CN108570501B - Multiple myeloma molecular typing and application - Google Patents

Abstract

The invention discloses molecular typing and application of multiple myeloma. The invention provides products, including substances for obtaining or detecting the expression of 97 genes in patients with multiple myeloma tumors to be tested; the above products also include equipment for running a Bayesian classifier of multiple myeloma; the invention identifies a The MCL1 gene co-expressed gene module (referred to as MCL1‑M for short) was used to divide multiple myeloma into two main subtypes, MCL‑M‑High and MCL‑M‑Low. The two subtypes have significantly different prognostic and genetic features.

Description

Molecular typing and application of multiple myeloma

technical field

The invention belongs to the field of biotechnology, and in particular relates to molecular typing and application of multiple myeloma.

Background technique

Multiple Myeloma (MM) is a tumor caused by the malignant proliferation of plasma cells, and is the second most common blood tumor, with an incidence of 1-2/100,000 people in China. Multiple myeloma is a common disease in the elderly over 60 years old. With the aggravation of aging in my country, the incidence rate is increasing year by year, and it has become a disease that seriously threatens the health of the elderly. Multiple myeloma is typically characterized by the presence of large numbers of abnormally proliferating plasma cells in the bone marrow that secrete an abnormal immunoglobulin or immunoglobulin fragment, the M protein.

Survival in multiple myeloma has improved markedly with the introduction of proteasome inhibitors such as bortezomib and immunomodulatory drugs such as lenalidomide. However, multiple myeloma is still not completely cured. Multiple myeloma is biologically and clinically highly heterogeneous, and therefore, its response to multidrug combination therapy and post-treatment survival vary widely among different patients. The biological mechanisms responsible for this difference are currently not fully understood, which to a certain extent hinders the development of personalized precision therapy. Therefore, in order to deepen the understanding of the biological nature of multiple myeloma and assist clinical treatment decisions, it is urgent to develop a simple and reliable molecular typing system. At present, several molecular typing systems for multiple myeloma have been proposed internationally. For example, Bergsagel et al. identified 8 multiple myeloma subtypes with distinct Cyclin D (Cyclin D) expression and chromosomal translocations. Using unbiased and hypothetical transcriptome analysis, Zhan and Broyl et al. proposed that multiple myeloma has 7-10 molecular subtypes, which can be further simplified into high-risk and low-risk groups, depending on how long the patient survives. Group. In addition, prognostic-related gene expression signatures such as UAMS-70 and UAMS-17, UAMS-80, IFM-15, Millennium-100, EMC-92, gene amplification indices such as GPI-5, MRC-IX-6, and central A body expansion index was also proposed. However, these molecular typing and expression signatures do not predict drug treatment response, nor do they correlate with the development of plasma cells, and the association between the genes used for molecular typing and the etiology of multiple myeloma has not been elucidated.

SUMMARY OF THE INVENTION

An object of the present invention is to provide the use of obtaining or detecting substances expressed in 97 genes in a patient with multiple myeloma tumors to be tested.

The invention provides the application of obtaining or detecting the substances expressed in 97 genes of the multiple myeloma tumor patients to be tested in preparing a product for predicting the prognosis and survival rate of the multiple myeloma tumor patients to be tested.

The invention also provides the application of obtaining or detecting the substances expressed by 97 genes in the multiple myeloma tumor patients to be tested in preparing a product for predicting the prognosis and survival period of the multiple myeloma tumor patients.

The invention also provides the application of obtaining or detecting the substances expressed in 97 genes in the multiple myeloma tumor patients to be tested in preparing a product for predicting the prognosis and survival risk of the multiple myeloma tumor patients.

Another object of the present invention is to provide the use of a device for obtaining or detecting the expression of 97 genes in a patient with multiple myeloma tumors to be tested and for running a Bayesian classifier for multiple myeloma.

The invention provides equipment for obtaining or detecting 97 gene expression substances in patients with multiple myeloma tumors to be tested and equipment for running a Bayesian classifier of multiple myeloma. application in the product.

The invention also provides equipment for obtaining or detecting 97 gene expression substances in the multiple myeloma tumor patients to be tested and running the multiple myeloma Bayesian classifier in the preparation of a product for predicting the prognosis and survival period of the multiple myeloma tumor patients Applications.

