# US 2006/0217926 A1 · Method for processing a set of spectra, particularly NMR spectra

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2006/0217926 A1
- Document ID: US20060217926A1
- Status: Published application
- Kind: A1
- Application number: 11/386,442
- Filed: 2006-03-22
- Published: 2006-09-28
- Earliest claimed: 2005-03-24
- Assignees: Hoffmann-La Roche Inc.
- Inventors: Frank Dieterle; Alfred Ross; Goetz Schlotterbeck; Hans Senn
- Web page: https://patentgrep.com/patent/US20060217926A1

## CPC

- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G01N 24/08: by using nuclear magnetic resonance (G01N24/12 takes precedence)
- G01R 33/4625: {Processing of acquired signals, e.g. elimination of phase errors, baseline fitting, chemometric analysis}
- Y10T 436/24: Nuclear magnetic resonance, electron spin resonance or other spin effects or mass spectrometry
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12)
- Y10T: TECHNICAL SUBJECTS COVERED BY FORMER US CLASSIFICATION

## Abstract

A method for processing a set of spectra, particularly NMR spectra, comprises the steps of: a) selecting a principal spectral range; b) recording a plurality of principal spectra in said principal spectral range; c) obtaining a reference principal spectrum in said principal spectral range; d) carrying out, for each one of said principal spectra, a bin-wise division of the principal spectrum by said reference principal spectrum to obtain a corresponding set of spectral quotients; e) calculating, for at least one of said principal spectra, an associated set of statistical measures derived from the corresponding set of spectral quotients; and f) carrying out, for at least one of said sets of statistical measures, an outlier detection test.

## Claims

### Claim 1

A method for processing a set of spectra, particularly NMR spectra, comprising the steps of:
a) selecting a principal spectral range;
b) recording a plurality of principal spectra in said principal spectral range;
c) obtaining a reference principal spectrum in said principal spectral range;
d) carrying out, for each one of said principal spectra, a bin-wise division of the principal spectrum by said reference principal spectrum to obtain a corresponding set of spectral quotients; and
e) calculating, for at least one of said principal spectra, an associated set of statistical measures derived from the corresponding set of spectral quotients;
characterized in that it further comprises the step of:
f) carrying out, for at least one of said sets of statistical measures, an outlier detection test.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the set of statistical measures corresponding to a principal spectrum comprises the median of its spectral quotients and an interquartile difference obtained by subtracting the first quartile of its spectral quotients from the third quartile of its spectral quotients.

### Claim 3 (depends on 2)

The method according to claim 2, wherein said outlier detection test comprises determining whether said interquartile difference exceeds a predefined threshold width.

### Claim 4 (depends on 3)

The method according to claim 3, wherein said threshold width is determined from the distribution of said interquartile differences taken over the entire set of principal spectra.

### Claim 5 (depends on 2)

The method according to claim 2, wherein said outlier detection test comprises determining whether said median of spectral quotients differs from a constant by more than a predefined threshold quotient deviation.

### Claim 6 (depends on 5)

The method according to claim 5, wherein said threshold quotient deviation is determined from the distribution of said median of spectral quotients taken over the entire set of principal spectra.

### Claim 7 (depends on 1)

The method according to claim 1, further comprising the steps of:
a) selecting an auxiliary spectral range that does not overlap with said principal spectral range,
b) recording, with each one of said principal spectra in said principal spectral range, an associated auxiliary spectrum in said auxiliary spectral range;
c) obtaining a reference auxiliary spectrum in said auxiliary spectral range;
d) carrying out, for each one of said auxiliary spectra, a bin-wise division of the auxiliary spectrum by said reference auxiliary spectrum to obtain a corresponding set of auxiliary spectral quotients; and
e) calculating, for each one of said auxiliary spectra, an associated set of statistical measures derived from the corresponding set of auxiliary spectral quotients;
wherein said outlier detection test comprises comparing statistical measures of a principal spectrum and of an associated auxiliary spectrum.

### Claim 8 (depends on 1)

The method according to claim 1, wherein step 1e) is carried out for each one of said principal spectra so as to obtain a complete set of statistical measures from which is derived a set of global statistical measures, and wherein said outlier detection test is carried out for said set of global statistical measures.

### Claim 9 (depends on 2)

The method according to claim 2, wherein said set of global statistical measures comprises:
a) the median of the complete set of said interquartile differences; and
b) the interquartile difference of the complete set of said interquartile differences.

### Claim 10 (depends on 1)

The method according to claim 1, wherein each one of said principal or auxiliary spectra is subjected to a normalization procedure before carrying out said bin-wise division.

### Claim 11 (depends on 10)

The method according to claim 10, wherein said normalization procedure for any one of said principal or auxiliary spectra comprises the steps of:
a) applying to said spectrum a pre-processing to obtain a pre-processed spectrum;
b) calculating an integrated intensity of said pre-processed spectrum; and
c) multiplying said pre-processed spectrum by a normalization factor that is proportional to the inverse of said integrated intensity.

### Claim 12 (depends on 1)

The method according to claim 1, wherein said reference principal or auxiliary spectrum is obtained as the median of a plurality of blank or control spectra recorded in the corresponding principal or auxiliary spectral range, respectively.

### Claim 13 (depends on 1)

The method according to claim 1, wherein said reference principal or auxiliary spectrum is obtained from a subset of said principal or auxiliary spectra.

### Claim 14 (depends on 8)

The method according to claim 8, wherein said set of global statistical measures comprises:
a) the median of the complete set of said interquartile differences; and
b) the interquartile difference of the complete set of said interquartile differences.

### Claim 15 (depends on 9)

The method according to claim 9, wherein each one of said principal or auxiliary spectra is subjected to a normalization procedure before carrying out said bin-wise division.

### Claim 16 (depends on 11)

The method according to claim 11, wherein said reference principal or auxiliary spectrum is obtained as the median of a plurality of blank or control spectra recorded in the corresponding principal or auxiliary spectral range, respectively.

### Claim 17 (depends on 11)

The method according to claim 11, wherein said reference principal or auxiliary spectrum is obtained from a subset of said principal or auxiliary spectra.
