KR20190126606A - IDENTIFYING METHOD FOR TUMOR PATIENT BASED ON miRNA IN EXOSOME AND APPARATUS FOR THE SAME - Google Patents

Abstract

The present invention relates to a method for discriminating a cancer patient based on exosome miRNA, which comprises the following steps: receiving, by a computer device, target expression data of a miRNA set isolated from the target exosome; reducing, by the computer device, a dimension of the target expression data; and determining, by the computer device, the degree of similarity between the dimensionally-reduced target expression data and previously prepared reference expression data. The reference expression data is data for a miRNA set isolated from at least one exosome of a patient and a normal person.

Description

Method and apparatus for identifying cancer patients based on exosome miRNA {IDENTIFYING METHOD FOR TUMOR PATIENT BASED ON miRNA IN EXOSOME AND APPARATUS FOR THE SAME}

The technology described below relates to a technique for determining whether a cancer patient using exosome miRNA expression data.

Recently, liquid biopsy, which obtains genetic information related to tumors in the blood, which can be obtained repeatedly instead of biopsy of cancer tissues, is drawing attention. In particular, there is a great interest in non-coding RNA (functional noncoding RNA) that performs a function in the RNA state, such as microRNA (miRNA). In particular, research on exosomes made in cancer cells and secreted into the serum is active.

International Publication WO2015-190586 (2015.12.17.)

The miRNA of exosomes is also studied as a marker for diseases such as tumors. However, the conventional research is only to find a specific miRNA as a marker for a specific disease. Techniques described below are intended to provide a technique for identifying cancer on the basis of the expression data for the entire miRNA of the exosomes.

A method for determining a cancer patient based on an exosome miRNA includes the steps of a computer device receiving target expression data of a miRNA set separated from a target exosome, the computer device reducing the dimension of the target expression data, and the computer. And determining, by the device, the similarity between the reduced target expression data and previously prepared reference expression data. The reference expression data is data for a miRNA set isolated from at least one exosome of a patient and a normal person.

An apparatus for determining a cancer patient based on an exosome miRNA stores an input device for receiving target expression data of a miRNA set separated from a exosome of a target, and stores reference expression data prepared in advance, and the target expression data and the reference expression. A storage device for storing a program for determining the similarity of data, and a computing device for reducing the dimension of the target expression data using the program, and calculates the similarity. The reference expression data is data for a miRNA set isolated from at least one exosome of a patient and a normal person, and the miRNA set includes all miRNAs isolated from the exosomes.

The technique described below simply analyzes the entirety of the exosome miRNA to determine whether it is onset without isolating a particular miRNA. The technology described below provides easy criteria for identifying based on similarity with reference data.

1 is a filtering example for a sample exosome miRNA.
2 is an example of a process of processing sample exosome miRNA expression data.
3 is an example hierarchical clustering of sample exosome miRNA expression data.
4 is an example of performing multidimensional scaling on sample exosome miRNA expression data.
5 is an example of a flowchart for a method of determining cancer patients based on exosome miRNA data.
6 is an example of performing multidimensional scaling on a target miRNA set.
7 is an example of a computer device for discriminating cancer patients based on exosome miRNA data.

The following description may be made in various ways and have a variety of embodiments, specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the technology described below to specific embodiments, it should be understood to include all changes, equivalents, and substitutes included in the spirit and scope of the technology described below.

The terms first, second, A, B, etc. may be used to describe various components, but the components are not limited by the terms, but merely for distinguishing one component from other components. Used only as For example, the first component may be referred to as the second component, and similarly, the second component may be referred to as the first component without departing from the scope of the technology described below. The term and / or includes a combination of a plurality of related items or any item of a plurality of related items.

As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It is to be understood that the present invention means that there is a part or a combination thereof, and does not exclude the presence or addition possibility of one or more other features or numbers, step operation components, parts or combinations thereof.

Prior to the detailed description of the drawings, it is to be clear that the division of the components in the present specification is only divided by the main function of each component. That is, two or more components to be described below may be combined into one component, or one component may be provided divided into two or more for each function. Each of the components to be described below may additionally perform some or all of the functions of other components in addition to the main functions of the components, and some of the main functions of each of the components are different. Of course, it may be carried out exclusively by.