The invention also provides equipment for obtaining or detecting 97 gene expression substances in the multiple myeloma tumor patients to be tested and running the multiple myeloma Bayesian classifier in the preparation of a product for predicting the prognosis and survival risk of multiple myeloma tumor patients Applications.

The above 97 genes are as follows:

ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593.

The above Bayesian classifier for multiple myeloma is obtained according to a method comprising the following steps:

1) Obtain the expression data of 97 genes of n multiple myeloma samples;

2) dividing the expression data of the 97 genes of the n multiple myeloma samples into two subtypes, MCL1-M-High and MCL1-M-Low, using the ConsensusClustering clustering algorithm;

3) Based on the two subtypes in step 2, the expression data of 97 genes of n multiple myeloma samples in step 1), and the prognosis and survival data of n multiple myeloma samples, use the Naive Bayes method Build the prime Bayes classifier.

The above step 3) is to first randomly divide the n multiple myeloma samples into a training set and a verification set according to a ratio of the number of samples greater than 1:1; then use the expression data of the 97 genes in the training, and Combined with the MCL1-M-High and MCL1-M-Low subtype labels of each sample obtained by the Consensus Clustering clustering algorithm, the naive Bayes algorithm in the R language machine learning package klaR package was used to establish a prediction of single patient MCL1-M - Bayesian classifier of multiple myeloma of High subtype and MCL1-M-Low subtype;

The method for obtaining the expression data of the 97 genes of each multiple myeloma sample described above is detection or obtaining from a database.

The third object of the present invention is to provide a product.

The products provided by the present invention include obtaining or detecting substances expressed in 97 genes in the multiple myeloma tumor patients to be tested and equipment for running the multiple myeloma Bayesian classifier.

Among the above-mentioned products, the product has at least one of the following functions:

1) Predict the prognosis and survival rate of patients with multiple myeloma tumors to be tested;

2) Predict the prognosis and survival time of patients with multiple myeloma tumors;

3) Predict the prognosis and survival risk of patients with multiple myeloma tumors.

The above product also includes a carrier for recording the detection method;

The detection method includes the following steps: the detection method includes the following steps: obtaining 97 samples of the multiple myeloma tumor patient to be detected by using the obtained or detected substance expressed by 97 genes in the multiple myeloma tumor patient to be detected. Gene expression data; then the expression data of the 97 genes of the multiple myeloma tumor patients to be tested are classified by the multiple myeloma Bayesian classifier, which belong to the MCL1-M-High subtype. The predicted prognosis survival of patients with multiple myeloma tumors was significantly lower than that of patients with multiple myeloma tumors belonging to the MCL1-M-Low subtype.

In the above product, the multiple myeloma tumor patient to be tested is a single patient or multiple patients.

The fourth object of the present invention is to provide a method for constructing a model for the typing of multiple myeloma tumor patients.

The method provided by the present invention comprises the following steps:

1) Obtain the expression data of 97 genes of n multiple myeloma samples;

2) dividing the expression data of the 97 genes of the n multiple myeloma samples into two subtypes, MCL1-M-High and MCL1-M-Low, using the ConsensusClustering clustering algorithm;

3) Based on the two subtypes in step 2), the expression data of 97 genes of n multiple myeloma samples in step 1), and the prognosis and survival data of n multiple myeloma samples, use Naive Bayesian The method constructs the prime Bayes classifier, which is the target model.

The above products also include equipment for running the Bayesian classifier of multiple myeloma (the equipment can be a CD-ROM or a computer, etc.);

The above product also includes a carrier for recording the detection method;

The detection method comprises the following steps: obtaining or detecting the expression data of the 97 genes in the multiple myeloma tumor patient to be detected by using the substance obtained or detecting the expression of the 97 genes in the multiple myeloma tumor patient to be detected; The expression data of the 97 genes of the multiple myeloma tumor patients to be tested are classified by the multiple myeloma Bayesian classifier, and the multiple myeloma tumor patients belonging to the MCL1-M-High subtype are predicted to be tested. Prognostic survival was significantly lower than patients with multiple myeloma tumors under test belonging to the MCL1-M-Low subtype.

In the above product, the multiple myeloma tumor patient to be tested is a single patient or multiple patients.

Among the above products, the n multiple myeloma samples are 551 samples.

Or the ratio greater than 1:1 is to randomly divide the training set and the validation set according to the ratio of 2:1.

