mass spectrometry and chromatography is U.S. Pat. No. 5,672,869 to Windig et al. (“the '869 patent”). The '869 patent describes a data processing approach which separates spurious peaks and noise by smoothing the raw data. This approach then compares processed and raw data. If a mass trace contains only background noise, the difference between raw and processed data is emphasized, and the algorithm assigns a low mass chromatographic quality (“MCQ”) value to that particular mass trace. On the other hand, mass traces containing a peak are assigned high MCQ values. The '869 patent then teaches selecting only mass traces that possess a MCQ above an appropriate threshold value.
However, it is not necessarily clear what is an appropriate threshold value, especially for complex and/or noisy data. For example, by selecting a threshold which is too high, some relevant information on low intensity signals may be lost, while setting too low a threshold may select many “signals” that are actually just background noise. As a result, extensive visual examination of raw and processed data by trained personnel may be required to address this problem, and thereby lower data processing efficiency and speed.
A need therefore exists for a data processing technique that provides more efficient and clear data processing.
SUMMARY OF THE INVENTION
The present invention adapts an information content theory and combines it with data smoothing to provide a measure of data quality that better facilitates the efficient and clear evaluation of data. In particular, the present invention provides a measure of data quality based on what is referred to herein as an entropy value. The entropy value approach of the present invention improves data processing by providing less ambiguous thresholds for data selection. As a result, for example, the present invention speeds data processing by decreasing the amount of time trained personnel may be required to personally inspect and select data.
The present invention provides a method of data processing in which the separation between spurious peaks and noise on the one hand, and relevant data on the other hand, takes place more accurately and clearly, thereby shortening the data analysis time. Consequently, trained personnel can use their time interpreting the data. At the same time, the present invention provides the option of generating fingerprints of complex mixtures, which are increasingly being used in various fields (chemistry, pharmacy, medicine, biology, biotechnology, and the like), but particularly in the life sciences, for example, from the analysis of biological materials oriented towards DNA fragments, proteins and metabolic components.
In one aspect, the present invention provides a method of data processing and evaluation comprising the steps of smoothing the data points of a chromatogram and determining an entropy value for the smoothed-chromatogram. In one embodiment, the method also comprises the step of correcting the data points of a chromatogram for baseline prior to determination of an entropy value for the smoothed-corrected chromatogram. The chromatogram may be either a mass chromatogram or a total ion current (“TIC”) chromatogram. It should be realized that the order of the smoothing and baseline correcting steps is unimportant to the present invention. That is, a chromatogram may be smoothed then baseline corrected, or baseline corrected and then smoothed. Accordingly, it is to be understood that the term “smoothed-corrected chromatogram” does not imply a specific order of practice.
In another embodiment, the method of the invention further determines a quality factor (i.e., an “IQ value”) for a chromatogram based on the evaluation of entropy values for a plurality of chromatograms of a data set. In a preferred embodiment, the method selects individual chromatograms (of either corrected-chromatograms and/or smooth-corrected chromatograms) based on their IQ values. The method then uses these selected chromatograms to generate a reconstructed total ion current (“RIC”) chromatogram. The method may further exclude from the RIC chromatogram one or more mass signals. In one embodiment, the one or more mass signals are selected for exclusion based on a mass signal quality value for the individual mass signals. In another embodiment, the method uses these selected chromatograms to generate a reconstructed mass chromatogram for one or more mass values. Further, in various embodiments, the RIC chromatograms are used as a fingerprint for comparison to other chromatograms of the same or other data sets.
In another aspect, the present invention provides a method of data processing and evaluation that correlates either a smoothed-chromatogram or a smoothed-corrected chromatogram with a plurality of chromatograms of a data set. The chromatogram of the smoothed-corrected-chromatogram (or smoothed-chromatogram) and the data set chromatograms may be, for example, a mass chromatogram, total ion current chromatogram, or a RIC chromatogram. In a preferred embodiment, the step of determining a correlation comprises using a multivariate analysis. Suitable forms of multivariate analysis include, but are not limited to, principal component analysis (“PCA”), discriminant analysis (“DA”), partial least squares (“PLS”), predictive linear discriminant analysis (“PLDA”), neural networks, and pattern recognition techniques.
In another embodiment of the present invention, the entropy values of a plurality of smoothed mass chromatograms are each calculated and stored, followed, if desired, by processing of these entropy values or, as the case may be, components selected according to these entropy values by means of chemometric and biometric methods. Preferred forms of component selection include multivariate analysis techniques (PCA, DA, PLS, PLDA, neural networks), pattern recognition techniques and Fourier transform techniques. In another embodiment, the selected components are further used to generate a fingerprint and that is used, in conjunction with chemometric and biometric techniques, as a characterization method for complex mixtures of various origins.
In another aspect, the present invention prov
Nederlandse Organisatie voor Toegepast-Natuurwetenschappelijk On
Testa Hurwitz & Thibeault LLP
Therkorn Ernest G.
LandOfFree
Method and system for identifying and quantifying chemical... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for identifying and quantifying chemical..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for identifying and quantifying chemical... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3302945