JP2006277611A - Multiple sample gene expression data analysis system, method, program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an analysis technique and analysis system for analyzing gene expression data without using level data and without correcting gene expression data distribution among samples in analysis on gene expression data among multiple samples. <P>SOLUTION: The multiple sample gene expression data analysis system comprises: an input device 1 for inputting gene expression data of multiple samples, a data analysis device 2; and an output device 3 for outputting an analysis result. The data analysis device 2 comprises: an order statistics calculation means 21 for calculating order statistics of a gene expression data set for one sample from among N sets of gene expression data from the input device 1; a level data generation means 22 for assigning a level for the gene expression data for every of the (N-1) samples and generating the (N-1) sets of level data; a data conversion means 23 for converting the gene data expression data sets using the order statistics and the (N-1) sets of level data; and a gene expression estimation means 24 for estimating a gene expression situation using the converted (N-1) sets of gene expression data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an analysis method, analysis system, and recording medium for gene expression data of a plurality of samples, and in particular, an analysis system, method, program, and recording for analyzing gene expression data without correcting the distribution of gene expression data among samples. It relates to the medium.

  In recent years, with the development of oligo- and Stanford-type DNA microarrays and macroarrays, comprehensive gene analysis has been widely studied and put into practical use. For example, DNA microarrays for gene expression analysis, DNA chips for SNP typing, CGH arrays, etc. are widely used for genome research.

  For example, in the study of disease-related genes, comprehensive gene expression analysis using microarrays is performed on samples of disease groups and healthy groups, and it is examined whether the expression level of genes changes between groups. Search for genes related to diseases. In studies on drug responsiveness, drugs are administered to cultured cells while changing the dose, and genes whose expression level changes with changes in dose are examined.

  In research using such a microarray, information on thousands to tens of thousands of genes can be obtained from one sample. Then, using the obtained large amount of gene information (such as gene expression), the relationship between the disease or drug responsiveness, side effect, etc. and the gene is analyzed using statistical analysis or data mining techniques.

  In microarray experiments, the experimental process is very complex, and various experimental errors and systematic distortions occur at each stage of the experiment. For this reason, several methods for correcting experimental errors have been proposed. For example, referring to the following Patent Document 1 and Non-Patent Document 1, the experimental error of the microarray is corrected using an order statistic or a nonparametric smoothing method. Furthermore, referring to Patent Document 2 below, it is possible to quantitatively analyze whether or not the expression is different between samples by estimating the expression status of the gene using a mathematical model and obtaining the posterior probability of the expression status. ing.

  Analysis methods for gene expression data obtained from multiple samples include, for example, statistical tests using t-tests and U tests to determine whether there is a difference in expression levels between groups, and expression between groups I check how many times the amount has increased or decreased. However, as described above, in a microarray experiment, experimental errors and systematic distortions occur between samples, so that the distribution of gene expression data between different samples is different. Therefore, in order to compare gene expression data between samples, it is necessary to correct the distribution of gene expression data between the samples.

  On the other hand, for example, referring to Patent Document 3 below, a rank corresponding to the expression intensity of a gene is assigned to each sample, and a nonparametric test is performed based on the rank data. Thus, the analysis is performed while normalizing the expression intensity value and reducing the influence of experimental error.

JP 2004-325419 A JP 2005-38279 A Japanese Patent Laid-Open No. 2003-21634 Yang et. Al, 2002, Nucleic Acids Research, Vol. 30, No. 4 (Nucleic Acids Research, 2002, Vol. 30, No. 4)

  The first problem in the prior art is that the bias in the distribution of gene expression data must be corrected for each sample comparison. The reason is that the distribution of gene expression data in each sample is different.

  The second problem is that the above-mentioned Patent Document 3 uses the rank information without considering the absolute value of the gene expression data. Microarray data is an index of gene expression as mRNA production, but its absolute value is assumed to have a different range of variation for each gene, and it is appropriate to apply a common scale. Absent. Replacing with ranking assumes that all compared values have a common scale.

  An object of the present invention is to analyze an absolute value of gene expression data without using rank data and correcting distribution of gene expression data between samples in analysis of gene expression data among a plurality of samples.・ To provide an analysis method.