The above products have at least one of the following functions:

1) Predict the prognosis and survival rate of patients with multiple myeloma tumors to be tested;

2) Predict the prognosis and survival time of patients with multiple myeloma tumors;

3) Predict the prognosis and survival risk of patients with multiple myeloma tumors.

Another object of the present invention is to provide a method of constructing a model for the typing of patients with multiple myeloma tumors.

The method provided by the present invention comprises the following steps:

1) Obtain the expression data of 97 genes of n multiple myeloma samples;

2) dividing the expression data of the 97 genes of the n multiple myeloma samples into two subtypes, MCL1-M-High and MCL1-M-Low, using the ConsensusClustering clustering algorithm;

3) Based on the two subtypes in step 2), the expression data of 97 genes of n multiple myeloma samples in step 1), and the prognosis and survival data of n multiple myeloma samples, use Naive Bayesian The method constructs the prime Bayes classifier, which is the target model.

The above-mentioned substances for obtaining or detecting the expression of 97 genes in patients with multiple myeloma tumors to be tested and/or the equipment for running the Bayesian classifier of multiple myeloma or the models obtained by the above-mentioned methods are used to predict the multiple myeloma tumors to be tested. The application in the product of the prognosis and survival rate of tumor patients is also the scope of protection of the present invention.

The above-mentioned substances for obtaining or detecting the expression of 97 genes in patients with multiple myeloma tumors to be tested and/or the equipment for running the Bayesian classifier of multiple myeloma or the models obtained by the above-mentioned methods are used to prepare and predict multiple myeloma tumors. The application in the product of patient prognosis and survival is also the scope of protection of the present invention.

The above-mentioned substances for obtaining or detecting the expression of 97 genes in patients with multiple myeloma tumors to be tested and/or the equipment for running the Bayesian classifier of multiple myeloma or the models obtained by the above-mentioned methods are used to prepare and predict multiple myeloma tumors. The application in the product of patient prognosis survival risk is also the scope of the protection of the present invention.

The present invention also provides a method for classifying patients with multiple myeloma tumors to be tested, comprising the following steps:

Detect or obtain the expression data of the 97 genes of the multiple myeloma tumor patient to be tested; and then use the above-mentioned multiple myeloma Bayesian classification for the expression data of the 97 genes of the multiple myeloma tumor patient to be tested. According to the classification of the multiple myeloma tumor patients to be tested, whether they belong to the MCL1-M-High subtype or the MCL1-M-Low subtype.

The present invention also provides a method for predicting the prognosis and survival risk of multiple myeloma tumor patients, comprising the steps of: detecting or obtaining the expression data of 97 genes of the multiple myeloma tumor patients to be detected; The expression data of 97 genes in patients with multiple myeloma tumors were classified by the above-mentioned Bayesian classifier for multiple myeloma, and the patients with multiple myeloma tumors belonging to the MCL1-M-High subtype were predicted to predict the prognosis and survival rate. lower than in patients with multiple myeloma tumors to be tested belonging to the MCL1-M-Low subtype.

The present invention also provides a method for predicting the prognosis and survival risk of multiple myeloma tumor patients, comprising the steps of: detecting or obtaining the expression data of 97 genes of the multiple myeloma tumor patients to be detected; The expression data of 97 genes were measured in patients with multiple myeloma tumors. The above Bayesian classifier for multiple myeloma was used for classification. The patients with multiple myeloma tumors belonging to the MCL1-M-High subtype had a poor prognosis and survival rate. The risk of poor prognosis in patients with multiple myeloma tumors belonging to the MCL1-M-Low subtype is small.

The expression levels of the above genes are all gene expression levels in tumor cells.

To overcome the above drawbacks, the inventors explored whether gene co-expression networks surrounding key signaling pathways conserved during germinal center (GC) plasma cell development could aid in the elucidation of MM pathogenesis and its application in molecular typing of MM. The inventors focused on finding a gene network in multiple myeloma that controls the dysregulation of B cell differentiation into plasma cells, as it may play a key role in the development of multiple myeloma. After the above analysis, a gene module co-expressed with the MCL1 gene (referred to as MCL1-M) was identified and applied to divide multiple myeloma into two main subtypes, namely MCL-M-High subtype and MCL -M-Low subtype. The two subtypes have significantly different prognostic and genetic features, and more importantly, the classification system predicts patient response to bortezomib treatment and correlates with the developmental stage of plasma cells. These findings could pave the way for future implementation of individualized precision therapy and improve understanding of the etiology of multiple myeloma.

