Breast cancer prognosis evaluation method and system and electronic equipment

Abstract

The invention provides a prognosis evaluation method and system for breast cancer and electronic equipment, and relates to the technical field of biologication.The prognosis evaluation method comprises the steps that first sample information and second sample information are collected, and the first sample information comprises sequencing data of intestinal microbiome; the second sample information comprises gene data and survival data of historical breast cancer patients; determining a key path according to the first sample information; searching a target gene according to the gene data of the historical breast cancer patient and the key pathway; constructing and training a prognosis evaluation model based on the target gene and the second sample information; acquiring basic information of a user; the basic information comprises an expression value of a target gene; and substituting the basic information into the prognosis evaluation model to obtain a prognosis evaluation result so as to assist in predicting the prognosis condition of the breast cancer patient and improve the prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the field of biotechnology, and in particular to a method, system and electronic equipment for evaluating the prognosis of breast cancer. Background Art

[0002] Breast cancer is a common malignant tumor in women. It originates from the uncontrolled proliferation of mammary epithelial cells. Tumor types are divided into non-invasive cancer and invasive cancer. Non-invasive cancer (carcinoma in situ) usually has a better prognosis because they have not invaded surrounding tissues or metastasized to other parts of the body. Invasive cancer may have a higher risk of recurrence and metastasis. The number of breast cancer cases has been increasing year by year in recent years.

[0003] Breast cancer is a complex disease whose occurrence, development and prognosis are affected by multiple factors, including genes, environment, lifestyle, etc. The interaction of these factors makes the prognosis prediction of breast cancer complicated and has low accuracy.

[0004] Therefore, a method, system and electronic device for evaluating the prognosis of breast cancer need to be proposed. Summary of the invention

[0005] This specification provides a prognosis assessment method, system and electronic device for breast cancer. By analyzing intestinal microbiome data, key pathways and target genes related to breast cancer prognosis are identified, and an effective prognosis assessment model is constructed based on these genes and survival data to assist in predicting the prognosis of breast cancer patients and improve the accuracy of the prediction.

[0006] The present application provides a method for evaluating the prognosis of breast cancer using the following technical solution, including:

[0007] Collecting first sample information and second sample information, wherein the first sample information includes sequencing data of the intestinal microbiome; the second sample information includes gene data and survival data of historical breast cancer patients; determining a key pathway based on the first sample information; searching for a target gene through the gene data of the historical breast cancer patients and the key pathway; and constructing and training a prognosis evaluation model based on the target gene and the second sample information;

[0008] Obtaining basic information of the user; the basic information includes the expression value of the target gene;

[0009] Substituting the basic information into the prognosis evaluation model, a prognosis evaluation result is obtained.

[0010] Optionally, determining a key path according to the first sample information includes:

[0011] constructing a plurality of original sample groups according to the first sample information;

[0012] Determine the significant bacterial genus corresponding to the original sample group;

[0013] In combination with the characteristic information of the significant bacterial genus, a first clustering is performed on the significant bacterial genus corresponding to each of the original sample groups to obtain a first clustering result;

[0014] Predicting the functional modules of significant bacterial genera, and extracting similar functional modules based on the first clustering results;

[0015] According to the similar functional modules, a path corresponding to the similar functional modules is searched and used as the key path.

[0016] Optionally, determining the significant bacterial genus corresponding to the original sample group includes:

[0017] Comparing the control sample with the experimental sample in the same original sample group to obtain the differential bacterial genera corresponding to the original sample group;

[0018] Calculating the importance of each of the differential bacterial genus according to the characteristic information of the differential bacterial genus;

[0019] According to the importance of each differential bacterial genus, the average importance was calculated;

[0020] Based on the average importance value, the significant genera corresponding to the original sample group were screened out from the differential genera.

[0021] Optionally, combining the characteristic information of the significant bacterial genus and performing a first clustering on the significant bacterial genus corresponding to each of the original sample groups to obtain a first clustering result includes:

[0022] Constructing a first clustering model, determining the number of nodes and the weight vectors of the nodes;

[0023] Calculate the similarity between each of the significant bacterial genus and the weight vectors of all the nodes;

[0024] Assigning the significant bacterial genus to the most similar node;

[0025] All significant genera assigned to the same node were considered to belong to the same cluster.

[0026] Optionally, the first clustering result includes a plurality of clusters;

[0027] The predicting of the functional modules of the significant bacterial genera and extracting similar functional modules in combination with the first clustering result include:

[0028] Predicting the functional modules of the significant bacterial genus to obtain first functional information of each cluster, wherein the first functional information includes: all functional modules corresponding to the cluster;

[0029] Finding a common functional module among clusters in the same original sample group as a first functional module corresponding to the original sample group;

[0030] Finding a first functional module common to each of the original sample groups as a second functional module common to all the original sample groups;

[0031] Collecting preset function module information, comparing the preset function module information with the second function module, and taking the overlapping second function modules as the similar function modules.

[0032] Optionally, constructing and training a prognosis evaluation model based on the target gene and the second sample information includes:

[0033] Performing a second clustering on the basic samples by using a second clustering model to obtain a second clustering result; the second clustering result includes a plurality of clusters;

[0034] All the survival data in each of the clusters are aggregated to determine the prognosis result corresponding to the cluster.

[0035] The present application provides a breast cancer prognosis assessment system that adopts the following technical solutions, including:

[0036] Model building module;

[0037] An information collection module is used to obtain basic information of the user; the basic information includes the expression value of the target gene;

[0038] The prognosis evaluation module is used to substitute the basic information into the prognosis evaluation model to obtain the prognosis evaluation result.

[0039] Optional, model building modules, including:

[0040] A first collection submodule is used to collect first sample information and second sample information, wherein the first sample information includes sequencing data of the intestinal microbiome; and the second sample information includes gene data and survival data of historical breast cancer patients;

[0041] A key path determination submodule, used to determine the key path according to the first sample information;

[0042] A target gene search submodule, used to search for target genes through the gene data of the historical breast cancer patients and the key pathways;

[0043] A model building submodule, used to build and train a prognosis evaluation model based on the target gene and the second sample information;

[0044] Optionally, the key path determination submodule includes:

[0045] A sample group construction unit, configured to construct a plurality of original sample groups according to the first sample information;

[0046] A significant bacterial genus determination unit, used to determine the significant bacterial genus corresponding to the original sample group;

[0047] A first clustering unit is used to perform first clustering on the significant bacterial genus corresponding to each of the original sample groups in combination with the characteristic information of the significant bacterial genus to obtain a first clustering result;

[0048] A function prediction unit, used to predict the function modules of significant bacterial genera and extract similar function modules in combination with the first clustering result;

[0049] The key path determining unit is used to search for paths corresponding to similar function modules according to the similar function modules as the key paths.

