JP2004078371A - Data processor, data processing method and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a mechanism for properly clustering data. <P>SOLUTION: A plot processing part 104 of an SNP data processor 1 plots emitted SNP fluorescent light quantity on coordinates, and an angle information processing part 107 obtains a straight line connecting each of the plotted SNP data with a reference point, and calculates the angle of the straight line with respect to the predetermined reference line, and a clustering processing part 108 clusters the emitted SNP fluorescent light quantity data based on the angle information. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a technique for cluster analysis of data, and more particularly to a technique suitable for clustering data distributed radially from a predetermined reference point such as SNP (Single Nucleotide Polymorphism) data. .
[0002]
[Prior art]
At present, in the field of bioinformatics, research on individual differences based on differences in base sequences of genes is being conducted.
This individual difference is considered to be caused by a difference in the sequence of about 0.1% of the nucleotide sequence of the gene, and this difference is called a genetic individuality (polymorphism). And, among these polymorphisms, those in which only one base is different are referred to as SNP ("single nucleotide polymorphism").
This SNP is presumed to be present at a ratio of about one in several hundred to 1,000 bases, and it is considered that three to ten million SNPs exist in the genome.
By identifying this SNP and analyzing it against factors such as clinical data and the environment, it is possible to say that certain drugs do not work for people with a specific single nucleotide, , Can be used for treatment.
[0003]
As described above, in order to realize the analysis using the SNP, it is necessary to analyze at what location on the genome and at what frequency the SNP exists.
When performing SNP frequency analysis, the TaqMan method, the Invader method, and the like are used. In these methods, the amount of fluorescence emitted from SNP is measured using two kinds of special reagents, and two homozygotes (for example, when the base is A or G, AA and GG) are used based on the fluorescence emission amount data. It is necessary to classify one type of heterozygote (for example, AG) into two or three classifications and determine the frequency of appearance for each classification.
Conventionally, in order to perform SNP frequency analysis, each data is plotted on two-dimensional coordinates based on the amount of light emitted from each reagent, and points on the plotted coordinates are visually clustered into two or three by a person. I was going.
[0004]
[Problems to be solved by the invention]
However, in the conventional method, since the data plotted on the two-dimensional coordinates is visually clustered by a person, the criterion of the clustering is not constant, and not only does the clustering vary, but also the person performs the clustering visually. There are problems such as enormous amount of time required for clustering and human error.
[0005]
Because of this, automation of clustering is required, and it is conceivable to use a general clustering method such as the K-means method.
However, since the SNP data has a characteristic of being radially distributed around a certain center point when plotted on two-dimensional coordinates, it is based on the amount of light emission using a general clustering method such as the K-means method. Since clustering simply clusters data that are close to each other, correct classification cannot be performed, such as formation of clusters that originally span different clusters, such as homo and hetero, making automation difficult. There was a problem.
[0006]
The present invention has been made to solve the above problems and problems, and can appropriately cluster data. In particular, data distributed radially around a predetermined reference point such as SNP data can be obtained. It is an object to provide a mechanism that can appropriately perform clustering.
[0007]
[Means for Solving the Problems]
In order to solve the above-described problem, a data processing device according to one aspect of the present invention is a device for clustering a plurality of data, and a plot processing unit that plots the data on two-dimensional coordinates. An angle information processing means for obtaining a straight line connecting the plotted individual data and a predetermined reference point and obtaining an angle between the straight line and the predetermined reference line; and a clustering process for clustering each data based on the angle information And a data processing device.
[0008]
Further, the data is data on the amount of fluorescence emitted from the SNP of the gene by two predetermined reagents, and the plotting means uses the amount of fluorescence emitted by each reagent as an axis of a two-dimensional coordinate system, and You may make it plot on a two-dimensional coordinate.
[0009]
Further, a predetermined hypothetical point is taken on the coordinates, an angle between a straight line connecting the hypothetical point and each data on the coordinates and a predetermined line passing through the hypothetical point is obtained, and the data is obtained based on the angle information. And a reference point processing means for determining the reference point from the intersection of the principal component straight lines of the respective clusters passing through the obtained cluster centers.
[0010]
In addition, before performing the cluster analysis, a means for extracting a point existing at a predetermined distance from the reference point and excluding the extracted point from the target of the cluster analysis may be further provided.