In addition, in carrying out the method or operation method, each process constituting the method may occur differently from the stated order unless the context clearly indicates a specific order. That is, each process may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

Exosomes correspond to nanovesicles secreted by eukaryotic organisms to exchange information between cells. Exosomes are a type of vesicles (extracellular vesicles) that cells release outside the cell. Exosomes are vesicular particles having a diameter of about 20 to 100nm. Exosomes also serve to transport integrins, MHC molecules, cytoskeleton proteins and the like. As such, exosomes have been studied as biomarkers containing various types of proteins.

miRNAs are small, uncoded sequences that are expressed in animals, plants, and viruses. They are RNA sequences consisting of about 22 nucleotides. miRNA also plays a role in directly controlling gene expression in eukaryotes by inhibiting the translation of mRNA. miRNAs are well conserved in both plants and animals and have been found to control many mRAN mechanisms.

Exosomes and microvesicles can deliver mRNA and miRNA from the originating cell (donor cell) to the recipient cell, thereby affecting the function of the recipient cell through protein expression or protein expression suppression. Exosomes and microvesicles containing mRNA and miRNA can enable intercellular communication, regardless of the distance between them. MRNAs and miRNAs associated with these exosomes and microvesicles also exhibit disease-specific onset factors. The miRNA contained in exosomes can act as a biomarker in metabolic diseases of mRNA and miRNA.

Techniques described below are techniques for predicting or diagnosing cancer by analyzing sample data. Techniques described below determine the onset of the disease by analyzing the miRNA of the exosomes. The degree of expression of exosomal miRNA is analyzed.

Currently, the next generation sequencing technology (NGS) is used to confirm the expression level of RNA. NGS decodes nucleotide sequences by generating reads, which are usually short sequence fragments of about 100 bases. NGS stores the decoded sequences in a file in FASTQ format. This is usually called raw data.

The length of the NGS read is about 100 bp, which is shorter than the 500-1,000 bp of the conventional Sanger type, the sequencing error is relatively large, and may include platform-dependent errors. The FASTQ file generated by the NGS platforms includes the accuracy of a decoded base (quality score or error rate) in the FASTA format, a text-based standard base data format that represents an existing DNA sequence.

First, the experiment that led to the following description will be described. 1 to 4 relate to the experiment. The experiment was conducted on 89 samples. The sample consisted of 33 normal samples (Control) and 56 samples of colorectal cancer patients (Test).

It is assumed that a process of separating exosomes from a sample (serum, etc.) and a process of separating (identifying) miRNAs from exosomes. Exosomes and miRNA isolation can utilize a variety of techniques, and can also be separated using a commercial kit. A total of 2,588 mature miRNAs were isolated from exosomes.

Expression data for miRNAs of sample exosomes were used. The expression data for the entire miRNA was used, not the specific miRNA of the sample exosomes. Expression data for miRNA can be derived through NGS analysis. The results of NGS analysis are in the form of computer readable files. Therefore, the processing of miRNA expression data is performed in a computer device.

Expression data can be expressed by various criteria. The analysis may be performed based on RPKM (Reads per kilo base per million mapped reads) and RPM (Reads per million mapped reads). RPM is mainly used to normalize miRNA expression. RPM, unlike RPKM, does not consider gene length. This is because the size of miRNAs is generally between 20-25.

First you can preprocess the raw data.

Preprocessing may include filtering data that did not help with the analysis. The computer device may remove some miRNAs from the mature miRNAs. Find the miRNA with zero RPM in the whole sample. The computer device may exclude miRNAs having an RPM of zero or more of the reference samples of the total samples. For example, miRNAs having an RPM of at least 50% of the total samples may be excluded from the analysis.

1 is a filtering example for a sample exosome miRNA. 1 is an example of filtering miRNAs having an RPM of at least 50% of the total samples. 2,390 of 2,588 mature miRNAs contained in exosomes were filtered out. After that, 198 mature miRNAs are analyzed. 1 shows the miRNA from which a portion indicated in red is removed. Of course, unlike FIG. 1, the RPM may not be based on a value of 0, and samples having a RPM below a specific reference value may be removed.

In general, miRNAs with low expression values are associated with low disease. Therefore, filtering may be advantageous for rapid analysis. However, the filtering process is not essential. In some cases, miRNA patterns that are not expressed may also be useful for specific disease analysis.

