Disease prediction method, system, computer device and medium

Abstract

The present invention provides a disease prediction method, system, computer device and medium, and the method includes the following steps: obtaining mutation site information of chromosomes; extracting data features from a chromosome database according to the mutation site information, binarizing the feature data to obtain pathogenicity prediction features; inputting the pathogenicity prediction features into a pre-generated disease prediction model, and outputting pathogenicity data; obtaining pathogenicity prediction results according to the pathogenicity data; the scheme integrates biological knowledge and statistical knowledge, and greatly mines the relationship between mutation point information and target diseases, which is conducive to improving the accuracy of disease prediction, and has stronger applicability and stronger scalability, and can provide doctors and patients with accurate references for judging the pathogenicity of a mutation site, and can be widely used in the field of bioinformatics technology.

Description

Technical Field

[0001] The present invention belongs to the field of bioinformatics, and in particular to a disease prediction method, system, computer equipment and storage medium. Background Art

[0002] With the update and iteration of genome sequencing technology, the sequencing cost is getting lower and lower, making the commercialization of whole genome sequencing services possible. This means that in the future, everyone can master their own unique genetic code through commercial gene sequencing technology. However, the popularity of this sequencing technology brings great challenges to downstream data analysis and interpretation. Because the value that simple sequencing can bring is limited, each patient or individual hopes to understand their own physical condition and even the diagnosis and treatment plan of the disease through sequencing technology. Therefore, the interpretation of genomic information is very important.

[0003] Although there are some solutions or methods based on genome sequencing in the prior art, the existing technical solutions for disease prediction based on genome usually have shortcomings and problems such as low accuracy, weak applicability, and poor scalability. Summary of the invention

[0004] In view of this, in order to at least partially solve one of the above-mentioned technical problems, an embodiment of the present invention aims to provide a disease prediction method with higher accuracy and stronger applicability, as well as a system, computer device and storage medium capable of implementing the method.

[0005] In a first aspect, the technical solution of the present application provides a disease prediction method, the steps of which include:

[0006] Obtain the mutation site information of chromosomes;

[0007] Extracting data features from a chromosome database according to the mutation site information, and binarizing the feature data to obtain pathogenicity prediction features;

[0008] Inputting the pathogenicity prediction feature into a pre-generated disease prediction model, and outputting pathogenicity data;

[0009] A pathogenicity prediction result is obtained according to the pathogenicity data.

[0010] In a feasible embodiment of the present application, before binarizing the feature data to obtain pathogenicity prediction features, the disease prediction method further includes at least one of the following steps:

[0011] Removing data irrelevant to pathogenicity prediction from the feature data;

[0012] Filling the missing values ​​in the feature data.

[0013] In a feasible embodiment of the present application, before the step of inputting the pathogenicity prediction feature into a pre-generated disease prediction model, the disease prediction method further includes:

[0014] Acquire a number of data from a gene mutation disease database to form a first data set;

[0015] Preprocess the first data set, construct feature engineering to filter the preprocessed first data set to obtain a training data set and a test data set;

[0016] The disease prediction model is trained using the training data set, and the trained disease prediction model is adjusted using the test data set.

[0017] In a feasible embodiment of the solution of the present application, the feature engineering step of screening the preprocessed first data set to obtain a training data set and a test data set includes the following steps:

[0018] Performing data cleaning on the first data according to a preset quality value threshold and a credibility threshold;

[0019] Completing the data with missing values ​​in the first data set after data cleaning;

[0020] The first data set after the missing values ​​are filled is binarized, and the binarized data is divided into the training data set and the test data set.

[0021] In a feasible embodiment of the solution of the present application, after completing the data with missing values ​​in the first data set after data cleaning, the disease prediction method further includes the following steps:

[0022] The first data after cleaning is labeled to distinguish and obtain second data containing pathogenic mutation sites and third data containing benign mutation sites.

[0023] In a feasible embodiment of the present application, the step of obtaining a pathogenicity prediction result according to the pathogenicity data includes:

[0024] The probability interval of pathogenicity is determined by confusion matrix and receiver operating characteristic curve.

[0025] The pathogenicity prediction result is obtained according to the pathogenicity data and the pathogenicity probability interval.

[0026] In a feasible embodiment of the present application, the step of extracting data features from the chromosome database according to the mutation site information includes:

[0027] Generate an index indicator based on the mutation site information, and obtain target gene information through whole genome comparison;

[0028] The ratio of guanine to cytosine is obtained according to the base information upstream and downstream of the mutation site, and the epigenetic characteristics are obtained by matching the target gene information as the data characteristics.

[0029] In a second aspect, the technical solution of the present invention further provides a disease prediction system, the system comprising:

[0030] An information acquisition module is used to obtain the mutation site information of the chromosome;

[0031] A data processing module, used to extract data features from a chromosome database according to the mutation site information, and perform binarization processing on the feature data to obtain pathogenicity prediction features;

[0032] A model prediction module, used to input the pathogenicity prediction feature into a pre-generated disease prediction model, output pathogenicity data; and obtain a pathogenicity prediction result based on the pathogenicity data;

[0033] The data storage module is used to store chromosome information.