[0011]
Further, the cluster processing means may further include a display processing means for performing clustering of a plurality of patterns and displaying a result of the clustering of the plurality of patterns performed by the cluster processing means in a selectable parallel manner.
Further, the data is data on the amount of fluorescence emitted from the SNP of the gene by two predetermined reagents, and the plotting means uses the amount of fluorescence emitted by each reagent as an axis of a two-dimensional coordinate system, and You may make it plot on a two-dimensional coordinate.
[0012]
A data processing method according to one aspect of the present invention is a method for classifying a plurality of data into a predetermined class by a computer, wherein the processing of plotting the data on coordinates and the plotting of individual data are performed. And a process for obtaining an angle between the straight line and the predetermined reference line, and a process for clustering each data based on the angle information.
[0013]
A computer program according to one aspect of the present invention is a computer program for classifying a plurality of data into a predetermined class for a computer. And a process of obtaining a straight line connecting the individual data plotted and a predetermined reference point, obtaining an angle between the straight line and a predetermined reference line, and a process of clustering each data based on the angle information. It is characterized by being executed.
[0014]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment in which a data processing device and a computer program according to the present invention are applied to an SNP data processing system will be described with reference to the drawings.
FIG. 1 shows an example of the SNP data processing system according to the present embodiment.
As shown in FIG. 1, the present system includes an SNP data processing device 1 and a result confirmation processing device 2 which constitute a data processing device according to the present invention, a preprocessing device 3, a correction processing device 4, a merge processing device 5, a SNP It is composed of a management device 6.
Each of these devices may be configured to be mutually connectable by a LAN (Local Area Network) or the like, or may be configured to exchange data via a predetermined medium such as a CD-ROM, FD, or MO. May be.
[0015]
The pre-processing device 3 is a computer that performs processing by the TaqMan method or the like.
The pre-processing device 3 performs, for example, a process of measuring the fluorescence emission amount of SNPs on a pair of chromosomes using two kinds of special reagents and generating fluorescence emission amount data of each SNP.
The fluorescence emission amount data is composed of a color and its emission intensity. As for the color, for example, if the gene is homozygous for AA, one of the two reagents emits light, and if the gene is homologous for GG, another reagent emits light and becomes another color. In this case, each reagent emits light, so that the color is an intermediate color.
[0016]
The correction processing device 4 is a computer for performing a correction process on the data rejected by the result confirmation processing device 2.
As the correction processing, for example, the correction processing device 4 refers to the rejected data file (reject file), acquires the fluorescence emission amount data that is the basis of the rejected data from the preprocessing device 3, By displaying this on two-dimensional coordinates, a process is performed so that the operator can visually confirm the distribution of data and perform manual clustering.
[0017]
The merging processing device 5 merges the data confirmed by the result confirmation device 2, the corrected data corrected by the correction processing device 4, and the uncorrectable reject data to create management SNP data. Computer.
The SNP management device 6 is a computer that stores the created management SNP data and manages the SNP data by making it into a database.
[0018]
The SNP data processing device 1 is a device for clustering SNP data. The SNP data processing device 1 is configured by a computer, and includes a CPU (Central Processing Unit), a computer program executed by the CPU, an internal memory such as a RAM and a ROM for storing the computer program and predetermined data, and a hard disk drive. The functional block shown in FIG. 2 can be configured by the external storage device or the like.
The function block shown in FIG. 2 includes an assay data storage unit 101, a call data storage unit 102, a setting processing unit 103, a plot processing unit 104, a reference point processing unit 105, a labeling area processing unit 106, an angle information processing unit 107, and clustering. It comprises a processing unit 108 and a suitability processing unit 109.
[0019]
The assay data storage unit 101 is a storage unit that stores fluorescence emission amount data of each SNP data.
The assay data storage unit 101 can store, for example, the file name of each SNP data, the fluorescence emission amount data of the SNP, and the like.
[0020]
The call data storage unit 102 can store data obtained as a result of performing the clustering process automatically by the SNP data processing device 1.
The call data storage unit 102 can store, for example, each SNP data and cluster information of each SNP data.
[0021]
The setting processing unit 103 can perform setting processing of an initial file and the like.
The plot processing unit 104 performs a process of plotting this on coordinates based on the fluorescence emission amount data of each SNP stored in the assay data storage unit 101. This process can be performed, for example, by the plot processing unit 104 plotting each SNP data on two-dimensional coordinates where the light emission amount of one reagent is on the X axis and the light emission amount of the other reagent is on the Y axis.
