Data processing: measuring – calibrating – or testing – Calibration or correction system
Reexamination Certificate
1999-09-16
2002-05-07
Grimley, Arthur T. (Department: 2852)
Data processing: measuring, calibrating, or testing
Calibration or correction system
C250S339090, C250S339120, C702S028000, C702S032000, C436S023000
Reexamination Certificate
active
06385549
ABSTRACT:
This application is the national phase under 35 U.S.C. §371 of PCT International Application No. PCT/DK98/00110 which has an International filing date of Mar. 20, 1998 which designated the United States of America.
TECHNICAL FIELD
The present invention relates to measurement of small amounts of a specified component in a fluid, and especially acetone and/or acetoacetate in milk.
PRIOR ART
It has for some time been desirable to be able to measure the content of acetone in milk in order to have a tool for an early detection of ketosis (a metabolic disease) in dairy cows. Methods for the determination are in fact available, but generally they are time-consuming. Besides the content of acetone there will typically be a related amount of acetoacetate. When ever in this specification the word acetone appears this means acetone and/or acetoacetate.
Recently IR spectrometry has become a preferred method for analysing milk, and accordingly it would be advantageous also to use IR spectrometry for finding the content of acetone and/or acetoacetate.
The application of IR spectral data for determination of concentrations of components in a composition is known e.g. from: WO 9516201 (Foss Electric A/S), WO 9624832 (Foss Electric A/S), U.S. Pat. No. 5,121,337, (Brown), U.S. Pat. No. 5,252,829, (Nygaard et al), U.S. Pat. No. 5,606,164 (Price et al) and EP 0751388 (Kyoto Dai-Ichi). From WO 9516201 it is known to determine added water and the related freezing point depression from IR spectral data From this document it is also known to be advantageous to increase the leverage of the known calibration samples by adding extra water to natural samples. U.S. Pat. No. 5,121,337 (Brown) discloses a method for correcting spectral data for data due to the spectral measurement process itself. Further it discloses how to estimate an unknown property and/or composition data of a sample by use of such method. U.S. Pat. No. 5,252,829 (Nygaard), owned by the applicant, discloses a method of determining urea in milk. The content of urea in milk is generally above zero but fairly low, and the successful urea determination described in the patent is based on thorough compensation for the influence by other components on the urea measurement, through use of contemporary determinations of the contributions from the other components, i.e. fat, lactose and protein. The samples used for calibration are generally samples having a considerable amount of urea, i.e. within the intended measuring range. U.S. Pat. No. 5606164 (Price et al) discloses a method and apparatus for biological fluid analyte concentration measurement using generalized distance outlier detection.
However, it has until now been considered impossible to use IR spectrometry for obtaining reliable detections of the amounts of acetone appearing in cows suffering from ketosis, cf. Hendrik-Jan Luinge, B. Lutz, P. Dobbelaar and Y. H. Schukken: “Infrared spectrometry as a sensor for the early detection of ketosis in cows”, 1996. Poster presentation, S.O.N. Analytische Chemie, Lunteren, Nov. 6-7, 1995.
The present invention provides a method for calibration of an IR spectrometry apparatus for providing and evaluating IR spectra in order to determine very low contents of specific components in a fluid, such as milk, and especially low contents of acetone, such as from about 0.5 mM to about 2.5 mM acetone in milk.
The art of extracting information on the chemical content of fluid compositions from measured spectra has for some time been based on a process of “leaming” or “calibrating” the IR spectrometry apparatus to enable the data processor of the apparatus to recognize certain components in a fluid. Normally, the spectra of about 15-20 known samples are measured and used together with the known values of the content in the samples to derive a calibration for the IR spectrometry apparatus. It is general knowledge to people in the art that the set of samples used for calibration must be representative for the desired range of measurements. There are several methods of calculating such calibrations, and many methods (such as PCR, MLR or PLS regression) are well known to people in the art. Hitherto the methods and apparatus avaiable have generally only been able to provide reliable measurements of contents which happen to appear in substantial amounts, i.e. that makes up a considerable fraction of the fluid, such as fat, protein and lactose in milk. In the case of acetone and/or acetoacetate the content in milk is generally zero or about zero.
SUMMARY OF THE INVENTION
The present invention provides a method for calibration of a spectrometry apparatus for providing and evaluating spectra for determination of very low concentrations of a specific component in a fluid in a specified measuring range above 0, (e.g. from 0.5 mM to 4.0 mM), by which calibration method a number of variables and corresponding coefficients (so-called B coefficients) are determined according to methods for multivariate calibration, such as PCR, MLR or PLS regression, comprising selecting and measuring a set of calibration samples including at least 30-50 known samples, e.g. from 50-300 known samples, for the calibration. According to the invention the set of calibration samples includes a number of samples representing the fluid without any substantial content of the specific component for the calibration, i.e. samples being below the specified measuring range.
The new method is based upon the use of a great number of known samples, which are measured by a spectrometry apparatus, providing a spectrum of each of the known samples (the content of the “known” samples either being known or determined by a reference method) and providing a calibration (e.g. by applying known calibration calculation methods such as PCR, MLR or PLS regression and by applying principal variables or genetic algorithms for variable selection).
The method was specifically developed for the determination of acetone in milk by use of IR spectroscopy. However, it is contemplated that the method is applicable to other types of spectroscopy for the determination of small amounts of other specific components.
Regarding the acetone and/or acetoacetate content, experience has indicated that a substantial number, e.g. about 20-30, and even better about 100 or 150, i.e. the vast majority of the calibration set (the known samples) may represent samples having none or almost no content of acetone.
The method according to the invention is advantageous because the vast majority of available samples are samples having no or almost no content of acetone. Only the few cows suffering from ketosis will provide samples which are representative for the range to be determined in order to be able to decide whether a cow suffer from ketosis or not. The most obvious solution to that problem would be to enrich a great number of natural samples with suitable amounts of acetone to provide a good calibration set.
According to the inventors experience such extended enrichment is not necessary. A few enriched samples and/or natural samples from cows suffering from ketosis will do, and the vast majority of the calibration samples can be natural samples from healthy cows, i.e. samples without any significant amount of acetone, and samples outside the desired measuring range. In this way a fairly accurate determination is possible. Further the provision of the calibration set is fairly easy, as the vast majority of samples may be natural samples.
According to a further advantageous method some of the calibration samples may be enriched samples, i.e. samples having zero or almost zero content of the specified component, whereto a number of predetermined, known amounts of the specified component being added. Accordingly a preferred set of calibration samples comprises a great number of samples having almost no content content of the specified component, and a small representative selection of samples covering the intended measuring range.
Further scope of the applicability of the present invention will become apparent from the detailed description g
Birch & Stewart Kolasch & Birch, LLP
Foss Electric A/S
Grimley Arthur T.
Le John
LandOfFree
Measurement of acetone in milk using IR spectroscopy does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Measurement of acetone in milk using IR spectroscopy, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Measurement of acetone in milk using IR spectroscopy will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2875913