  Referring to FIG. 1, a gene expression data analysis system according to the present invention includes an input device 1 that inputs a plurality of samples of gene expression data (gene expression data set), a data analysis device 2 that operates under program control, and an output device. 3 is included. The data analysis device 2 includes an order statistic calculation unit 21 for calculating an order statistic of a gene expression data set for a certain sample from N gene expression data sets given from the input device 1; Rank data generation means 22 for ranking the gene expression data for each (N-1) samples and generating a rank data set, calculated order statistics, and (N-1) rank data sets Is used to convert the gene expression data set, and the gene expression estimation means 24 is used to estimate the gene expression status using the converted gene expression data set. The analysis result is output to the output device 3.

  The gene expression data analysis method according to the present invention includes a step of inputting a plurality of samples of gene expression data (gene expression data set), a step of analyzing the gene expression data, and a step of outputting the analysis result. The data analysis step includes calculating an order statistic of the gene expression data set for one sample from the N gene expression data sets input in the input step, and (N-1) pieces. Assigning a ranking to the gene expression data for each of the samples, generating (N-1) ranking data sets, calculating the calculated order statistics and the (N-1) ranking data sets. And converting the gene expression data set and using the converted (N-1) gene expression data sets Characterized by the step of estimating the state of expression.

  According to the present invention, the gene expression data is analyzed without correcting the bias of the distribution of the gene expression data every time the samples are compared because the processing for replacing the gene expression data is performed using the order statistics of the gene expression data. can do. As a result, the calculation time spent for correcting the distribution in the analysis of gene expression data can be reduced, and the throughput can be improved.

  Furthermore, according to the present invention, instead of using the rank data of gene expression data, the order statistics of gene expression data are used, so that gene expression data can be analyzed using the absolute amount of gene expression data. it can. As a result, it is possible to perform data analysis on genes having various fluctuation ranges of expression intensity in consideration of the unique fluctuation areas.

  A first embodiment of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, the first embodiment of the present invention includes an input device 1 that inputs a gene expression intensity data set, a data analysis device 2 that operates under program control, and an output device such as a display device or a printing device. 3 is included. The data analysis apparatus includes order statistics calculation means 21, rank data generation means 22, data conversion means 23, and gene expression estimation means 24.

  The order statistic calculation means 21 calculates the order statistic of the gene expression data set for one sample given from the input device 1, and sends the calculated order statistic to the data conversion means 23.

  The rank data generation means 22 assigns the rank of gene expression level for each sample to the gene expression data set for (N-1) samples given from the input device, and (N-1) ranks. A data set is generated, and the generated rank data set is sent to the data conversion means 23.

  The data conversion means 23 uses the order statistic sent from the order statistic calculation means 21 and the (N-1) rank data sets given from the rank data generation means 22 to generate a gene expression data set. The converted gene expression data set is sent to the gene expression estimating means 24.

  When the gene expression estimation means 24 is referred to the above-mentioned patent document 1, a data analysis device (FIG. 2 of the above-mentioned patent document 1) that estimates the expression status of the gene using a mathematical model estimates the gene expression status. The posterior probability of the expression state for each gene is sent to the output device 3.

  Next, the operation of the present embodiment will be described in detail with reference to FIGS.

とし、

とする。ここで、

をｋ番目の遺伝子発現データセットにおけるｉ番目の遺伝子の発現量とする。 Now, N gene expression data sets

age,

And here,

Is the expression level of the i-th gene in the k-th gene expression data set.

を計算する（図２のステップB２）。ここで、

である。 Among the N gene expression data sets input from the input device 1, one gene expression data set C _k (for example, a sample of interest) is sent to the order statistic calculation means 21. The order statistic calculation means 21 performs the order statistic on the sent gene expression intensity data set C _k .

Is calculated (step B2 in FIG. 2). here,

It is.

に対して、各々の遺伝子発現データセットごとに遺伝子発現量に順位をふり、順位データセット

を生成する（ステップＢ３）。 The rank data generation means 22 includes (N-1) gene expression data sets other than the gene expression data set C _k used for the order statistics.

For each gene expression data set, rank the gene expression level,

Is generated (step B3).

とし、

をｊ番目の遺伝子発現データセットにおけるｉ番目の遺伝子の発現量の順位とする。 However,

age,

Is the rank of the expression level of the i-th gene in the j-th gene expression data set.

と順序統計量

を用いて、（Ｎ−１）個の遺伝子発現データセットを以下の数５式に従って変換し、（Ｎ−１）個の変換された遺伝子発現データセット

を生成する。 The data conversion means 23 has (N-1) ranking data sets.

And order statistics

(N-1) gene expression data sets are converted according to the following equation (5), and (N-1) converted gene expression data sets:

Is generated.