[0022]
The reference point processing unit 105 performs a process of obtaining a reference point serving as a reference when measuring an angle on the coordinates of each SNP data.
In this process, for example, the reference point processing unit 105 sets the origin on the coordinates as a temporary reference point (assumed point), and calculates the angle between the straight line connecting each SNP data on the coordinates from the origin and the X axis. The SNP data is classified into a plurality of clusters on the basis of the data, a principal component line at the center of each cluster is obtained, and the intersection of the principal component lines is set as the next assumption point, and the same processing is repeated again, so that the assumption point becomes constant. The point converged as a result can be set as a reference point.
It should be noted that the reference point may be determined not by determining the point at which the hypothetical point converges, but by determining the point obtained by repeating the above-described processing a predetermined number of times as the reference point. Good. Alternatively, when the convergence does not occur, the reference point processing unit 105 may determine the first reference point (such as the origin) as the reference point.
[0023]
The labeling area processing unit 106 performs processing for separating a labeling area to be subjected to cluster analysis and a non-labeling area to be excluded from cluster analysis.
As this processing, for example, the labeling area processing unit 106 takes an intermediate point (x_median_median) between the maximum value and the minimum value of the SNP data, and if the distance from this intermediate point is larger than the threshold value (S), In the case where the size is small, the non-labeling area can be used.
The threshold value (S) for performing labeling may be determined by the operator each time an operation is performed, or may be set in advance by default.
[0024]
The angle information processing unit 107 performs a process of obtaining an angle between a straight line connecting each SNP data from a reference point on the coordinates and a predetermined reference line.
The reference line is a line serving as a reference when measuring the angle. For example, the X-axis or the Y-axis may be used as a reference, or another axis may be used as the reference line, which is optional.
[0025]
The clustering processing unit 108 performs a process of clustering the SNP data based on the angle information of the SNP data.
This clustering process can be performed using an existing clustering algorithm such as the k-means method based on one-dimensional angle information.
When clustering is performed, for example, the clustering processing unit 108 can perform two patterns of processing, that is, two clusters and three clusters in parallel.
[0026]
The matching degree processing unit 109 performs a process of calculating a matching degree indicating whether or not the result of the clustering performed by the classing processing unit 108 matches the clustering of the SNP data.
In this processing, for example, the fitness processing unit 109 sets the distance between the clusters to α and the data variance in each cluster to β, and the larger the ratio of α / β (F ratio), the larger the value. The suitability can be determined as appropriate clustering has been performed. Further, as another fitness of clustering, for example, the fitness processing unit 109 can determine the fitness based on whether or not the allele frequency has a distribution according to the law of Hardy Weinberg equilibrium. The law of Hardy Weinberg equilibrium means that, for example, when the frequencies of alleles M and N are p and q (where p + q = 1), the genotype frequency is represented by the product of the constituent allele frequencies. Is p ² , NN frequency is q ² , MN frequency is a ratio of 2 pq.
[0027]
The result confirmation processing device 2 is a device that performs a process for causing an operator to confirm the processing result obtained by the clustering performed by the SNP data processing device 1.
As shown in FIG. 3, the result confirmation processing device 2 is connected with a display 20 and an input device 30 such as a keyboard and a mouse. The result confirmation processing device 2 is constituted by a computer, and includes a CPU, a computer program executed by the CPU, an internal memory such as a RAM and a ROM for storing the computer program and predetermined data, and an external storage device such as a hard disk drive. Thus, the functional blocks shown in FIG. 3 can be configured.
The functional block shown in FIG. 3 includes a confirmation data storage unit 201, a data input / output processing unit 202, and a display control unit 203.
[0028]
The confirmation data storage unit 201 is a storage unit for storing data for which confirmation of the clustering result has been completed by the operator. The confirmation data storage unit 201 can store the SNP data and cluster information obtained by clustering the SNP data.
[0029]
The data input / output processing unit 202 can perform a process of receiving input of clustering data of the SNP generated by the SNP data processing device 1 and a process of outputting confirmed data to the merge processing device 5.
[0030]
The display control unit 203 displays the clustered result data on the display 20 and performs a process of requesting confirmation from the operator.
When the SNP data processing apparatus 1 classifies two patterns of two rasters and three clusters, the display control unit 203 displays the results of these two patterns on the display 2 in parallel, Can confirm and select the result data.