Expression data of a sample miRNA includes respective expression values for a plurality of miRNAs. The computer device may process to reduce the dimension of the expression data. In multidimensional data analysis, dimension reduction techniques are diverse. Therefore, one of the various dimension reduction techniques can be used in this process. The computer device may also perform normalization on the expression data.

2 is an example of a process of processing sample exosome miRNA expression data. 2 is an example of the dimension reduction and normalization process for the sample exosome miRNA expression data.

Figure 2 (A) is an example showing the RPM for the expression data of the sample exosome miRNA. 2 (A) is an example for some expression data. 2 (A) is an example of raw data. Figure 2 (A) may be an example for the expression data of the sample exosome miRNA filtered in FIG.

The computer device can reduce the dimension in the expression data of the sample exosome miRNA. For example, data processing may be performed using Equation 1 below. Equation 1 reduces the dimension by taking the log (log ₂ ) in the RPM. The dimension reduced value is displayed as RPM _log . If the raw data values are distributed over a wide range and the distribution is skewed to a relatively low value, taking a log can reduce the range of data values and make the data evenly distributed throughout, facilitating data validation.

RPM on the right side of Equation 1 is the initial value of the expression data of exosome miRNA. Equation 1 is the logarithm of the RPM plus one. This is because there may be data with zero RPM.

Figure 2 (B) is an example showing the results of taking a log in the miRNA expression data of Figure 2 (A). In FIG. 2B, the graph visually representing the value of the expression data is in the form of a bar having a range of minimum and maximum values. The black solid line on the bar corresponds to the median (average) of the sample.

Further, the computer device may perform normalization on expression data of a sample exosome miRNA. NGS analysis results may have different values depending on the analyte sample or the system environment, so that regular normalization of expression data is desirable. Computer devices may process data using various normalization techniques. For example, the computer device may perform quantile normalization. Describing one of the displacement value normalizations, the computer device sorts the data in a certain criterion (ascending or descending order) in different data sets and changes the values of the data in the same order in the sorted data sets to be the same. For example, the average of two values can change the value of the data. This allows computer devices to regularly normalize data sets that are different from each other. In some cases, the computer device may perform subsequent analysis based only on the values of the top few percent. FIG. 2 (C) shows an example in which displacement value normalization is performed on the data of FIG. 2B. Referring to FIG. 2C, it can be seen that the median value of the sample data is kept constant.

The computer device now performs an analysis on the preprocessed data. The computer device performs analysis on the data including information about the plurality of objects.

The computer device may perform so-called cluster analysis. Cluster analysis is a technique that classifies high-parity individuals into clusters based on their similarity without prior information on categories. The computer device may classify sample exosome miRNA expression data through hierarchical clustering.

Hierarchical clustering measures similarity between individuals, classifying the highly similar individuals into one group. Various algorithms exist for hierarchical clustering. Hierarchical clustering is briefly described. Hierarchical clustering starts with one cluster and divides the objects belonging to the cluster into a plurality of groups based on similarity.

The similarity of an entity is defined by its distance. Distance is a measure of dissimilarity between individuals. Computer devices may use various distances. For example, Euclidean street, Manhattan street, Pearson street, etc. can be used. Euclidean distance is the shortest distance between two entities. Pearson distance is a constant normalized by dividing distance by the variance of the variable.

The similarity between a group and an entity is defined as linkage. The computer device may use various connections. For example, a single connection, a full connection, a center connection, an average connection can be used. A single connection is the distance of the closest of the two clusters. Fully connected means the distance of the two groups that are farthest from each other. The central link is the distance between the mean values of the clusters. The average linkage means the average of each distance between one clustered entity and another clustered entity.

3 is an example hierarchical clustering of sample exosome miRNA expression data. 3 shows an example of performing hierarchical clustering using Euclidean distance and full connectivity. 3 is an example of performing hierarchical clustering based on the aforementioned RPM _log value. In FIG. 3, the normal sample (Control) is shown in red, and the patient sample (Test) is shown in blue. Referring to the results of FIG. 3, it can be seen that patient samples and normal samples are classified into different groups. The computer device stores the hierarchical clustering results for the sample exosome miRNA expression data. The saved hierarchical clustering result is called a reference cluster. The computer device may then analyze the newly input data based on the reference cluster. That is, by analyzing the exosome miRNA expression data, it can be determined whether a particular sample is a patient or normal.