[0050] Optionally, the significant bacterial genus determination unit includes:

[0051] A comparison subunit, used to compare the control sample and the experimental sample in the same original sample group to obtain the different bacterial genera corresponding to the original sample group;

[0052] An importance calculation subunit, used for calculating the importance of each of the differential bacterial genus according to the characteristic information of the differential bacterial genus;

[0053] The average value calculation subunit is used to calculate the importance average value according to the importance of each differential bacterial genus;

[0054] The significant bacterial genus determination subunit is used to screen out the significant bacterial genus corresponding to the original sample group from the differential bacterial genus based on the average importance value.

[0055] Optionally, the first clustering unit includes:

[0056] A first clustering model building subunit is used to build a first clustering model and determine the number of nodes and the weight vectors of the nodes;

[0057] A similarity calculation subunit, used to calculate the similarity between each of the significant bacterial genus and the weight vectors of all the nodes;

[0058] An allocating subunit, used to allocate the significant bacterial genus to the most similar node;

[0059] All significant genera assigned to the same node were considered to belong to the same cluster.

[0060] Optionally, the first clustering result includes a plurality of clusters;

[0061] Optional, function prediction unit, including:

[0062] A first prediction subunit is used to predict the functional modules of the significant bacterial genus to obtain first functional information of each cluster, wherein the first functional information includes: all functional modules corresponding to the cluster;

[0063] A first screening subunit, configured to search for a common functional module among clusters in the same original sample group as a first functional module corresponding to the original sample group;

[0064] A second screening subunit is used to find the first functional module common to each of the original sample groups as the second functional module common to all the original sample groups;

[0065] The third screening subunit is used to collect preset function module information, compare the preset function module information with the second function module, and use the overlapping second function modules as the similar function modules.

[0066] Optional, model building submodules, including:

[0067] A second clustering unit, used for performing a second clustering on the basic samples by using a second clustering model to obtain a second clustering result; the second clustering result includes a plurality of clusters;

[0068] The summarizing unit is used to summarize all the survival data in each of the clusters to determine the prognosis result corresponding to the cluster.

[0069] This specification also provides an electronic device, wherein the electronic device includes:

[0070] processor; and,

[0071] A memory storing computer executable instructions, which when executed cause the processor to perform any of the above methods.

[0072] The present specification also provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, any of the above methods is implemented.

[0073] In the present application, by collecting first sample information and second sample information, the first sample information includes sequencing data of the intestinal microbiome; the second sample information includes genetic data and survival data of historical breast cancer patients; the key pathway is determined according to the first sample information; the target gene is found through the genetic data of the historical breast cancer patients and the key pathway; based on the target gene and the second sample information, a prognosis evaluation model is constructed and trained; the basic information of the user is obtained; the basic information includes the expression value of the target gene; the basic information is substituted into the prognosis evaluation model to obtain a prognosis evaluation result, so as to assist in predicting the prognosis of breast cancer patients and improve the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic diagram of the principle of a method for evaluating the prognosis of breast cancer provided in an embodiment of this specification;

[0075] Figure 2 A schematic diagram of significant bacterial genera provided in the examples of this specification;

[0076] Figure 3 A schematic diagram of overlapping significant bacterial genera in each original sample group provided in the embodiments of this specification;

[0077] Figure 4 A schematic diagram of the enrichment of key pathways in each cluster provided in the embodiments of this specification;

[0078] Figure 5 Validation of four clusters of breast cancer provided in the embodiments of this specification Figure 1 ;

[0079] Figure 6 Validation of four clusters of breast cancer provided in the embodiments of this specification Figure 2 ;

[0080] Figure 7 A schematic diagram of the structure of a breast cancer prognosis assessment system provided in an embodiment of this specification;

[0081] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of this specification;

[0082] Fig. 9 A schematic diagram of a computer-readable storage medium provided in accordance with an embodiment of the present specification. DETAILED DESCRIPTION

[0083] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0084] Exemplary embodiments of the present invention will now be described more fully with reference to the accompanying drawings. However, exemplary embodiments can be implemented in a variety of forms, and should not be construed as limiting the present invention to the embodiments set forth herein. On the contrary, providing these exemplary embodiments enables the present invention to be more comprehensive and complete, and is more convenient for fully conveying the inventive concept to those skilled in the art. The same reference numerals in the figures represent the same or similar elements, components or parts, and thus their repeated description will be omitted.

[0085] Under the premise of being consistent with the technical concept of the present invention, the features, structures, characteristics or other details described in a specific embodiment do not exclude that they can be combined in one or more other embodiments in a suitable manner.

[0086] In the description of specific embodiments, the features, structures, characteristics or other details described in the present invention are intended to enable those skilled in the art to fully understand the embodiments. However, it does not exclude that those skilled in the art can practice the technical solutions of the present invention without one or more of the specific features, structures, characteristics or other details.

[0087] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0088] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0089] The term "and / or" or "and / or" includes all combinations of any one or more of the associated listed items.

[0090] Figure 1 A schematic diagram of the principle of a method for evaluating the prognosis of breast cancer provided in an embodiment of this specification, the method comprising:

[0091] S1 constructs a prognostic evaluation model;

[0092] S2 obtains basic information of the user; the basic information includes the expression value of the target gene;

[0093] S3 substitutes the basic information into the prognosis evaluation model to obtain a prognosis evaluation result.

[0094] The intestinal microbiota is a complex ecosystem that contains a large number of microbial species and complex interactions. At the same time, different diseases may involve different microbial species and complex microbe-host interactions, which makes it very difficult to directly confirm diseases through the intestinal microbiota. Although many studies have shown that the intestinal microbiota is closely related to the occurrence and development of various diseases, there are few studies on clinical subtypes based on the intestinal microbiota, especially in breast cancer.

