Mobility Normalizer, Normalizing Method, Normalizing Program, Self-Organized Map, Substance Detecting Method, Detecting Program, Detection Rule Creating Method, and Data Structure

• US 20090136139A1
• Filed: 07/11/2006
• Published: 05/28/2009
• Est. Priority Date: 11/17/2005
• Status: Active Grant

• Current Assignee: Kochi University
• Original Assignee: Kochi University
• Inventors: Kataoka, Hiromi
• Granted Patent: US 7,991,569 B2
• Time in Patent Office: Days
• US Class Current: 382/207
• CPC Class Codes: G01N 27/44717 Arrangements for investigating the separated zones, e.g. localising zones
• CPC Class Codes: G01N 30/8644 Data segmentation, e.g. time windows
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 2218/02 Preprocessing

Abstract

The object of the present invention is to provide an apparatus, a method, a program, and a self-organizing map which are capable of normalizing mobility without using a marker, as well as a substance detection method, a program, a detection rule creating method, and a data structure which use normalized mobility. The mobility normalizing method comprises the steps of determining a plurality of warping functions converting data to be corrected, which is unit time sequence data obtained by measuring mobility, to the respective plurality of reference waveform data sets, and a DTW distance associated with each warping function; evaluating a minimum value of the plurality of DTW distances, and determining the warping function associated with the determined minimum DTW distance; determining a slope and an intercept of a straight line approximating the determined warping function; and correcting the data to be corrected using a linear function specified by the slope and the intercept.

Claims

1. A mobility normalizing apparatus comprising a recording unit containing a plurality of reference waveform data sets, and a processing unit;

wherein

the processing unit;

when externally acquired mobility measurement data is unit time sequence data, defines the measurement data as data to be corrected;

when externally acquired mobility measurement data is a plurality of two-dimensional data sets associated with time information and measured values, interpolates these two-dimensional data sets to create the data to be corrected, which is unit time sequence data;

reads the plurality of reference waveform data sets from the recording unit, and determines a plurality of warping functions converting the data to be corrected to the respective plurality of reference waveform data sets, and a DTW distance associated with each warping function;

evaluates a minimum value of the plurality of DTW distances, and determines the warping function associated with the determined minimum DTW distance;

determines a slope and an intercept of a straight line approximating the determined warping function; and

corrects the data to be corrected using a linear function specified by the slope and the intercept.

2. A mobility normalizing apparatus according to claim 1, wherein the processing unit calculates a slope between given vertices for each determined warping function, excludes slopes included in predetermined areas at both ends when the plurality of slopes between vertices are arranged in order of increasing size, and determines the slope of the straight line approximating the warping function based on the slopes between vertices which are not excluded.

3. A mobility normalizing apparatus according to claim 2, wherein the processing unit determines a median of the plurality of slopes between vertices as the slope of the straight line approximating the warping function.

4. A mobility normalizing apparatus according to claim 1, wherein the plurality of reference waveform data sets are data obtained by correcting mobility measurement data for an analyte containing a marker substance, using the mobility of the marker substance.

5. A mobility normalizing apparatus according to claim 1, wherein the plurality of reference waveform data sets are data forming a self-organizing map created using a plurality of mobility measurement data sets.

6. A mobility normalizing method comprising:

a first step of determining a plurality of warping functions converting data to be corrected, which is unit time sequence data obtained by measuring mobility, to respective plurality of given reference waveform data sets, and a DTW distance associated with each warping function;

a second step of evaluating a minimum value of the plurality of DTW distances, and determining the warping function associated with the determined minimum DTW distance;

a third step of determining a slope and an intercept of a straight line approximating the determined warping function; and

a fourth step of correcting the data to be corrected using a linear function specified by the slope and the intercept.

7. A mobility normalizing method according to claim 6, wherein the data obtained by measuring mobility is a plurality of two-dimensional data sets associated with time information and measured values;

and wherein

the data to be corrected is unit time sequence data created by interpolating the plurality of two-dimensional data sets.