[0031]
Next, an embodiment of a data processing method according to the present invention will be described with reference to the drawings.
First, the flow of processing of the entire system will be described with reference to FIG.
In FIG. 1, the preprocessing device 3 acquires the fluorescence emission amount data of the SNP using the TaqMan method or the like (S1).
The SNP data processing device 1 acquires the fluorescence emission amount data of the SNP from the preprocessing device 3, performs clustering of the SNP data, and labels the SNP data indicating the cluster (S2).
[0032]
When the clustering is completed, the result confirmation processing device 2 displays the result on the two-dimensional coordinates based on the clustering result data generated by the SNP data processing device 1 and requests the operator to confirm (S3).
As a result of the confirmation, if it is determined that proper clustering has not been performed by the operator, the reject data name is written to a reject file and provided to the correction processing device 4 (S4).
[0033]
The correction processing device 4 acquires the fluorescence emission amount data of the SNP from the preprocessing device 3 based on the reject data name, displays the acquired data on a predetermined display, and allows the operator to visually correct the clustering. (S5).
[0034]
The merge processing device 5 acquires the data determined to be appropriate clustering by the operator's confirmation in the processing of S3, the data corrected by the processing of S5, and the reject data determined to be uncorrectable, and The management data is created by merging the data (S6).
[0035]
When the merge process is completed, the merged management data is provided to the SNP management device 6, and the SNP management device 6 manages the management data by creating a database or the like (S7), and ends the process.
[0036]
Next, a detailed process when the SNP data processing device 1 performs clustering will be described with reference to FIG.
In FIG. 4, first, the setting processing unit 103 refers to the initialization file input by the operator and performs initialization such as predetermined constants for performing clustering (S101).
The initialization file includes, for example, an output file name, an input file name, the number of setting files representing the number of files to be processed, the number of records in one file, and a threshold value of a labeling area (S ), Data such as a rotation angle for placing data in a continuous area of the tangent.
[0037]
The setting processing unit 103 opens the output file stored in the call data storage unit 102 (S102), and determines whether the number of read SNP data files has reached the number of setting files to be processed this time (S102). S103).
As a result of the determination, if the number of set files has been reached, it is assumed that the processing has been completed for all the SNP data files targeted for the current processing, and the processing ends.
[0038]
If the number of set files has not been reached as a result of the determination, the plot processing unit 104 opens the input file stored in the assay data storage unit 101, reads the fluorescence emission amount data of the SNP, and A plot is made on two-dimensional coordinates based on the intensity of the emission color (S104).
The two-dimensional coordinates are obtained by taking the light emission amount of one reagent on one axis (X axis) and the light emission amount of the other reagent on the other axis (Y axis).
[0039]
When the plot is completed, the reference point processing unit 105 converts each SNP data into polar coordinates, and performs one-dimensional clustering using the angle information of the polar coordinates (S105).
For example, the polar coordinates of each data can be obtained based on an angle and a distance between the straight line and the X axis by the reference point processing unit 105 obtaining a straight line connecting each data with the origin on the coordinates as a pole. The clustering can be performed, for example, by the reference point processing unit 105 using a known clustering method such as the K-means method.
[0040]
When the clustering is completed, the reference point processing unit 105 performs a principal component analysis of each cluster by a predetermined principal component analysis algorithm, finds a principal component straight line passing through the center of each cluster, and determines an intersection of these principal component straight lines. (S106).
Then, the reference point processing unit 105 performs the processing of S105 and S106 again a predetermined number of times (n times) using this intersection as a temporary reference point (assumed point), and the intersection n from the obtained intersection 1 converges to a fixed position. It is determined whether or not there is (S107).
If the result of determination is that the intersection has not converged, the process returns to the above-described step S105, and the process is repeated again.
If it is determined in the process of S107 that the intersection has converged to a fixed position, the intersection n is set as a reference point (S108).
[0041]
Instead of determining whether or not the intersection has converged, the number of times of repeating the calculation of the intersection is set in advance, and the reference point processing unit 105 returns to the processing of S105 until the set number of times and repeats the processing. The reference point may be determined based on the result.
[0042]
When the reference point setting is completed, the labeling area processing unit 106 acquires the coordinates (x_median, y_median) of the center point when the SNP data is plotted on the two-dimensional coordinates (S109).