[0095] Therefore, the present invention constructs a prognostic evaluation model based on intestinal microbial-related metabolic pathways, human gene expression profiles, and patient prognosis and survival status, so as to more accurately predict the prognosis of breast cancer patients and provide more powerful support for clinical decision-making. Specifically, the method includes:

[0096] S1 constructs a prognostic evaluation model;

[0097] The intestinal microbiome is known to change significantly in breast cancer and other cancers, and shows significant associations with these diseases. In order to determine whether the intestinal microbiome in breast cancer shares common functional features with the microbiome in other cancers, and then accurately construct an evaluation model, first, the present invention analyzes the 16S rRNA sequencing data of the intestinal microbiome of various cancers.

[0098] S11 collects first sample information, where the first sample information includes sequencing data of an intestinal microbiome;

[0099] In one embodiment of the present specification, the first sample information includes 16SrRNA sequencing data of the intestinal microbiome of four public datasets.

[0100] S12 determines a key path according to the first sample information;

[0101] S121 constructing a plurality of original sample groups according to the first sample information;

[0102] The present invention collects 16S rRNA sequencing data of intestinal microbiomes from four public datasets and divides them into three original sample groups, each of which corresponds to a cancer category.

[0103] Among them, the four public datasets include: PRJNA658160 dataset, PRJNA817689 dataset, PRJNA639644 dataset and PRJNA86188 dataset. The three original sample groups include: breast cancer (BC) sample group, gastric cancer (GC) sample group and colorectal cancer (CRC) sample group.

[0104] Specifically, the 16S rRNA sequencing data of the intestinal microbiome of the PRJNA658160 dataset was collected to construct a breast cancer (BC) sample group, which included 308 first control samples of normal subjects and 350 first experimental samples of breast cancer patients.

[0105] The 16S rRNA sequencing data of the gut microbiome of the PRJNA817689 dataset and the PRJNA639644 dataset were collected to construct a gastric cancer (GC) sample group, which included 140 second control samples from normal subjects and 124 second experimental samples from GC patients.

[0106] The 16S rRNA sequencing data of the intestinal microbiome of the PRJNA86188 dataset were collected to construct a colorectal cancer (CRC) sample group, which included 260 third control samples from normal subjects and 428 third experimental samples from colorectal cancer patients.

[0107] S122 determining the significant bacterial genus corresponding to the original sample group;

[0108] S122-1 compares the control sample and the experimental sample in the same original sample group to obtain the different bacterial genera corresponding to the original sample group;

[0109] In one embodiment of the present specification, the first control sample is compared with the first experimental sample to find the differential bacterial genus corresponding to the breast cancer sample group. The second control sample is compared with the second experimental sample to find the differential bacterial genus corresponding to the gastric cancer sample group. The third control sample is compared with the third experimental sample to find the differential bacterial genus corresponding to the colorectal cancer sample group.

[0110] S122-2 obtains characteristic information of each differential bacterial genus in the original sample group, and calculates the importance of each differential bacterial genus according to the characteristic information of the differential bacterial genus;

[0111] S122-3, calculating the importance average according to the importance of each differential bacterial genus;

[0112] Specifically, the importance of all differential bacterial genera is added up, and then divided by the total number of species of the differential bacterial genera to obtain the average importance. Since the average importance is determined based on the differential bacterial genera in the original sample group, the average importance may be different in different original sample groups.

[0113] S122-4 screens out the significant bacterial genera corresponding to the original sample group from the differential bacterial genera based on the average importance value.

[0114] Specifically, the importance of each differential bacterial genus is compared with the average importance value. If the importance of the differential bacterial genus is greater than the average importance value, the differential bacterial genus is regarded as a significant bacterial genus.

[0115] Preferably, a random forest model is used for screening.

[0116] Taking the breast cancer sample group as an example, the characteristic information of each differential bacterial genus in the breast cancer sample group is obtained, and the importance of the differential bacterial genus is calculated according to the characteristic information; the importance average is calculated according to the importance of each differential bacterial genus; based on the importance average, the significant bacterial genus corresponding to the breast cancer sample group is screened out from the differential bacterial genus. Among them, the characteristic information includes the abundance of the genus.

[0117] In one embodiment of this specification, Figure 2 As shown, the breast cancer sample group corresponds to 66 significant genera, of which the first control sample corresponds to 59 significant genera and the first experimental sample corresponds to 7 significant genera. The gastric cancer sample group corresponds to 47 significant genera, of which the second control sample corresponds to 17 significant genera and the second experimental sample corresponds to 30 significant genera. The colorectal cancer sample group corresponds to 66 significant genera, of which the third control sample corresponds to 60 significant genera and the third experimental sample corresponds to 6 significant genera.

[0118] S123, combining the characteristic information of the significant bacterial genus, performing first clustering on the significant bacterial genus corresponding to each of the original sample groups by using a first clustering model to obtain a first clustering result; the first clustering result includes a plurality of clusters;

[0119] In one embodiment of the present specification, a first clustering model is constructed. The first clustering model is a K-means clustering model.

[0120] S123-1 randomly selects significant bacterial genera with corresponding numbers according to the optimal number of clusters as the original central bacterial genera;

[0121] S123-2 randomly selects a non-central genus, determines the similarity between the non-central genus and all central genus, and assigns the non-central genus to the cluster where the central genus with the highest similarity is located.

[0122] S123-3 calculates the median of all significant bacterial genera in the cluster and re-determines the central bacterial genus in the cluster;

[0123] S123-4 loops through steps S123-2 and S123-3 until a maximum number of iterations or a convergence threshold is reached; a cluster is determined based on each central bacterial genus and its corresponding cluster members.

[0124] In order to facilitate better fitting and improve the feasibility of interpreting the results, in another embodiment of the present specification, S123-5 constructs a first clustering model, which is a SOM unsupervised learning model.

[0125] Determine the number of nodes and the weight vectors of the nodes;

[0126] Initialize a two-dimensional grid, each grid represents a node. Randomly initialize a weight vector for each node. The dimension of the input data is determined based on the characteristic information of the significant bacterial genus.

[0127] Train the first clustering model and adjust the weight vector. Specifically, obtain training samples and perform the following operations on each sample data in the training samples: calculate the similarity between the sample data and the weight vectors of all nodes, find the node that is most similar to the sample data point as the winning node. Update the weight vectors of the winning node and its surrounding nodes according to the learning rate and neighborhood function.

[0128] Repeat the above process, perform iterative training, and adjust the weight vector.