とし、

をｊ番目の遺伝子発現データセットにおけるｉ番目の遺伝子の発現量をｋ番目の遺伝子発現データセットの順序統計量によって置き換えた値である。

However,

age,

Is a value obtained by replacing the expression level of the i-th gene in the j-th gene expression data set by the order statistics of the k-th gene expression data set.

here,

Is

The m th order statistic at

. That is, in the data conversion unit 23, (N-1) pieces of the gene expression data set on the basis of the order of the gene expression level, replaced with the gene expression data of a gene expression data set C _k (step B4).

The gene expression estimation means 24 includes (N-1) gene expression data sets converted by the data conversion means 23.

And C _k are used to calculate the posterior probability of the gene expression state. Here, the estimation of gene expression status is

And _Ck for all combinations (N-1 times).

Now

And C _k , the posterior probability of the expression state of each gene in sample k

And Similarly, the posterior probability of the expression state is obtained for all combinations (step B5).

Next, the gene expression estimation means 24 uses the posterior probabilities of p × (N−1) expression states for the sample k, for example, to obtain the average of the posterior probabilities for all genes by the following equation (6) ( Step B6).

here,

Let denote an average operation on a vector x such as arithmetic average or logit average.

  Next, the effect of this embodiment will be described. In the present embodiment, the data of the remaining (N-1) gene expression data sets are replaced based on the rank information by the order statistics of the gene expression data set of a specific sample. Thereby, the distribution of the gene expression data among the respective samples becomes equal, and the gene expression data can be analyzed without correcting the distribution of the gene expression data for each comparison between the samples. Furthermore, by repeating the comparison with a plurality of samples, the posterior probability of the expression state of each gene in sample k can be accurately estimated.

  Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

The second embodiment of the present invention is that the gene expression data is used to visualize the fluctuation of the expression level between samples by a graph. Referring to FIG. 3, the configuration of the second embodiment is that expression variation graphing means 25 is used instead of the gene expression estimating means 24 in the configuration of the first embodiment shown in FIG. Different from the configuration of the first embodiment.

  Next, the operation of the present embodiment will be described in detail with reference to FIGS.

In the operation of the present embodiment, steps B1 to B4 in the flowchart of the first embodiment shown in FIG. 2 are common to steps C1 to C4 in FIG.

After step C4, the expression variation graphing means 25 converts the converted (N-1) gene expression data.

And C _k are used to generate a graph of the variation in the expression level of gene j when the sample is changed, and send it to the output device 3 (step C5). The output device 3 displays or outputs a graph of expression fluctuation (step C6).

  Next, the effect of this embodiment will be described. In the present embodiment, changes in the expression level of gene i are displayed in a graph using the order statistics of gene expression data. Thereby, the absolute value of gene expression data can be graphed without correcting between samples.

  In addition, when examining changes in gene expression levels while changing the dose of a drug to a cell line, or examining time-series changes in expression levels, reference gene expression data (for example, before administration or at time 0) The remaining gene expression data is replaced on the basis of the rank information by the order statistics of the gene expression data). Then, using the replaced gene expression data, the change in the expression level of each gene when the dose or time is changed is graphed. The effect in this example is that the change in the expression level of the gene can be examined without correcting the distribution of the expression level of the gene data in each sample and without using the rank data.

  Next, a third embodiment of the present invention will be described in detail with reference to the drawings. In the present embodiment, statistical testing is performed using genetic data. Referring to FIG. 5, the configuration of the third embodiment is that statistical test means 26 is used instead of the gene expression estimation means 24 in the configuration of the first embodiment shown in FIG. Different from the configuration of the first embodiment.

  Next, the operation of the present embodiment will be described in detail with reference to FIGS.

  In the operation of the present embodiment, steps B1 to B4 in the flowchart of the first embodiment shown in FIG. 2 are common to steps D1 to D4 in FIG.

After step D5, the statistical test means 26 converts the converted N-1 gene expression data.

And with respect to C _k, using a statistical test method such as t- test and U test extracts genes significantly a difference in the expression level of the gene between groups, sends to the output device 3 (step D5).

  Next, the effect of this embodiment will be described. In the present embodiment, statistical tests are performed using order statistics of gene expression data. Thereby, a statistical test can be performed without correcting the distribution of gene expression data between samples.

  Next, a fourth embodiment of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 7, the fourth embodiment of the present invention includes an input device, a data analysis device, and an output device as in the first embodiment of the present invention. A recording medium 4 on which a computer program for execution is recorded is provided. The recording medium 4 may be either a portable type or a fixed type, and may be a magnetic disk, a semiconductor memory, a CD-ROM or other recording medium. In addition, a computer program for executing this data analysis method can be stored in a recording device of a computer connected to a network, and transferred to another computer via the network. The providing medium for providing the computer program for executing the present algorithm can be distributed as a computer-readable medium in various formats, and is not limited to a specific type of medium.