As shown in FIG. 6, the center point (M) finds the maximum value Max and the minimum value Min in the SNP fluorescence emission amount data, and can set the middle (x_median, y_median) as the center point.
[0043]
The labeling area processing unit 106 determines whether the number of read records has reached the set number of records, that is, whether the processing has been performed for all records (S110).
As a result of the determination, if the number has not reached the set number, the labeling area processing unit 106 determines that the value of the coordinates (x, y) of one SNP fluorescence emission amount data is (x / x_median) ² + (Y / y_median) ² It is determined whether or not the threshold value (S) is satisfied, that is, whether or not the distance from the center point is within a predetermined threshold value S (S111).
As a result of the determination, for SNP data having coordinates smaller than the threshold value, the SNP data is set as a non-labeling area, and a flag for excluding the SNP data from target data to be clustered is set (S112).
Also, as a result of the determination, a flag indicating that the SNP data having coordinates larger than the threshold value belongs to the labeling area is set (S113).
FIG. 6 shows an example of this labeling process. As shown by an arc-shaped line in FIG. 6, the area inside the area is a non-labeling area, and the area outside the arc-shaped line is a labeling area.
[0044]
When the number of read records reaches the set number as a result of the determination processing of the labeling area and the determination in S110, the angle information processing unit 107 sets the reference point to the extreme with respect to the SNP fluorescence emission amount data of the labeling area. Is converted to the polar coordinates (S114).
In this process, as shown in FIG. 7, for example, a reference point (origin in the illustrated example) is used as a pole, and an angle between a straight line connecting the reference point and each data and a reference line (for example, X axis) is determined. The SNP fluorescence emission amount data can be converted into polar coordinates based on the measured values.
[0045]
Subsequently, in FIG. 5, when the conversion into the polar coordinates is completed, the clustering processing unit 108 performs one-dimensional clustering by a predetermined clustering algorithm based on the polar coordinate angle information of each SNP data (S115).
At this time, the clustering processing unit 108 performs clustering in two cases, that is, clustering into two clusters and clustering into three clusters. Note that this clustering process can use an existing clustering algorithm such as the k-means method.
FIG. 8 shows an example in which a histogram is created based on angles. In FIG. 8, the horizontal axis represents the angle (radian), and the vertical axis represents the number of data.
[0046]
When the clustering is completed, the clustering processing unit 108 stores the polar coordinates and the clustered two-dimensional coordinates in the call data storage unit 102 (S116).
[0047]
The fitness processing unit 109 calculates the F ratio of the cluster on the coordinates for each of the two clusters and the three clusters (S117).
The fitness processing unit 109 calculates the allele ratio and the like for each of the two clusters and the three clusters, and calculates the degree to which this ratio conforms to the Hardy Weinberg equilibrium rule (S118).
[0048]
Then, the suitability processing unit 109 determines whether proper clustering is performed or rejected data for which proper clustering cannot be performed, based on the F ratio of each cluster and the degree of conformity to the law of Hardy Weinberg equilibrium (S119). ).
This discriminating process can be performed by determining in advance a threshold value of the degree of conformity with the F ratio and the law of Hardy Weinberg equilibrium, and discriminating reject data when these calculation results are lower than the threshold value.
[0049]
When it is determined that the data is reject data as a result of the determination, the matching degree processing unit 109 creates a reject file in which a data name, a sample ID, and the like are described (S120).
[0050]
If it is determined that the clustering is appropriate clustering, or if the process of S120 is completed, the label, the F ratio, and the degree of conformity of the Hardy Weinberg equilibrium rule are stored in the call data storage unit 102 as output files. (S121), the process returns to S103, and the process is repeated until the number of set files is reached, and the process ends.
[0051]
Next, a process when the operator checks the clustering data created by the result checking device 2 will be described with reference to FIG.
In FIG. 9, when the result confirmation processing device 2 provides two patterns of clustering result data from the SNP data processing device 1 via a predetermined network or the like, the data input / output processing unit 202 deletes the provided data. Accept (S201).
The display control unit 203 determines whether the data is clustered or rejected data without clustering based on the F ratio of each cluster of the output file and the fitness data of Hardy Weinberg equilibrium law (S202). ).
[0052]
If the result of the determination is that the data has been properly clustered, the display control unit 203 displays the clustering result with a high degree of conformity as an active state as a two-dimensional graph as shown in FIG. In the example, the left side is an active state), and the clustering result with low fitness is displayed in parallel on the display 2 in a shaded state (S203).