The computer device may perform another analysis on sample exosome miRNA expression data. For example, the computer device may perform Multi-Dimensional Scaling (MDS). Multidimensional scaling is a data abbreviation multivariate analysis technique that seeks to implicitly represent data by finding structures inherent in cluster analysis. In multidimensional scaling, the criteria that can be used to evaluate similarity or dissimilarity between objects are identified, and each object is visually represented in multidimensional space for each criterion. Multidimensional scaling also has a variety of techniques. For example, traditional multidimensional scaling, metric MDS, non-metric MSD and the like.

The multidimensional scaling method calculates similarity or dissimilarity based on a specific value of an object and expresses it as a point in a multidimensional (two or three dimensional) space. The computer device may determine the similarity or dissimilarity between the samples based on the RPM _log values described above. Dissimilarity can be defined based on the aforementioned distance (eg, Euclidean distance). The multidimensional scaling method can calculate the goodness of fit for the distance between the determined objects. Usually stress measures are used. The computer device may repeat the optimization process of adjusting the position of the object by obtaining new coordinates that may reduce stress. Since there are various well-known algorithms for the multidimensional scaling method, a detailed description thereof will be omitted.

4 is an example of performing multidimensional scaling on sample exosome miRNA expression data. 4 is an example of performing a multi-dimensional scaling method based on the above RPM _log value. In FIG. 4, the degree of similarity between samples is represented by coordinates in two-dimensional space using two components (component 1 and component 2) which best describe the dispersibility of the entire data. The element here is a dimension to represent dispersibility. In FIG. 4, the normal sample (Control) is shown in red, and the patient sample (Test) is shown in blue. Referring to FIG. 4, it can be seen that the patient sample and the normal sample have different patterns.

The computer device stores multidimensional scaling results for sample exosome miRNA expression data. Saved multidimensional scaling results are called reference patterns. Thereafter, the computer device may analyze data newly input based on the reference pattern. That is, by analyzing the exosome miRNA expression data, it can be determined whether a particular sample is a patient or normal.

The above description is a process and an analysis result of analyzing data for colorectal cancer patients. However, the analysis according to the above process is not necessarily limited to colon cancer. On the basis of the expression level of the exo-some miRNA may determine whether the specific disease.

5 is an example of a flowchart for a method 100 for determining a cancer patient based on exosome miRNA data. The method 100 for determining a cancer patient is performed in a computer device. The computer device first analyzes sample exosome miRNA data to prepare reference data (110). Meanwhile, the sample exosome miRNA data is called sample data. As described above, miRNA expression data of samples exosomes are analyzed. The computer apparatus may perform hierarchical clustering based on similarity or dissimilarity of the sample to prepare a reference cluster through the above-described process. In addition, the computer device may prepare a reference pattern using a multidimensional scaling method based on the similarity or dissimilarity of the sample. The reference data may include at least one of a reference cluster and a reference pattern. Reference data may include an identifier as to whether a particular sample is patient or normal.

The computer device now obtains data about the target sample to be analyzed. The data for the target sample is called target data. Target data includes expression data for miRNAs throughout the exosomes of the target. Thus, the target data includes data for a plurality of miRNAs. Multiple miRNAs are called target miRNA sets.

The computer device preprocesses 120 expression data (target data) of the target exosome miRNA. The computer device may perform preprocessing, such as filtering and normalization, as described above.

The computer device analyzes 130 the similarity between the sample data and the target data. Similarity may be determined based on the RPM _log described in Equation 1. The computer device may determine the similarity (or dissimilarity) based on the Euclidean distance of the sample data and the target data.

The computer device generates 140 a result of the analysis based on the similarity between the sample data and the target data. The computer device may perform hierarchical clustering of the sample data and the target data. Alternatively, the computer device may perform multidimensional scaling on the sample data and the target data.

The computer device may diagnose the target based on the analysis result (150). Meanwhile, this process may be performed by a separate expert using the analysis result generated in step 140 by the computer device. The computer device may perform diagnosis based on the cluster to which the target data belongs in the reference cluster as shown in FIG. 3. For example, when the target data belongs to the cluster shown in blue in FIG. 3, the target may be diagnosed as a cancer patient.