Description of drawings

Figure 1 shows the ROC diagram of the Bayesian classifier typing results in the GSE2658 validation set.

Figure 2 shows the ROC diagram of the Bayesian classifier typing results in the MMRF dataset.

Figure 3 is the ROC diagram of the Bayesian classifier typing results in the GSE19784 dataset.

Figure 4 is an overall survival curve of multiple myeloma MCL1-M-High and MCL1-M-Low molecular subtypes in GSE2658.

Figure 5 is an overall survival curve of multiple myeloma MCL1-M-High and MCL1-M-Low molecular subtypes in GSE2658.

Figure 6 is an overall survival curve (upper panel) and progression-free survival curve (lower panel) of multiple myeloma MCL1-M-High and MCL1-M-Low molecular subtypes in GSE19784.

Figure 7 shows that patients with MCL1-M-High and MCL1-M-Low subtypes in GS19784 have different responses to bortezomib treatment.

Detailed ways

The experimental methods used in the following examples are conventional methods unless otherwise specified.

The materials, reagents, etc. used in the following examples can be obtained from commercial sources unless otherwise specified.

Example 1. Screening of molecular diagnostic markers for multiple myeloma and implementation of molecular typing

Using the multiple myeloma expression data set GSE2658 provided by the NCBI GEO public database, through Pearson correlation analysis, 87 genes co-expressed with MCL1 were obtained. -M gene enriched in multiple myeloma samples. For more stable molecular typing, 36 of the 133 genes with low classification efficiency were further screened, and finally 97 classified genes with stable differential expression and relatively high abundance were retained.

The names of the 97 genes are as follows:

ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593.

These 97 genes will then serve as classification factors for typing. Using the expression data of these 97 genes in 551 multiple myeloma samples from the GSE2658 dataset, we first used the Consensus Clustering clustering algorithm to classify the 551 multiple myeloma cases into MCL1-M- There are two subtypes, High and MCL1-M-Low. However, cluster-based classification methods cannot perform molecular typing of individual samples. In order to implement individualized diagnosis, the 551 samples were randomly divided into training set (369 cases) and validation set (182 cases) according to the ratio of 2:1, which were used to build and evaluate individualized classifiers. Sampling adopts a stratified sampling method to ensure that the ratio of the two subtypes of MCL1-M-High and MCL1-M-Low in the training set and the test set is consistent with the original.

According to the expression data of 97 classified genes of 369 samples in the training set and the two subtype labels of MCL1-M-High and MCL1-M-Low divided by the Consensus Clustering clustering algorithm, the machine learning package klaR package in R language is used to provide The Naive Bayes algorithm built a multiple myeloma Bayesian classifier that predicted the MCL1-M-High and MCL1-M-Low subtypes in individual patients.

The classification accuracy was evaluated using 182 samples in the validation set.

According to the returned accuracy, the model is iteratively optimized, and finally the classification accuracy exceeds 95%. The accuracy data of the classifier is shown in Table 1, and the receiver operating curve (ROC) is shown in Figure 1.

Table 1. The accuracy of the classifier built with the GSE2658 dataset in the validation set

In order to confirm that the classifier built using the GSE2658 dataset can be generalized and applied. The inventors then used this classifier to predict the molecular subtypes of the samples in the multiple myeloma large dataset MMRF and the GEO multiple myeloma expression dataset GSE19784 published by NCI.

The MMRF dataset is different from GSE2658 in that the expression levels of its genes were obtained by RNA-seq instead of microarray. The predicted results are shown in Table 2, and the ROC diagram is shown in Figure 2.

Table 2. Accuracy of classifiers built with GSE2658 dataset in MMRF dataset

The results show that the classifier can maintain a high accuracy even across platforms, which shows that it has high promotion and application value.

GSE19784 is also a multiple myeloma expression dataset. Like GSE2618, the U1332.0Plus chip is used to measure gene expression. However, the two origins are detected at different times in different experiments, and the experimental conditions may be different, which results in significantly different data distributions and noise levels between the two. The subtype prediction results of the database are shown in Table 3, and the ROC curve is shown in Figure 3.

Table 3. Accuracy of classifiers built with GSE2658 dataset on GSE19784 dataset

The results show that the classifier can better overcome the above problems and still maintain a high accuracy.