  The data analysis program is read from the recording medium 4 into the data analysis device 5, controls the operation of the data analysis device 5, and the data file input from the input device 1 is processed by the data processing device 2 in the first embodiment. The same process as the process is executed.

  Hereinafter, specific examples of the present invention will be specifically described with reference to results of actual data. Such an example corresponds to the first embodiment described above.

  The microarray data used here is data obtained by an oligo type microarray, and one gene expression data set is composed of 19958 gene expression data. In this example, for simplicity, two gene expression data sets are used, which are referred to as sample A and sample B, respectively.

FIG. 8 shows the results of plotting the gene expression data of Sample A and Sample B. The horizontal axis of FIG. 8 represents the sum of logarithmic values of gene expression intensities in two samples, and the vertical axis represents the difference in logarithmic values of gene expression intensities in two samples (referred to as SD plot). The curve in the SD plot is, for example, a nonparametric smoothing curve expressed by the following equation (7).

  Here, u represents the sum of logarithmic values of gene expression in two samples (horizontal axis in SD plot), and v represents the difference in logarithmic values of gene expression intensities in two samples (vertical axis in SD plot). Shall.

  As can be seen from the shape of the scatter plot and the nonparametric smoothing curve, the distribution of the gene expression data in the two samples is biased. Usually, it is necessary to correct such a bias in the distribution of gene expression data using a nonparametric smoothing curve expressed by the above equation (7).

  FIG. 9 shows an SD plot after the sample B gene expression data is replaced by the order statistics of the sample A gene expression data. Note that the curve in the SD plot is a nonparametric smoothing curve represented by the above equation (7). Compared to FIG. 8, it can be seen that there is no bias in the distribution of gene expression data.

  FIG. 10 shows a result estimated by estimating the gene expression status in Patent Document 2 after correcting the distribution of gene expression data for sample A and sample B using a nonparametric smoothing curve.

  The horizontal axis represents the sum of logarithmic values of gene expression intensities of sample A and sample B, and the vertical axis represents the difference between them. The shapes of five types of plots (1) to (5) are shown in the lower part of FIG. The area of the white circle plot in (1) represents a point where the probability of heterogeneity between the expression states of sample A and sample B is 0 to 0.2. The x-plot region in (2) represents a point where the heterogeneity probability of the expression state of sample A and sample B is 0.2 to 0.4. The region of the triangular plot of (3) represents a point where the probability of heterogeneity in the expression state of sample A and sample B is 0.4 to 0.6. The area of the square plot in (4) represents the point where the heterogeneity probability of the expression state of sample A and sample B is 0.6 to 0.8. The area of the black circle plot in (5) represents a point where the probability of heterogeneity in the expression state of sample A and sample B is 0.8 to 1.0.

  FIG. 11 shows the result of estimating the gene expression status with respect to FIG. The meaning content of the shape of the plot in FIG. 11 is the same as the shape of the plot in FIG.

  Comparing FIG. 10 and FIG. 11, the estimation result of the gene expression state for the gene expression data corrected by the nonparametric smoothing curve and the estimation result of the gene expression state for the gene expression data replaced by the order statistic are almost equal. It can be seen that the same results are shown.

It is the block diagram which showed the structure of the analysis system which concerns on the 1st Embodiment of this invention. It is a processing flowchart of the analysis system of FIG. It is the block diagram which showed the structure of the analysis system which concerns on the 2nd Embodiment of this invention. FIG. 4 is a process flow diagram of the analysis system of FIG. 3. It is the block diagram which showed the structure of the analysis system which concerns on the 3rd Embodiment of this invention. FIG. 6 is a process flow diagram of the analysis system of FIG. 5. It is the block diagram which showed the structure of the analysis system which concerns on the 4th Embodiment of this invention. It is SD plot by the sample A and the sample B. FIG. It is an SD plot in which the gene expression data of sample B is replaced by the order statistics of sample A. It is the figure which showed the estimation result of the gene expression condition after correction | amendment by nonparametric smoothing. It is the figure which showed the estimation result of the gene expression condition after replacing with the order statistic of the sample A.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Input device 2 Data analysis device 3 Output device 4 Recording medium 5 Data analysis device 21 Order statistic calculation means 22 Rank data generation means 23 Data conversion means 24 Gene expression estimation means 25 Expression variation graphing means 26 Statistical test means