[0129] S123-6 calculating the similarity between each of the significant bacterial genus and the weight vectors of all the nodes;

[0130] S123-7 assigns the significant bacterial genera to the most similar nodes; all significant bacterial genera assigned to the same node are considered to belong to the same cluster.

[0131] In one embodiment of the present specification, the breast cancer sample group corresponds to three clusters, namely, breast cancer-cluster 1 (BC-G1), breast cancer-cluster 2 (BC-G2), and breast cancer-cluster 3 (BC-G3). That is, according to the experimental results, the optimal number of clusters is determined to be 3.

[0132] The significant bacterial genera corresponding to the gastric cancer sample group were clustered to obtain the clustering results, among which the gastric cancer sample group corresponded to three clusters, namely gastric cancer-cluster 1 (GC-G1), gastric cancer-cluster 2 (GC-G2), and gastric cancer-cluster 3 (GC-G3).

[0133] The significant bacterial genera corresponding to the colorectal cancer sample group were clustered to obtain clustering results, among which the colorectal cancer sample group corresponded to three clusters, namely colorectal cancer-cluster 1 (CRC-G1), colorectal cancer-cluster 2 (CRC-G2), and colorectal cancer-cluster 3 (CRC-G3). These clusters showed significantly different intestinal microbiome characteristics.

[0134] By comparing the overlapping genera in the three cancers, it was found that the differences between the overlapping genera in the three cancers were not extensive, and almost half were unique to each cancer, which may be related to tumor specificity. In other words, although there are some common genera in the three different cancers, the performance or difference of these common genera in each cancer is not significant or obvious.

[0135] like Figure 3 As shown, the top represents the number of significant bacterial genera in each original sample group (bactertial GeneraCounts). The data below represents the number of overlaps of bacterial genera in different original sample groups (Intersection Size). Among the three original sample groups, there are 25 significant bacterial genera that exist only in the breast cancer (BC) sample group. There are 24 significant bacterial genera that exist only in the gastric cancer (GC) sample group. There are 19 significant bacterial genera that exist only in the colorectal cancer (CRC) sample group.

[0136] There are 29 significant genera overlapping between the colorectal cancer (CRC) sample group and the breast cancer (BC) sample group. There are 11 significant genera overlapping between the colorectal cancer (CRC) sample group and the gastric cancer (GC) sample group. There are 5 significant genera overlapping between the breast cancer (BC) sample group and the gastric cancer (GC) sample group.

[0137] There are 7 significant genera overlapping in the colorectal cancer (CRC) sample group, the breast cancer (BC) sample group, and the gastric cancer (GC) sample group.

[0138] Although Escherichia coli-Shigella is a common genus in breast cancer, gastric cancer, and colorectal cancer, its enrichment pattern in these three cancers is not the same. It is enriched in the second cluster (BC-G2) in the breast cancer sample group; enriched in the third cluster (GC-G3) in the gastric cancer sample group, and enriched in the first cluster (CRC-G1) in the colorectal cancer sample group. That is, it belongs to different clusters (or groups, subgroups) in different cancers, and even in the genera shared between cancers, their commonalities in the clusters are limited. This means that the role of a certain bacterial genus in one cancer cannot be simply generalized to all cancers, but it is necessary to specifically analyze its enrichment pattern and mechanism of action in each cancer.

[0139] Next, pathway enrichment analysis was performed on each cluster, and it was found that in clusters of the same sequence, the metabolic profiles remained consistent in different types of cancer. Specifically, the pathways enriched in the first cluster (G1) were mainly related to amino acid metabolism and carbohydrate metabolism. The pathways enriched in the second cluster (G2) were mainly related to carbohydrate metabolism. The pathways enriched in the third cluster (G3) were mainly involved in translation and amino acid metabolism.

[0140] Moreover, it is known that cysteine ​​and methionine metabolism have significant changes in breast cancer samples. In the present invention, it is significantly enriched in breast cancer-cluster 3 (BC-G3), and similarly, the pathway is also significantly enriched in gastric cancer-cluster 3 (GC-G3) and colorectal cancer-cluster 3 (CRC-G3). In other words, even if the clusters of different cancers are different in the composition of the taxonomic units, they are conservative / stable at the metabolic function level. For the impact of intestinal microorganisms on cancer, especially the impact on breast cancer, these key metabolic pathways should be paid attention to.

[0141] In general, although each sample group involves different microbial genera, within the cluster corresponding to the breast cancer sample group, the metabolic function of the intestinal microbiome is similar to the results of the gastric cancer sample group and the colorectal cancer sample group, which indicates that the effect of intestinal microorganisms on cancer is conservative at the metabolic level, and these functions may play a more important role in the development of cancer, which further illustrates the feasibility of the above method of the present invention.

[0142] S124 predicts the functional modules of the significant bacterial genera, and extracts similar functional modules based on the first clustering result;

[0143] When analyzing 16S rRNA sequencing data, we can usually only identify the distribution and number of species, but cannot obtain the specific functions of a species. Therefore, when understanding a genus of bacteria, we need to know what kind of environment they grow in and which genes and molecules are involved.

[0144] In one embodiment of the present specification, S124-1 predicts the functional modules of the significant bacterial genus through picrust to obtain the first functional information of each cluster, wherein the first functional information includes: all functional modules corresponding to the cluster. Based on this, the genetic type of each cluster is determined, and this information is matched with the known gene information to infer the ecological role and metabolic function of the cluster.

[0145] Specifically, characteristic information of significant bacterial genera corresponding to the original sample group is obtained;

[0146] The feature information includes 16S rRNA sequencing data of significant bacterial genera. The format of the original sequencing data is FASTQ format.

[0147] The reads obtained by sequencing are clustered to obtain several target clustering results.

[0148] The target clustering result can be an operational taxonomic unit (OTU) or an amplicon sequence variant (ASV).

[0149] In one embodiment of the present specification, paired-end reads are preprocessed using Trimmomatic software to detect and truncate sequences containing ambiguous bases (N), and low-quality sequences with an average quality score below 20 are truncated using a sliding window trimming method.

[0150] After trimming, paired-end reads were assembled using FLASH software. Assembly parameters were: minimum overlap of 10 bp, maximum overlap of 200 bp, and maximum mismatch rate of 20%.