Example 2. Application of Bayesian classifier of multiple myeloma in predicting patient prognosis and survival rate

1. Database GSE2658

According to the expression data of 97 classified genes in 551 multiple myeloma patient samples (detected before treatment) in the GSE2658 database, the 551 samples were classified by the Bayesian classifier of multiple myeloma obtained in Example 1, and 249 Cases of MCL1-M-High subtype multiple myeloma and 302 MCL1-M-Low subtype multiple myeloma.

551 sample patients were followed up for 72 months after treatment. According to the follow-up results, survival analysis (K-M analysis and cox regression analysis) was performed. The results are shown in Figure 4. It can be seen that MCL1-M-High and MCL1-M-Low The two multiple myeloma subtypes had significantly different prognosis, with the MCL1-M-High subtype having a lower overall survival rate than the MCL1-M-Low subtype (log-rank test, p=0.0201, likelihood ratio test, Hazard ratio 1.588, p=0.0212).

Therefore, the use of Bayesian classifier to use the 97 genes of the MCL1 gene group for typing can be used to predict the prognosis of the patient to be tested.

2. Database MMRF

According to the expression data of 97 classified genes of MCL1 gene group in 534 samples of multiple myeloma patients in the MMRF dataset (detected before treatment), the Bayesian classifier of multiple myeloma obtained in Example 1 was used to separate the 534 samples. Divided into two subtypes, MCL1-M-High (231 cases) and MCL1-M-Low (303 cases).

534 sample patients were followed up for 48 months after treatment. According to the follow-up results, survival analysis (K-M analysis and cox regression analysis) was performed. The results are shown in Figure 5. It can be seen that in the MMRF data set, MCL1-M-High and The two subtypes of MCL1-M-Low also have significantly different prognosis, and the overall survival rate of the MCL1-M-High subtype is lower than that of the MCL1-M-Low subtype (log-rank test, p=0.006663, likelihood ratio test, hazard ratio 1.838, p=0.00706).

The results indicate that no matter which platform the gene expression data comes from, the use of Bayesian classifier to classify 97 genes of the MCL1 gene group can be used to predict the prognosis of the patients to be tested.

3. Database GSE19784

According to the expression data of 97 classified genes of the MCL1 gene group in the validation set of 304 multiple myeloma patient samples (pre-treatment detection) in the database GSE19784, the multiple myeloma Bayesian classifier obtained in Example 1 was used to separate the The 304 samples were divided into two subtypes, MCL1-M-High (107 cases) and MCL1-M-Low (196 cases).

The 304 sample patients were followed up for 96 months after treatment. According to the follow-up results, survival analysis (K-M analysis and cox regression analysis) was performed. The results are shown in Figure 6 (A is the overall survival rate, B is the progression-free survival rate). See, in the GSE19784 dataset, the two subtypes MCL1-M-High and MCL1-M-Low also have significantly different prognosis, the overall survival rate of the MCL1-M-High subtype is better than that of the MCL1-M-Low subtype Low (log-rank test, p<0.0001, likelihood ratio test, hazard ratio 1.91, p=0.0002). The GSE19784 dataset also includes disease progression information, so we also analyzed the difference in progression-free survival. Similarly, the MCL1-M-High subtype has a log-lower PFS rate than the MCL1-M-Low subtype. rank test, p=0.0282, likelihood ratio test, hazard ratio 1.36, p=0.031) This result once again shows that the use of Bayesian classifiers for typing with 97 genes of the MCL1 gene group can be used to predict the patient's disease. Prognosis.

Example 3. Molecular diagnostic markers and typing of multiple myeloma in predicting whether the patient to be tested can be treated with bortezomib

The GSE19784 multiple myeloma expression dataset was obtained from a clinical trial of a phase III drug (HOVON-65/GMMG-HD4), with the patient's drug regimen. The trial randomly divided patients into two groups to receive VAD (155 cases) and PAD (148 cases) two drug combinations, the difference between the two is that the PAD regimen has more bortezomib (trade name: Velcade). The gene expression data of the enrolled patients were collected before treatment.

The patients were stratified according to the above-mentioned MCL1-M molecular classification, and then were divided into two subgroups, MCL1-M-High (51 PAD, VAD 56) and MCL1-M-Low (PAD 104, VAD 92). The patients were divided into groups according to the drug treatment plan and the survival analysis (K-M analysis and cox regression analysis) was carried out.