At this time, the data belonging to each cluster may be represented by different colors. Further, the labeling data may be displayed in a round shape, and the non-labeling data may be displayed in a square shape.
[0053]
In this state, the operator compares the two clustering results and visually determines whether the clustering result is correct.
Then, the display control unit 203 determines that the operator has determined that the selected clustering result (the three clusters in the example of FIG. 10) is a correct clustering result, or that the operator has not selected the clustering result (the example of FIG. 10). Then, it is determined which of the two clusters has been judged to be more correct, or which of the two has not been judged to be an appropriate clustering result (S214).
[0054]
When the operator determines that the active clustering result is correct and instructs the “NEXT” radio button in FIG. 10, the display control unit 203 confirms the selected clustering result as an appropriate clustering result. The data is stored in the data storage unit 201 (S205), and the process ends.
If there is data to be processed next, the process returns to the process of S201 and repeats the process.
[0055]
When the operator determines that the clustering result of the shaded one is correct and designates the graph or the radio button (the “2 cluster” radio button in the example of FIG. 10), the display control unit 203 selects The displayed clustering result is switched to the display of the active state and displayed (S206). When the operator designates the NEXT button, the clustering data is stored in the confirmation data storage unit 201, and the process is terminated.
If the operator determines that none of the clustering results is appropriate, the process proceeds to S208 described below.
[0056]
If it is determined in the process of S202 that the data is reject data, the display control unit 203 displays the character “REJECTED” as an active state, and displays any clustering result with a shade ( S207).
In this state, the operator visually checks and determines whether or not the correction can be made, and the display control unit 203 determines whether or not the operator has given an instruction to correct the clustering (S208).
When the operator manually corrects the data, the data is provided to the correction processing device 4 by pointing to the “Manual call” radio button in FIG. 10 (S209), and the clustering result is manually input by the operator. And the process is terminated.
When the operator determines that the original data itself is improper data such as improper data, and instructs the “unable” radio button in FIG. 10, the display control unit 203 The data is stored as reject data in the confirmation data storage unit 201 (S210), and the process ends.
[0057]
As described above, according to the present embodiment, the plot processing unit 104 plots the SNP fluorescence emission amount data on coordinates, and the angle information processing unit 107 obtains a straight line connecting the plotted individual SNP data and the reference point. Since the angle between this straight line and a predetermined reference line is determined and the SNP fluorescence emission amount data is clustered based on the angle information by the clustering processing unit 108, the SNP data distributed radially around the reference point is obtained. Can be properly clustered.
As a result, the SNP data can be automatically clustered, so that uniform and error-free clustering can be performed and the amount of work of the person performing the clustering can be reduced as compared with the case where the SNP data is manually performed.
[0058]
Further, the reference point processing unit 105 takes a predetermined hypothetical point on the coordinate, obtains an angle between a straight line connecting the hypothetical point and each SNP fluorescence emission amount data on the coordinate and a predetermined reference line, and obtains the angle information. Are clustered based on the SNP fluorescence emission amount data, and the intersection of the obtained straight line passing through the center of gravity of each cluster is determined as the reference point. For example, the position where the SNP data is away from the origin of the coordinates Even when the clustering cannot be properly performed based on the origin of the coordinates, for example, when the distribution is distributed, the clustering of the SNP fluorescence emission amount data can be performed after setting an appropriate reference point.
[0059]
Before performing the cluster analysis by the labeling area processing unit 106, a point existing at a predetermined distance from the reference point is extracted and the extracted point is excluded from the target of the cluster analysis. , Clustering can be performed after previously excluding data in a non-labeling area where clustering is likely to cause an error, and more appropriate clustering can be performed.
[0060]
Also, the clustering processing unit 108 performs clustering of two patterns of two clusters and three clusters, and the display control unit 203 displays the clustering results of these two patterns in parallel in a selectable manner. Can compare two clusters, compare each clustering result, and select an appropriate cluster.
[0061]
In the above-described embodiment, an example will be described in which the SNP data processing device 1, the result confirmation processing device 2, the preprocessing device 3, the correction processing device 4, the merge processing device 5, and the SNP management device 6 are configured as separate devices. However, any or all of these functions may be implemented by a single device, which is optional.