[0151] The sequences were further denoised using QIIME software: sequences with ambiguity, homologous sequences or lengths less than 200 bp were discarded; sequences with more than 75% base quality greater than Q20 were retained; and sequences with chimeras were detected and removed.

[0152] The clean sequences were processed for primer sequence removal and clustering to generate operational taxonomic units (OTUs), and a 97% similarity cutoff was performed using Vsearch software.

[0153] The representative sequence of each OTU was selected using the QIIME software package. All representative sequences were annotated and blasted against the Silva database version 123 (or Greengens) (16s / 18s rDNA) using the RDP classifier (confidence threshold of 70%). All representative sequences were annotated and blasted against the Unite database (ITSs rDNA).

[0154] The target clustering results are compared with the species in the reference genome database; through the comparison, the species that each target clustering result (OTU / ASV) most likely corresponds to can be determined, and the genome information of the species can be obtained.

[0155] The reference genome database contains genome information of a large number of known species, including their gene composition and functional annotations.

[0156] Based on the comparison results, the genes that each target cluster result (OTU / ASV) may carry are predicted. Specifically, this is achieved by searching the genome information of the corresponding species in the reference genome database and assuming that the target cluster result (OTU / ASV) has a similar genetic composition to the species.

[0157] The predicted genes were annotated to functional categories in databases such as KEGG.

[0158] KEGG is a biological information database that contains a large number of known genes and genomes. The functional categories in it can help us understand the role of genes in biological processes.

[0159] By summarizing the functional annotations of all target clustering results (OTU / ASV) in each significant bacterial genus, the functional composition of the microbial community in the cluster can be obtained.

[0160] S124-2 searching for a common functional module among clusters in the same original sample group as a first functional module corresponding to the original sample group;

[0161] In one embodiment of the present specification, for different clusters in the same original sample group, the one-way variance is used to find the different functional modules, and then the functional modules shared by all clusters in the original sample group are determined, and the shared functional modules are used as the first functional modules corresponding to the original sample group. That is, several first functional modules corresponding to the breast cancer sample group, several first functional modules corresponding to the gastric cancer sample group, and several first functional modules corresponding to the colorectal cancer sample group are obtained.

[0162] S124-3 searching for a first functional module common to each of the original sample groups as a second functional module common to all the original sample groups;

[0163] In one embodiment of the present specification, for different original sample groups, one-way variance is used to find the first functional module of the difference, and then the first functional module common to the original sample groups is determined as the second functional module common to all original sample groups, that is, the first functional module common to the breast cancer sample group, the gastric cancer sample group and the colorectal cancer sample group is found as the second functional module.

[0164] S124-4 collects preset function module information, compares the preset function module information with the second function module, and uses the overlapping second function module as the similar function module.

[0165] The preset functional module information includes KEGG functional module information of humans.

[0166] The KEGG functional module information of human beings is collected, the KEGG functional module information of human beings is compared with the shared second functional module, and the shared functional module between the preset functional module information and the second functional module is taken as a similar functional module.

[0167] S125 searches for a path corresponding to a similar functional module according to the similar functional module, and uses the path as the key path. Figure 4 The right side of the figure is the critical path.

[0168] S13 collects second sample information, where the second sample information includes gene data and survival data of historical breast cancer patients; searches for target genes through the gene data of historical breast cancer patients and the key pathway;

[0169] The genetic data of historical breast cancer patients include: transcriptome sequencing data of breast cancer in the TCGA database (TCGA-BRCA RNA-seq).

[0170] Screen out genes related to key pathways from the genetic data of historical breast cancer patients as target genes; summarize all target genes as a target gene set;

[0171] The target gene is a gene significantly correlated with the survival rate. In one embodiment of the present specification, the target gene is as shown in Table 1.

[0172]

[0173]

[0174]

[0175] (Table 1)

[0176] S14 constructs and trains a prognosis evaluation model based on the target gene and the second sample information; S141 constructs a training set based on the target gene and the survival data of the historical breast cancer patients.

[0177] According to the gene data of the historical breast cancer patients in the second sample information, the expression values ​​of all target genes of the historical breast cancer patients are searched, and the target gene expression profiles of the historical breast cancer patients are obtained by summarizing.

[0178] The target gene expression profile and corresponding survival data of the same historical breast cancer patient are used as a basic sample. The basic samples are aggregated to obtain a training set.

[0179] Survival data include intervals associated with event status. Event status includes, but is not limited to, death, recurrence, and last follow-up. Intervals associated with event status include: the interval from a certain starting point (such as the start of diagnosis, the start of treatment, postoperatively, etc.) to the occurrence of the event status (such as death, recurrence, last follow-up, etc.).

[0180] In one embodiment of the present specification, the survival data includes: survival time and / or recurrence time. The survival time is the interval from surgery to death; the recurrence time is the interval from surgery to recurrence. The units of the intervals of each basic sample are unified.

[0181] The second basic sample is preprocessed, and the preprocessing includes: processing missing values ​​for the basic sample, processing outliers for the basic sample, and standardizing the basic sample to ensure that the training data and the test / application data are on the same scale.

[0182] S142 constructs a prognostic evaluation model;

[0183] The prognostic evaluation model is preferably a neural network model, a DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm model, or a COX regression algorithm model.

[0184] S143 uses the target gene expression profiles of historical breast cancer patients as input features and inputs them into the prognostic assessment model for training.

[0185] In one embodiment of the present specification, the prognostic assessment model is a neural network model.

[0186] The neural network model is preferably a BP neural network, and the neural network includes an input layer, a hidden layer, and an output layer.

[0187] Among them, the input layer: the number of neurons is equal to the number of target genes.

[0188] Hidden layer: Select 1 hidden layer with 12 neurons. Use ReLU (Rectified Linear Unit) as the activation function to solve the gradient vanishing problem and speed up training.

[0189] Output layer: Depending on the type of survival data (such as survival time or recurrence time), the output layer can have one or more neurons. If the output is a continuous value (such as survival time), an activation function is usually not required, or a linear activation function is used.

[0190] (1) The neural network is trained using basic samples; the output results are compared with the survival data of historical breast cancer patients, and the model is calibrated based on the loss function.

[0191] In one embodiment of the present specification, network weights and biases are initialized, usually using random small values ​​or a specific initialization method (such as He initialization). The pre-processed expression value of the target gene is received through the input layer; the hidden layer performs calculations and applies activation functions; and the output layer generates prediction results.