The results are shown in Figure 7, A is the overall survival rate of the MCL1-M-High group, B is the overall survival rate of the MCL1-M-Low group, C is the progression-free survival rate of the MCL1-M-High group, and D is the MCL1 - Progression-free survival in the M-Low group; it can be observed that the PAD drug with bortezomib only prolongs the survival of patients in the MCL1-M-High group, especially the progression-free survival (Fig. 7 left, MCL -M-High group, right MCL-M-Low group; top, overall survival curve, bottom, progression-free survival curve), which revealed that bortezomib can clinically delay the recurrence and deterioration of patients in the MCL-M-High group , but there was no effect in the MCL-M-Low group of patients. To sum up, the implementation of the molecular typing of the present invention can guide clinical medication and avoid the use of bortezomib in patients in the MCL1-M-Low group. Side effects from drug treatment.

Claims (4)

1. The application of obtaining or detecting the expression substances of 97 genes in patients with multiple myeloma tumors to be tested in the preparation of products for predicting the prognosis and survival rate of patients with multiple myeloma tumors to be tested; The 97 genes are as follows: ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593. 2. The application of obtaining or detecting the expression substances of 97 genes in patients with multiple myeloma tumors to be tested in the preparation of products for predicting the prognosis and survival of patients with multiple myeloma tumors; The 97 genes are as follows: ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593. 3. The application of obtaining or detecting the expression substances of 97 genes in patients with multiple myeloma tumors to be tested in the preparation of products for predicting the prognosis and survival risk of patients with multiple myeloma tumors; The 97 genes are as follows: ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593. 4. Obtaining or detecting the substances expressed by 97 genes in the multiple myeloma tumor patients to be tested and the equipment for running the multiple myeloma Bayesian classifier in the preparation of products for predicting the prognosis and survival rate of the multiple myeloma tumor patients to be tested Applications; Application of obtaining or detecting substances expressed in 97 genes in multiple myeloma tumor patients to be tested and equipment for running multiple myeloma Bayesian classifiers in the preparation of products for predicting the prognosis and survival time of multiple myeloma tumor patients Or the application of obtaining or detecting 97 gene expression substances in multiple myeloma tumor patients to be tested and equipment for running multiple myeloma Bayesian classifiers in the preparation of products for predicting the prognosis and survival risk of multiple myeloma tumor patients; The 97 genes are as follows: ACBD3, ADAR, ADSS, ALDH2, ANP32E, ANXA2, ATF3, ATP8B2, CACYBP, CAPN2, CCND1, CCT3, CDC42SE1, CERS2, CHSY3, CLIC1, CLMN, COPA, CSNK1G3, DAP3, DENND1B, ENSA, EPRS, EPSTI1, EVL, FAM13A, FAM49A, FLAD1, FRZB, GLRX2, HAX1, HDGF, HLA-A, HLA-B, HLA-C, HLA-F, HLA-G, IL6R, ISG20L2, JTB, KLF2, LAMTOR2, LDHA, MCL1, MOXD1, MRPL24, MRPL9, MVP, MYL6, NDUFS2, NOP58, NOTCH2NL, NTAN1, PAK1, PI4KB, PIEZO1, PIK3AP1, PIM2, PIP5K1B, PMVK, POGZ, PPIA, PRCC, PRKCA, PRRC2C, PSMB4, PSMD4, RAB29, RCBTB2, SCAMP3, SCAPER, SDHC, SEL1L3, SELPLG, SHC1, SIDT1, SSR2, STAP1, TAP1, TIMM17A, TLR10, TMCO1, TOR1AIP2, TOR3A, TP53INP1, TPM3, TRANK1, TROVE2, UAP1, UBE2Q1, UBQLN4, UHMK1, VPS45, YY1AP1, ZC3H11A, ZFP36 and ZNF593; The multiple myeloma Bayesian classifier is obtained according to a method comprising the following steps: 1) Obtain the expression data of 97 genes of n multiple myeloma samples; 2) dividing the expression data of the 97 genes of the n multiple myeloma samples into two subtypes, MCL1-M-High and MCL1-M-Low, using the Consensus Clustering clustering algorithm; 3) Based on the two subtypes in step 2, the expression data of 97 genes of n multiple myeloma samples in step 1), and the prognosis and survival data of n multiple myeloma samples, use the Naive Bayes method Build the prime Bayes classifier.

Images