Claims (24)

In a gene expression data analysis system including an input device that inputs gene expression data (gene expression data set) of a plurality of samples, a data analysis device that operates under program control, and an output device that outputs analysis results,
The data analysis apparatus comprises: an order statistic calculating means for calculating an order statistic of a gene expression data set for a sample from N gene expression data sets given from the input device; 1) rank the gene expression data for each sample, rank data generation means for generating (N-1) rank data sets, the calculated order statistics, and the generated (N -1) Data conversion means for converting gene expression data sets using the number of rank data sets, and gene expression for estimating gene expression status using the converted (N-1) gene expression data sets An analysis system for gene expression data of a plurality of samples, comprising an estimation means. N gene expression data sets

age,

age,

Is the expression level of the i-th gene in the k-th gene expression data set,
The order statistic calculation means 21 calculates the order statistic with respect to any one of the N gene expression data sets C _k input from the input device.

(here,

It is. The system for analyzing gene expression data of a plurality of samples according to claim 1. The rank data generating means includes (N-1) gene expression data sets other than the gene expression data set C _k used for the order statistics.

For each gene expression data set, rank the gene expression level,

(However,

age,

The analysis system for gene expression data of a plurality of samples according to claim 2, wherein the expression level of the i-th gene in the j-th gene expression data set is a ranking. The data conversion means includes (N-1) rank data sets.

And order statistics

(N-1) gene expression data sets are expressed by the following equation 1 (where,

Is

The m th order statistic at

. ), And (N-1) transformed gene expression data sets

(However,

age,

The expression value of a plurality of samples according to claim 3, wherein the expression level of the i-th gene in the j-th gene expression data set is replaced by the order statistic of the k-th gene expression data set) Analysis system for data.

The gene expression estimation means includes (N-1) gene expression data sets converted by the data conversion means.

Analysis system for the C _k with, the plurality of samples according to claim 4, wherein calculating the posterior probability of gene expression state gene expression data. The estimation of gene expression status is

_6. The analysis system for gene expression data of a plurality of samples according to claim 5, wherein the analysis is performed for all combinations (N-1 times) of Ck and Ck. The gene expression estimation means uses the posterior probabilities of p × (N−1) expression states for sample k,

Represents the average operation on the vector x. ) To find the average of posterior probabilities for all genes (

And C _k , the posterior probability of the expression state of each gene in sample k

And The analysis system for gene expression data of a plurality of samples according to claim 6.

In a gene expression data analysis system including an input device that inputs gene expression data (gene expression data set) of a plurality of samples, a data analysis device that operates under program control, and an output device that outputs analysis results,
The data analysis apparatus comprises: an order statistic calculating means for calculating an order statistic of a gene expression data set for a sample from N gene expression data sets given from the input device; 1) A gene expression data set is converted using rank data generation means for ranking the gene expression data for each sample, calculated order statistics, and (N-1) rank data sets. And a data conversion means and an expression fluctuation graphing means for generating an expression fluctuation graph of gene j when the sample is changed using the converted (N-1) gene expression data sets. Analysis system for gene expression data of multiple samples. The expression variation graphing means includes (N-1) converted gene expression data.

And with C _k, to generate the expression level variation graph of gene j when changing the sample, multiple samples of gene expression according to claim 8, wherein the sending the graph image data to the output device Analysis system for data. In a gene expression data analysis system including an input device that inputs gene expression data (gene expression data set) of a plurality of samples, a data analysis device that operates under program control, and an output device that outputs analysis results,
The data analysis apparatus comprises: an order statistic calculating means for calculating an order statistic of a gene expression data set for a sample from N gene expression data sets given from the input device; 1) A gene expression data set is converted using rank data generation means for ranking the gene expression data for each sample, calculated order statistics, and (N-1) rank data sets. Data conversion means and the converted (N-1) gene expression data

And C _k using a predetermined statistical test method, a gene having a significant difference in gene expression level between groups is extracted, and a statistical test means for sending the data to the output device is provided. An analysis system for gene expression data of multiple samples.   The analysis system for gene expression data of a plurality of samples according to claim 10, wherein the statistical test method is a t-test or a U test. In a method for analyzing gene expression data, comprising a step of inputting gene expression data of multiple samples (gene expression data set), a step of analyzing the gene expression data, and a step of outputting the analysis result,
The data analysis step includes
Calculating an order statistic of the gene expression data set for one sample from the N gene expression data sets input in the input step; and (N-1) gene expression data for each sample. And (N-1) ranking data sets, and using the calculated order statistics and the (N-1) ranking data sets, a gene expression data set And a method of estimating gene expression status using the converted (N-1) gene expression data sets, and a method for analyzing gene expression data of a plurality of samples. The order statistic calculation step includes N gene expression data sets.