8. A mobility normalizing method according to claim 6, wherein the third step comprises:

a fifth step of calculating a slope between given vertices for each warping function determined in the second step; and

a sixth step of excluding slopes included in predetermined areas at both ends when the plurality of slopes between vertices are arranged in order of increasing size, and determining the slope of the straight line approximating the warping function based on the slopes between vertices which are not excluded.

9. A mobility normalizing method according to claim 6, wherein the sixth step is a step of determining a median of the plurality of slopes between vertices as the slope of the straight line approximating the warping function.

10. A mobility normalizing method according to claim 6, wherein the plurality of reference waveform data sets are data obtained by correcting mobility measurement data for an analyte containing a marker substance, using the mobility of the marker substance.

11. A mobility normalizing method according to claim 6, wherein the plurality of reference waveform data sets are data forming a self-organizing map created using a plurality of mobility measurement data sets.

13. A computer-readable recording medium containing a self-organizing map used as the reference waveform data in the mobility normalizing method of claim 6;

the self-organizing map being created by a learning process executed by a computer, using waveform data obtained by visually correcting data to be corrected, which is unit time sequence data obtained by measuring mobility of an analyte, or using waveform data obtained by correcting data to be corrected, which is unit time sequence data obtained by measuring mobility of an analyte containing a marker substance, based on the mobility of the marker substance.

14. A substance detection method using a decision tree determined as a detection rule by learning that uses supervised data;

the decision tree being determined by learning that uses;

normalized data obtained by correcting a plurality of mobility measurement data sets for a plurality of analytes, using the mobility normalizing method of claim 6; and

the supervised data including information representing the presence or absence of a given substance, which is associated with each measurement data set.

15. A substance detection method according to claim 14, wherein

the substance is M protein;

and wherein,

under the condition that the mobility of albumin is 400 and the mobility of dimethylformamide in the gamma region is 800, the decision tree is described in the form of combinations of mobilities and determination signs as shown in FIG. 14 or FIG.

16. A computer-readable recording medium containing a substance detection program;

the program causing a computer to implement a function of detecting a substance, using a decision tree determined by learning that uses supervised data as a detection rule;

the decision tree being determined by learning that uses;

normalized data obtained by correcting a plurality of mobility measurement data sets for a plurality of analytes, using the mobility normalizing method of claim 6; and

the supervised data including information representing the presence or absence of a given substance, which is associated with each measurement data set.

17. A method for creating a detection rule for use in detecting a substance, the method comprising:

a first step of using a computer to correct a plurality of mobility measurement data sets for a plurality of analytes, using the mobility normalizing method of claim 6, to create normalized data; and

a second step of determining a decision tree as a detection rule by learning, using supervised data including information representing the presence or absence of a given substance, which is associated with each measurement data set, and the normalized data.

18. A computer-readable recording medium containing a data structure for use as a substance detection rule;

the data structure representing a decision tree determined by learning that uses;

normalized data obtained by correcting a plurality of mobility measurement data sets for a plurality of analytes, using the mobility normalizing method of claim 6; and

supervised data including information representing the presence or absence of a given substance, which is associated with each measurement data set.

19. A computer-readable recording medium containing a data structure according to claim 18, wherein

the substance is M protein;

and wherein,

under the condition that the mobility of albumin is 400 and the mobility of dimethylformamide in the γ

region is 800, the decision tree is described in the form of combinations of mobilities and determination signs as shown in FIG. 14 or FIG. 15.

12. A computer-readable recording medium containing a mobility normalizing program;

the mobility normalizing program being used in an apparatus comprising a recording unit containing a plurality of reference waveform data sets, and a processing unit; and

causing the processing unit to implement;

a first function of determining a plurality of warping functions converting data to be corrected, which is unit time sequence data obtained by measuring mobility, to the respective plurality of reference waveform data sets, and a DTW distance associated with each warping function;

a second function of evaluating a minimum value of the plurality of DTW distances, and determining the warping function associated with the determined minimum DTW distance;

a third function of determining a slope and an intercept of a straight line approximating the determined warping function; and

a fourth function of correcting the data to be corrected using a linear function specified by the slope and the intercept.