[0192] Huber Loss is used as the loss function to calculate the loss between the predicted results and the actual survival data. According to the gradient of the loss function, the chain rule is used to calculate the gradient of each layer parameter and update the weights and biases.

[0193] An optimization algorithm is used to improve the performance of the model. The optimization algorithm is preferably a genetic algorithm (GA).

[0194] Of course, you can also use an independent validation set (different from the training set) to evaluate the performance of the model.

[0195] In another embodiment of the present specification, the prognostic evaluation model is a DBSCAN clustering algorithm model, and the network structure, hyperparameters or training strategy are adjusted according to the verification results, and then the model is retrained.

[0196] (1) Determine the neighborhood size (eps) and the minimum number of points (minPts). The neighborhood size (eps) is preferably 18, and the minimum number of points (minPts) is preferably 35.

[0197] (2) The DBSCAN algorithm is used to cluster the basic samples and divide them into core points, boundary points and noise points.

[0198] Specifically, all basic samples are marked as unvisited. Select an unvisited basic sample p and perform the following operations on the basic sample p:

[0199] Calculate the ε neighborhood of the basic sample p. If there are at least MinPts basic samples in the ε neighborhood, create a new cluster and mark the basic sample p as a core point; otherwise, mark the basic sample p as a noise point.

[0200] Find all base samples that are density-reachable from base sample p: For each base sample q in the ε neighborhood of base sample p (including base sample p itself, if base sample p is a core point): If base sample q has not been visited, mark base sample q as visited and check the ε neighborhood of base sample q. If there are at least MinPts points in the ε neighborhood of base sample q, then base sample q is also considered a core point and continue to look for points that are density-reachable from base sample q. Otherwise, base sample q is considered a boundary point.

[0201] Repeat the above steps until no new base samples can be added to any cluster.

[0202] Mark the remaining unvisited basis samples: All basis samples that remain unvisited at the end of the clustering process are considered as noise points.

[0203] (3) Analyze the survival data of historical breast cancer patients in different clusters to observe whether there are significant differences. Use KM survival curves to compare the survival of historical breast cancer patients in different clusters.

[0204] Survival data include: patient survival time. Based on the average survival time of historical breast cancer patients in different clusters, the risk characteristics of each cluster are determined.

[0205] In one embodiment of the present specification, a number of default risk features are pre-set in order based on the length of the average survival time. According to the average survival time of historical breast cancer patients in different clusters, the clusters are sorted from long to short according to the average survival time, and survival risk features are assigned to each cluster based on the number of clusters. For example, if there are 3 clusters, according to the average survival time from long to short, the risk features of each cluster are: low, medium, and high.

[0206] In another embodiment of the present specification, the clusters are sorted from long to short according to the average survival time; and each cluster is assigned an increasing serial number as a risk feature, and the smaller the serial number, the lower the survival risk. For example, if there are 5 clusters, the clusters are sorted from long to short according to the average survival time, and the risk features of each cluster are: level 1, level 2, level 3, level 4, level 5, and the survival risk increases from level 1 to level 5.

[0207] In yet another embodiment of the present specification, the cluster with the shortest average survival time is identified as a high-risk group, and the other clusters are identified as average-risk groups.

[0208] Use a validation set or an independent dataset to verify the accuracy of clustering results and prognostic assessment.

[0209] In yet another embodiment of the present specification, the prognostic evaluation model is a COX regression algorithm model.

[0210] Specifically, statistical software (such as R's survival package and Python's lifelines library) is used to fit the prognostic evaluation model (Cox regression model); the expression values ​​of the target genes are used as independent variables (predictors) and the survival data are used as dependent variables (response variables). A model coefficient is assigned to each target gene.

[0211] Assess model assumptions, such as proportional hazards assumption, by plotting residuals, etc. Use cross-validation (e.g., k-fold cross-validation) to assess model stability and predictive performance. Check the model's C-index (consistency index) to assess the predictive performance of the prognostic assessment model.

[0212] In other embodiments of the present specification, a prognostic evaluation model may also be constructed by clustering and COX regression.

[0213] (1) performing a second clustering on the basic sample by using a second clustering model to obtain a second clustering result; the second clustering result includes a plurality of clusters;

[0214] Specifically, the second clustering model is preferably a distance-based k-means method. The k value (number of clusters) is determined. In one embodiment of the present specification, the k value is preferably 4. That is, the second clustering result includes four clusters: a first cluster (Cluster 1), a second cluster (Cluster 2), a third cluster (Cluster 3), and a fourth cluster (Cluster 4).

[0215] The target gene expression profile (expression values ​​of all target genes) of historical breast cancer patients is used as input data. K sample points are randomly selected as the initial cluster centers, where the sample points are quantitative representations of the gene expression profile. The cluster center is a certain "average" state of the expression values ​​of the basic samples in the cluster on each target gene.

[0216] The distance between each base sample and the cluster center is calculated and assigned to the nearest cluster center.

[0217] According to the expression values ​​of all target genes of the basic samples in each cluster, the cluster center is recalculated to determine the mean expression value of all target genes in the cluster.

[0218] Repeat the above steps until the cluster center no longer changes significantly or the preset number of iterations is reached.

[0219] Further analysis is performed based on the basic samples in the cluster. Identify the unique genes (differential genes) in each cluster. Determine the functions of differential genes and the biological processes involved through functional analysis (such as GO enrichment analysis, KEGG pathway analysis, etc.). For example, determine the differential genes and functions between any two clusters; determine the differential genes and functions between each cluster and other clusters; determine the differential genes and functions of triple-negative and non-triple-negative samples in the second cluster. In this manual, the independent and significant influencing factors of each cluster are determined by multivariate Cox analysis. By comparing the gene expression of the four clusters and the characteristics of the Hallmark pathway, the significant differences in gene expression and pathway activity between different target clusters are determined. For example, in the fourth cluster, the expression of Hallmark_p53_pathway is higher, while the expression in the first cluster is relatively low. In addition, in the second cluster, the expression of Hallmark_kras_signaling_up, Hallmark_inflammatory_response, and Hallmark_tnfa_signaling_via_nfkb is relatively high.

[0220] (2) Combining COX regression, summarizing all the survival data in each cluster to determine the prognosis result corresponding to the cluster.