[0062]
Further, in the above-described embodiment, the clustering processing of the SNP fluorescence emission amount data has been described. However, the present invention is not limited to this, and can be applied to data distributed radially from a predetermined reference point. .
[0063]
The computer program for the SNP data processing device 1 or the result confirmation processing device 2 of the present embodiment may be stored in a computer-readable medium (FD, CD-ROM, etc.) and distributed, or may be superimposed on a carrier wave. Distribution via a communication network is also possible.
When the OS (Operating System) shares the functions of the SNP data processing device 1 or the result confirmation processing device 2 or realizes the functions of the SNP data processing device 1 or the OS and the application program jointly, only the part other than the OS is used as a computer program. The computer program may be stored in a computer-readable medium, or the computer program may be distributed.
[0064]
【The invention's effect】
According to the present invention, data can be appropriately clustered, and in particular, data distributed radially around a predetermined reference point, such as SNP data, can be appropriately clustered.
[Brief description of the drawings]
FIG. 1 is a diagram showing an outline of an embodiment of a system using a data processing apparatus according to the present invention and a flow of processing.
FIG. 2 is a functional block diagram of the SNP data processing device according to the embodiment;
FIG. 3 is a functional block diagram of a confirmation processing device according to the embodiment;
FIG. 4 is a processing flow showing a processing flow of the SNP data processing device according to the embodiment;
FIG. 5 is a subsequent processing flow showing the processing flow of the SNP data processing device according to the embodiment;
FIG. 6 is an exemplary view showing the concept of a labeling process according to the embodiment;
FIG. 7 is an exemplary view showing a concept of a process of acquiring angle information of each data on coordinates according to the embodiment;
FIG. 8 is a view showing an example of a histogram based on angle information according to the embodiment;
FIG. 9 is an exemplary view showing the flow of the process of the confirmation processing device according to the embodiment;
FIG. 10 is an exemplary view showing an example of a screen of the confirmation processing device according to the embodiment.
[Explanation of symbols]
1 SNP data processing device
2 Result confirmation processing device
104 Plot processing unit
105 Reference point processing unit
106 Labeling area processing unit
107 Angle Information Processing Unit
108 Clustering processing unit
203 Display control unit

Claims (8)

An apparatus for clustering a plurality of data,
Plot processing means for plotting each of the data on two-dimensional coordinates;
Angle information processing means for obtaining a straight line connecting the individual data plotted and a predetermined reference point, and obtaining an angle between the straight line and a predetermined reference line,
Clustering processing means for clustering each data based on the angle information;
A data processing device comprising: The data is data distributed radially from the reference point when plotted on two-dimensional coordinates.
The data processing device according to claim 1. A predetermined hypothetical point is taken on the coordinates, an angle between a straight line connecting the hypothetical point and each data on the coordinates and a predetermined line passing through the hypothetical point is obtained, and clustering of the data is performed based on the angle information. And a reference point processing means for determining the reference point from the intersection of the principal component straight lines of the respective clusters passing through the obtained cluster centers.
The data processing device according to claim 1. Before performing the cluster analysis, a point that is present at a predetermined distance from the reference point is extracted, and further includes a unit that excludes the extracted point from a target of the cluster analysis.
The data processing device according to claim 1. The cluster processing means performs clustering of a plurality of patterns,
Display processing means for selectively displaying a plurality of pattern clustering results performed by the cluster processing means in a selectable parallel manner;
The data processing device according to claim 1. The data is fluorescence emission amount data of the SNP of the gene by two predetermined reagents,
The plotting means plots the fluorescence emission amount data on the two-dimensional coordinates, using the fluorescence emission amount of each reagent as an axis of the two-dimensional coordinates, respectively.
The data processing device according to claim 1. A method for classifying a plurality of data into a predetermined class by a computer,
A process of plotting each of the data on coordinates,
A process of obtaining a straight line connecting the individual data plotted and a predetermined reference point, and obtaining an angle between the straight line and a predetermined reference line;
A process of clustering each data based on the angle information;
A data processing method comprising: A computer program for classifying a plurality of data into a predetermined class for a computer,
Against the computer
A process of plotting each of the data on coordinates,
A process of obtaining a straight line connecting the individual data plotted and a predetermined reference point, and obtaining an angle between the straight line and a predetermined reference line;
A process of clustering each data based on the angle information;
A computer program that executes

Images