age,

age,

Is the expression level of the i-th gene in the k-th gene expression data set, with respect to any one gene expression data set C _{k of} the N gene expression data sets input in the input step. , Order statistics

(here,

It is. The method for analyzing gene expression data of a plurality of samples according to claim 12, further comprising: The rank data generation step includes (N−1) gene expression data sets other than the gene expression data set C _k used for the order statistics.

For each gene expression data set, rank the gene expression level,

Step to generate (however,

age,

The analysis method for gene expression data of a plurality of samples according to claim 13, characterized in that the expression level of the i-th gene in the j-th gene expression data set is included. The data conversion step includes (N-1) ranking data sets.

And order statistics

(N-1) gene expression data sets using the following equation (3)

Is

The m th order statistic at

. ), And (N-1) transformed gene expression data sets

Step to generate (however,

age,

The value obtained by replacing the expression level of the i-th gene in the j-th gene expression data set with the order statistic of the k-th gene expression data set). Analysis method for gene expression data.

The gene expression estimation step includes (N-1) gene expression data sets converted in the data conversion step.

And with C _k, 15. analysis method for multiple sample gene expression data, wherein further comprising the step of calculating a posterior probability of gene expression status. The estimation of gene expression status is

And all combinations analysis method on the gene expression data for a plurality of samples as claimed in claim 16 wherein the possible wherein performed on (N-1 times) of C _k. The gene expression estimation step uses the posterior probability of p × (N−1) expression states for sample k,

Represents the average operation on the vector x. ) To find the average of posterior probabilities for all genes (

And C _k , the posterior probability of the expression state of each gene in sample k

And 18. The method for analyzing gene expression data of a plurality of samples according to claim 17, further comprising:

In a method for analyzing gene expression data, comprising a step of inputting gene expression data of multiple samples (gene expression data set), a step of analyzing the gene expression data, and a step of outputting the analysis result,
The data analysis step includes
Calculating an order statistic of a gene expression data set for a sample from N gene expression data sets given from the input device;
(N-1) assigning ranks to the gene expression data for each of the samples, and generating (N-1) rank data sets;
Transforming a gene expression data set using the calculated order statistics and the (N-1) rank data sets;
Analysis of gene expression data of a plurality of samples, comprising a step of generating a graph of expression variation of gene j when the sample is changed using the converted (N-1) gene expression data sets Method. The expression variation graphing step includes (N-1) transformed gene expression data.

And using C _k, to generate the expression level variation graph of gene j when changing the sample, multiple samples of gene expression according to claim 19, wherein sending the graph image data to the output device Analysis method for data. In a method for analyzing gene expression data, comprising a step of inputting gene expression data of multiple samples (gene expression data set), a step of analyzing the gene expression data, and a step of outputting the analysis result,
The data analysis step includes
Calculating an order statistic of a gene expression data set for a sample from N gene expression data sets given from the input device;
(N-1) assigning ranks to the gene expression data for each of the samples, and generating (N-1) rank data sets;
Transforming a gene expression data set using the calculated order statistics and the (N-1) rank data sets;
The converted (N-1) gene expression data

And C _k using a predetermined statistical test method, a gene having a significant difference in gene expression level between groups is extracted, and a statistical test step of sending the data to the output device is provided. Analysis method for gene expression data of multiple samples.   The analysis method for gene expression data of a plurality of samples according to claim 21, wherein the statistical test method is a t-test or a U test. On the computer,
Inputting gene expression data of multiple samples (gene expression data set);
Calculating an order statistic of the gene expression data set for one sample from the N gene expression data sets input in the input step;
(N-1) assigning ranks to the gene expression data for each of the samples, and generating (N-1) rank data sets;
Transforming a gene expression data set using the calculated order statistics and the (N-1) rank data sets;
Estimating the expression status of genes using the transformed (N-1) gene expression data sets;
A gene expression data analysis program for executing the step of outputting analysis results. A computer-readable information recording medium (including a compact disk, a flexible disk, a hard disk, a magneto-optical disk, a digital video disk, a magnetic tape, or a semiconductor memory), wherein the program according to claim 23 is recorded. .

Images