[0221] In one embodiment of the present specification, a risk feature can be assigned to each cluster through the Cox regression model, and the risk feature represents the survival risk level of the cluster relative to other clusters. Specifically, the risk feature can be a risk score, wherein the higher the risk score, the worse the prognosis effect corresponding to the cluster.

[0222] The following combination Figure 5-6 , which further illustrates the interpretability of this prognostic assessment model.

[0223] like Figure 5 As shown, (A) is the flow chart of TCGA-BRCA dataset analysis. Based on the importance of the identified pan-cancer pathways, genes related to these pathways were selected and target genes were screened. (B) is the survival curve of the four clusters of breast cancer. Their prognostic results are significantly different, among which the fourth cluster shows the best prognosis and the fourth cluster has the worst prognosis. (CE) is the proportion and chi-square test of traditional molecular subtypes, stages and triple-negative breast cancer status in each breast cancer cluster. (EI) is a box plot of TMB, Aneuploidy score, variant genome portion and MSIsensor score at the gene level for breast cancer clusters. The second cluster has the highest score, while the fourth cluster, on the contrary, has the lowest score. (JM) is a box plot of CD4, CD8, neutrophils and myeloid dendritic cells at the immune level for the four clusters of breast cancer. The second cluster is the most abundant among all clusters. ****P<0.0001. ***P<0.001. **P<0.01. *P<0.05. ns P>0.05.

[0224] like Figure 6As shown, (A) is a violin plot depicting the distribution of scores in the four clusters. (B) is the survival curve of the low group and the high group. Among them, the high group showed worse survival results. (C) is the predictive value of the score in the TCGA-BRCA cohort (AUC: 0.857, 0.802, 0.634 and 0.671; 90-day, 180-day and 365-day and 3-year overall survival rates). (D) is a volcano map of differentially expressed genes between the high group and the low group, with genes significantly upregulated in the high group in red and genes significantly upregulated in the low group in blue. (E) is a bar chart of differential pathways between the high group and the low group. Pathways related to cancer are significantly enriched in the high group, accompanied by significantly higher scores. (F) is a comparison of immune cell expression in the second cluster between the high group and the low group. (GI) is a box plot of the high group and the low group between the variant genomic fraction, TMB, MSIMANTIS score and TMB nonsense mutations at the gene level for the four clusters of breast cancer. The high-score group in the second cluster had significantly higher values ​​at the gene level, indicating that high-score patients had a greater mutation burden. ***P<0.001. **P<0.01. *P<0.05.

[0225] S2 obtains basic information of the user; the basic information includes the expression value of the target gene;

[0226] In one embodiment of the present specification, a tissue sample of a user is first obtained, and then gene sequencing is performed on the tissue sample of the user to extract the gene sequence and obtain the expression value of the target gene.

[0227] To improve accuracy, the expression of the target gene can also be normalized.

[0228] S3 substitutes the basic information into the prognosis evaluation model to obtain the prognosis evaluation result.

[0229] In one embodiment of the present specification, the prognosis evaluation model is a neural network model. The target gene and the corresponding expression value of the target gene are extracted from the basic information; the expression value of the target gene is input into the prognosis evaluation model to obtain the predicted survival time, and the predicted survival time is used as the prognosis evaluation result.

[0230] In another embodiment of the present specification, the prognosis evaluation model is a DBSCAN clustering algorithm model. The target gene and the corresponding expression value of the target gene are extracted from the basic information; the expression value of the target gene is input into the prognosis evaluation model, the cluster to which the user belongs is determined, and the risk feature corresponding to the cluster is used as the prognosis evaluation result.

[0231] In yet another embodiment of the present specification, the prognostic evaluation model is a COX regression model.

[0232] Extract target genes and corresponding target gene expression values ​​from basic information;

[0233] Match the model coefficients corresponding to each target gene;

[0234] The linear prediction value (Linear Predictor) is calculated based on the expression value of the target gene and the model coefficient. The expression value of each target gene is multiplied by its corresponding model coefficient, and then all the products are added together. In one embodiment of the present specification, specifically, Linear Preictor = β1×Expr1+β2×Expr2+…+β n ×Expr n ; Among them, β i is the model coefficient of the ith target gene, Expr i is the expression value of the target gene in the user's first sample. n is the total number of target genes.

[0235] The linear prediction value was used as the risk score. The risk score was negatively correlated with the prognostic effect.

[0236] In other embodiments of the present specification, the prognostic evaluation model includes clustering and COX regression. The expression value of each target gene is input into the second clustering model, and the distance between the expression value of the user's target gene and each cluster center is calculated by the second clustering model. According to the calculated distance, the user is assigned to the cluster represented by the nearest cluster center to determine the cluster to which the user belongs; the risk characteristics corresponding to the cluster to which the user belongs are used as the prognostic evaluation result.

[0237] Figure 7 A schematic diagram of a prognosis evaluation system for breast cancer provided in an embodiment of this specification, the system comprises:

[0238] Model building module 701;

[0239] The information collection module 702 is used to obtain basic information of the user; the basic information includes the expression value of the target gene;

[0240] The prognosis evaluation module 703 is used to substitute the basic information into the prognosis evaluation model to obtain a prognosis evaluation result.

[0241] Optionally, the model building module 701 includes:

[0242] A first collection submodule is used to collect first sample information and second sample information, wherein the first sample information includes sequencing data of the intestinal microbiome; and the second sample information includes gene data and survival data of historical breast cancer patients;

[0243] A key path determination submodule, used to determine the key path according to the first sample information;

[0244] A target gene search submodule, used to search for target genes through the gene data of the historical breast cancer patients and the key pathways;

[0245] A model building submodule, used to build and train a prognosis evaluation model based on the target gene and the second sample information;

[0246] Optionally, the key path determination submodule includes:

[0247] A sample group construction unit, configured to construct a plurality of original sample groups according to the first sample information;

[0248] A significant bacterial genus determination unit, used to determine the significant bacterial genus corresponding to the original sample group;

[0249] A first clustering unit is used to perform first clustering on the significant bacterial genus corresponding to each of the original sample groups in combination with the characteristic information of the significant bacterial genus to obtain a first clustering result;

[0250] A function prediction unit, used to predict the function modules of significant bacterial genera and extract similar function modules in combination with the first clustering result;

[0251] The key path determining unit is used to search for paths corresponding to similar function modules according to the similar function modules as the key paths.