6 is an example of performing multidimensional scaling on a target miRNA set. 6 corresponds to a result of the computer apparatus performing a multidimensional scaling method using sample data and target data. If the analysis indicates that the location of the target miRNA matches or is close to the location of the patient sample, the computer device can diagnose that the target is a cancer patient. Conversely, if the analysis indicates that the location of the target miRNA matches or is closer to the position of the normal sample, the computer device can diagnose that the target is a cancer patient.

In this case, the computer device may compare and analyze the positions of the plurality of patient samples and the positions of the targets based on various criteria. For example, the computer device may make a diagnosis by comparing the distance between the average position of the patient sample and the position of the target or the distance between the average position of the normal sample and the position of the target. The computer device may determine the target as the patient if the location of the target is closer to the average position of the patient.

Alternatively, the computer apparatus may diagnose by comparing a value accumulated in the distance between the positions of the plurality of patient samples and the position of the target and a value accumulated in the distance between the positions of the plurality of normal samples and the position of the target. The computer device may determine that the target is for a target having a smaller cumulative distance from the patient.

In addition, the method for determining a cancer patient based on exosome miRNA data as described above may be implemented as a program (or application) comprising an executable algorithm that can be executed in a computer. The program may be stored and provided in a non-transitory computer readable medium.

The non-transitory readable medium refers to a medium that stores data semi-permanently and is readable by a device, not a medium storing data for a short time such as a register, a cache, a memory, and the like. Specifically, the various applications or programs described above may be stored and provided in a non-transitory readable medium such as a CD, a DVD, a hard disk, a Blu-ray disk, a USB, a memory card, a ROM, or the like.

7 is an example of a computer device for discriminating cancer patients based on exosome miRNA data. 7 is an example of the above-described computer device.

7A illustrates an example in which a computer device 100 such as a PC determines a cancer patient. Computer device 200 receives data for analysis from gene DB 50. Gene DB 50 holds the results of NGS analysis on the sample data and target data described above. Gene DB 50 holds expression data for exosome miRNA. Computer device 200 analyzes the expression data for the sample to establish a reference model. The reference model includes at least one of a reference cluster based on hierarchical clustering described above and a reference pattern based on multidimensional scaling described above. The computer device 200 performs preprocessing on the target data, and performs hierarchical clustering and / or multidimensional scaling based on similarity with the sample data. The computer device 200 may diagnose the target based on the result of the similarity analysis. The detailed procedure is as described above.

7B is an example in which a computer device 300 such as a server on a network determines a cancer patient. Computer device 300 receives data for analysis from client device 80. The client device 80 may be a personal PC, an NGS analysis device, or the like. The client device 80 holds NGS analysis results for the above-described sample data and target data. Client device 80 holds expression data for exosome miRNAs. Computer device 300 analyzes the expression data for the sample to establish a reference model. The reference model includes at least one of a reference cluster based on hierarchical clustering described above and a reference pattern based on multidimensional scaling described above. The computer device 300 performs preprocessing on the target data, and performs hierarchical clustering and / or multidimensional scaling based on similarity with the sample data. The computer device 300 may diagnose the target based on the result of the similarity analysis. The detailed procedure is as described above. The computer device 300 may transmit the analysis result and the diagnosis result to the client device.

7C is an example of a block diagram showing the configuration of the computer device 400. The computer device 400 corresponds to the computer device 200 or the computer device 300 described above. The computer device 400 includes an input device 410, a computing device 420, a storage device 430, and an output device 440.

The input device 410 receives data for analysis. The data received may include NGS analysis results for sample data and target data. The data received includes expression data for exosome miRNAs. The input device 410 may be a physical interface device such as a keyboard, a mouse, or a touch pad. Alternatively, the input device 410 may be a device that receives gene expression data stored from an external storage medium (USB, etc.). Alternatively, the input device 410 may be a communication module for receiving gene expression data from an external network.

Storage device 430 stores a program for diagnosing the onset of cancer by analyzing the expression data of the aforementioned exosome miRNA. The program stored in the storage device 430 may store source code for analyzing expression data of exosome miRNA according to the above description. The storage device 430 may store a reference model prepared by analyzing sample data.

The operation unit 420 performs preprocessing on the target data using the input sample data / target data and the program stored in the storage device 430, and hierarchical clustering and / or multidimensional scaling based on similarity with the sample data. Do this. The operation device 420 may diagnose the target by comparing the result of the similarity analysis with the reference model. The computing device 420 refers to a processor device that processes a specific operation through a program such as a CPU, an application processor (AP), or the like.