[0252] Optionally, the significant bacterial genus determination unit includes:

[0253] A comparison subunit, used to compare the control sample and the experimental sample in the same original sample group to obtain the different bacterial genera corresponding to the original sample group;

[0254] An importance calculation subunit, used for calculating the importance of each of the differential bacterial genus according to the characteristic information of the differential bacterial genus;

[0255] The average value calculation subunit is used to calculate the importance average value according to the importance of each differential bacterial genus;

[0256] The significant bacterial genus determination subunit is used to screen out the significant bacterial genus corresponding to the original sample group from the differential bacterial genus based on the average importance value.

[0257] Optionally, the first clustering unit includes:

[0258] A first clustering model building subunit is used to build a first clustering model and determine the number of nodes and the weight vectors of the nodes;

[0259] A similarity calculation subunit, used to calculate the similarity between each of the significant bacterial genus and the weight vectors of all the nodes;

[0260] An allocating subunit, used to allocate the significant bacterial genus to the most similar node;

[0261] All significant genera assigned to the same node were considered to belong to the same cluster.

[0262] Optionally, the first clustering result includes a plurality of clusters;

[0263] Optional, function prediction unit, including:

[0264] A first prediction subunit is used to predict the functional modules of the significant bacterial genus to obtain first functional information of each cluster, wherein the first functional information includes: all functional modules corresponding to the cluster;

[0265] A first screening subunit, configured to search for a common functional module among clusters in the same original sample group as a first functional module corresponding to the original sample group;

[0266] A second screening subunit is used to find the first functional module common to each of the original sample groups as the second functional module common to all the original sample groups;

[0267] The third screening subunit is used to collect preset function module information, compare the preset function module information with the second function module, and use the overlapping second function modules as the similar function modules.

[0268] Optional, model building submodules, including:

[0269] A second clustering unit, used for performing a second clustering on the basic samples by using a second clustering model to obtain a second clustering result; the second clustering result includes a plurality of clusters;

[0270] The summarizing unit is used to summarize all the survival data in each of the clusters to determine the prognosis result corresponding to the cluster.

[0271] The functions of the device in the embodiment of the present invention have been described in the above method embodiment, so for details not covered in the description of this embodiment, please refer to the relevant description in the above embodiment, and no further description will be given here.

[0272] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0273] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0274] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0275] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0276] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for evaluating the prognosis of breast cancer, characterized in that: include: Collecting first sample information and second sample information, wherein the first sample information includes sequencing data of the intestinal microbiome; the second sample information includes gene data and survival data of historical breast cancer patients; determining a key pathway based on the first sample information; and searching for a target gene through the gene data of the historical breast cancer patients and the key pathway; Based on the target gene and the second sample information, constructing and training a prognosis evaluation model; Get basic information of the user; The basic information includes the expression value of the target gene; Substituting the basic information into the prognosis evaluation model, a prognosis evaluation result is obtained.

2. A method for evaluating the prognosis of breast cancer according to claim 1, characterized in that: The determining the key path according to the first sample information includes: constructing a plurality of original sample groups according to the first sample information; Determine the significant bacterial genus corresponding to the original sample group; In combination with the characteristic information of the significant bacterial genus, a first clustering is performed on the significant bacterial genus corresponding to each of the original sample groups to obtain a first clustering result; Predicting the functional modules of significant bacterial genera, and extracting similar functional modules based on the first clustering results; According to the similar functional modules, a path corresponding to the similar functional modules is searched and used as the key path.

3. A method for evaluating the prognosis of breast cancer according to claim 2, characterized in that: Determining the significant bacterial genus corresponding to the original sample group includes: Comparing the control sample with the experimental sample in the same original sample group to obtain the differential bacterial genera corresponding to the original sample group; Calculating the importance of each of the differential bacterial genus according to the characteristic information of the differential bacterial genus; According to the importance of each differential bacterial genus, the average importance was calculated; Based on the average importance value, the significant genera corresponding to the original sample group were screened out from the differential genera.

4. A method for evaluating the prognosis of breast cancer according to claim 2, characterized in that: The first clustering of the significant bacterial genus corresponding to each of the original sample groups is performed in combination with the characteristic information of the significant bacterial genus to obtain a first clustering result, including: Constructing a first clustering model, determining the number of nodes and the weight vectors of the nodes; Calculate the similarity between each of the significant bacterial genus and the weight vectors of all the nodes; Assigning the significant bacterial genus to the most similar node; All significant genera assigned to the same node were considered to belong to the same cluster.

5. A method for evaluating the prognosis of breast cancer according to claim 2, characterized in that: The first clustering result includes a plurality of clusters; The predicting of the functional modules of the significant bacterial genera and extracting similar functional modules in combination with the first clustering result include: Predicting the functional modules of the significant bacterial genus to obtain first functional information of each cluster, wherein the first functional information includes: all functional modules corresponding to the cluster; Finding a common functional module among clusters in the same original sample group as a first functional module corresponding to the original sample group; Finding a first functional module common to each of the original sample groups as a second functional module common to all the original sample groups; Collecting preset function module information, comparing the preset function module information with the second function module, and taking the overlapping second function modules as the similar function modules.

6. A method for evaluating the prognosis of breast cancer according to claim 1, characterized in that: The constructing and training a prognosis evaluation model based on the target gene and the second sample information includes: Performing a second clustering on the basic samples by using a second clustering model to obtain a second clustering result; the second clustering result includes a plurality of clusters; All the survival data in each of the clusters are aggregated to determine the prognosis result corresponding to the cluster.

7. A prognosis evaluation system for breast cancer, characterized in that: include: A model building module is used to collect first sample information and second sample information, wherein the first sample information includes sequencing data of the intestinal microbiome; the second sample information includes gene data and survival data of historical breast cancer patients; determine a key pathway based on the first sample information; and search for a target gene through the gene data of the historical breast cancer patients and the key pathway; Based on the target gene and the second sample information, constructing and training a prognosis evaluation model; Information collection module, used to obtain basic information of users; The basic information includes the expression value of the target gene; The prognosis evaluation module is used to substitute the basic information into the prognosis evaluation model to obtain the prognosis evaluation result.

8. An electronic device, wherein: The electronic device includes: processor; and, A memory storing computer executable instructions which, when executed, cause the processor to perform the method according to any one of claims 1-6.

9. A computer-readable storage medium, wherein: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method of any one of claims 1 to 6 is implemented.

Images