The output device 440 is a device for outputting an analysis result for sample data and target data. The output device 440 may be a display device for outputting an image or a printer for outputting text. Furthermore, the output device 440 may be a communication module for transmitting the result of analysis to another device.

The embodiments and the drawings attached to this specification are merely to clearly show a part of the technical idea included in the above-described technology, and those skilled in the art can easily make it within the scope of the technical idea included in the description and the drawings of the above-described technology. It will be apparent that both the inferred modifications and the specific embodiments are included in the scope of the above-described technology.

50: gene DB
80: client device
200: computer device
300: computer device
400: computer device
410: input device
420: computing device
430: storage device
440: output device

Claims (15)

Receiving, by the computer device, target expression data of the miRNA set separated from the target exosome;
The computer device narrowing the dimension of the target expression data; And
Determining, by the computer device, the degree of similarity between the dimensionally reduced target expression data and previously prepared reference expression data,
The reference expression data is a method for determining a cancer patient based on the exosome miRNA that is the data for the set of miRNA isolated from at least one exosome of the patient and normal people. The method of claim 1,
The miRNA set is a method for discriminating cancer patients based on exosome miRNA including all miRNA isolated from exosomes. The method of claim 1,
The computer device targets a cancer patient based on an exosome miRNA that determines the similarity of the miRNA set from which a miRNA having a reads per million mapped reads (RPM) of 0 (zero) is removed from a plurality of samples. How to determine. The method of claim 1,
In the step of reducing the dimension, the computer device determines a cancer patient based on an exosome miRNA that calculates a log (log ₂ ) of a value obtained by adding a specific value to RPMs (Reads per million mapped reads) of the target expression data. How to. The method of claim 4, wherein
And the computer device performing quantile normalization on the logarithmic result. The method of claim 1,
The computer device may determine a cancer patient based on exosome miRNA that determines the similarity by performing hierarchical clustering on the target expression data and the reference expression data. The method of claim 1,
The computer device may determine a cancer patient based on an exosome miRNA that determines the similarity by MultiDimensional Scaling on the target expression data and the reference expression data. The method of claim 7, wherein
The computer apparatus may determine a cancer patient based on exosome miRNA comparing a pattern of multidimensional scaling analysis results of reference expression data of miRNAs of a plurality of patient exosomes with multidimensional scaling results of the target expression data. . A computer-readable recording medium having recorded thereon a program for executing a method for determining a cancer patient based on the exosome miRNA according to any one of claims 1 to 8 on a computer. An input device for receiving target expression data of a miRNA set separated from an exosome of a target;
A storage device for storing reference expression data prepared in advance and storing a program for determining a similarity degree between the target expression data and the reference expression data; And
A computing device that reduces the dimension of the target expression data using the program and calculates the similarity;
The reference expression data is data for a miRNA set isolated from at least one exosome of a patient and a normal person, wherein the miRNA set is a device for discriminating a cancer patient based on an exosome miRNA including all miRNAs separated from the exosome . The method of claim 10,
The arithmetic unit targets a cancer patient based on the exosome miRNA that determines the similarity of the miRNA set from which a miRNA having a reads per million mapped reads (RPM) of 0 (zero) is removed from a plurality of samples. Device to determine. The method of claim 10,
The computing device calculates a log (log ₂ ) of a value obtained by adding a specific value to RPMs (Reads per million mapped reads) of the target expression data, and an exo for quantile normalization of the result of the log operation. Some device to identify cancer patients based on miRNA. The method of claim 10,
The computing device is a device for determining a cancer patient based on the exosome miRNA to determine the similarity by performing a hierarchical clustering (Herarchical clustering) of the target expression data and the reference expression data. The method of claim 10,
The computing device is a device for discriminating cancer patients based on the exosome miRNA to determine the similarity by the multi-dimensional scaling (MultiDimensional Scaling) for the target expression data and the reference expression data. The method of claim 14,
The computing device is a device for discriminating cancer patients based on exosome miRNA comparing a pattern of multidimensional scaling analysis results of reference expression data of miRNAs of a plurality of patient exosomes with multidimensional scaling results of the target expression data. .

Images