[0034] In a third aspect, the technical solution of the present invention further provides a disease prediction computer device, which comprises:

[0035] at least one processor;

[0036] at least one memory for storing at least one program;

[0037] When at least one program is executed by at least one processor, the at least one processor runs the method described in the first aspect.

[0038] In a fourth aspect, the technical solution of the present invention further provides a storage medium, in which a program executable by a processor is stored. When the program executable by the processor is executed by the processor, it is used to run a disease prediction method as described in the first aspect.

[0039] The advantages and beneficial effects of the present invention will be partially given in the following description, and the other parts can be understood through the specific embodiments of the present invention:

[0040] The technical solution of this application can accurately predict the pathogenicity of gene mutation sites through necessary preprocessing and feature engineering processing based on the mutation point information of chromosomes, and then through the disease prediction model pre-built by machine learning algorithms. The solution obtains an efficient and accurate disease prediction model through feature screening and data processing fitting training, integrating biological knowledge and statistical knowledge, and greatly mining the relationship between mutation point information and target diseases, which is conducive to improving the accuracy of disease prediction, and the built prediction model can be directly put into use without adaptation and adjustment, with stronger applicability and stronger expansion performance, and can provide doctors and patients with accurate reference for judging the pathogenicity of a certain mutation site. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 A flowchart of a disease prediction method provided by an embodiment of the present invention;

[0043] Figure 2 A flowchart of another disease prediction method provided by an embodiment of the present invention;

[0044] Figure 3 Flow chart of the steps of the disease prediction model training process in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limitations of the present invention. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0046] First of all, it should be explained that the technical solution of this application is based on the latest version of chromosome data, and the disease prediction model built by the machine learning algorithm further accurately predicts the pathogenicity of the gene mutation site. Although there are some pathogenicity prediction schemes and methods based on genome sequencing data and corresponding products in the prior art, they generally have problems such as low accuracy and low chromosome versions supported. There are currently two versions of chromosome data, namely GRCh37 and GRCh38. Most of the prediction models in the prior art are built based on the GRCh37 version of chromosomes, while the GRCh38 version of chromosomes can provide more information, which is crucial to improving the accuracy of genetic disease prediction. Therefore, the technical solution of this application builds a disease prediction model based on the GRCh38 version of chromosomes, and improves the accuracy of gene mutation pathogenicity prediction.

[0047] In the first aspect, Figure 1 As shown, based on the establishment of a disease prediction model based on the GRCh38 version of chromosomes, the disease prediction method provided by the technical solution of the present application may include the following main steps S100-S400:

[0048] S100, obtaining the mutation site information of the chromosome;

[0049] Specifically, in the embodiment, it is first necessary to obtain the chromosomes of the subject and the gene mutation information carried in the chromosomes. Exemplarily, the embodiment can obtain the genome information of the subject through genome sequencing, and then obtain the gene mutation information by comparing it with the chromosomes of the GRCh38 version. For example, a string of data strings obtained in the embodiment is: (chr3, 888888, C, T); this data string indicates that the base at the 888888 position of chromosome 3 of the subject is T, but the base at the 888888 position of chromosome 3 on the GRCh38 version of the genome is C, that is, the 888888 position of chromosome 3 of the subject has mutated, from C base to T base.

[0050] S200, extracting data features from the chromosome database according to the mutation site information, and binarizing the feature data to obtain pathogenicity prediction features;

[0051] Specifically, the chromosome database used in the embodiment can be a CADD database, which is a public database used to provide features for gene sites; in the embodiment, according to the mutation site information of the chromosome obtained in step S100, indexing and matching are performed in the CADD database to obtain the gene features, protein features and epigenetic features corresponding to the mutation site, and these features are binarized as pathogenicity prediction feature data to obtain a data record with a more unified format and content and a more intuitive expression. Exemplarily, in this group of features of Consequence in the embodiment, what is returned after matching by the CADD database are nonsense mutations, synonymous mutations, intron mutations, etc., and such features need to be binarized. It should be noted that in this embodiment, a total of 196 features for pathogenicity prediction are finally obtained.

[0052] In some optional embodiments, the method may include steps S210-S220 in the process of extracting data features from the chromosome database according to the mutation site information in step S200:

[0053] S210, generating index indicators according to the mutation site information, and obtaining target gene information through whole genome comparison;

[0054] S220, obtaining the ratio of guanine to cytosine based on the base information upstream and downstream of the mutation site, and obtaining epigenetic features as data features by matching the target gene information;

[0055] Specifically, in step S200 of the embodiment, data features can be extracted through the CADD database to obtain a total of 135 data features; in this process, the CADD database will use the chromosome version and mutation site as index indicators to retrieve all information features on the mutation site. First, for the input mutation site information, the CADD database can obtain the gene information where the mutation is located through whole genome alignment, and calculate the GC content within 10bp upstream and downstream of the site, that is, the ratio of guanine and cytosine; then, according to the data in the CADD database, match the epigenetic features of the site, such as acetylation, methylation, etc. If a feature that is not recorded in the database appears during the matching process of the embodiment, the feature will be returned as a null value.

[0056] In some optional embodiments, before the method performs binarization processing on the feature data to obtain the pathogenicity prediction feature, the disease prediction method further includes at least one of steps S220 and S230:

[0057] S220, removing data irrelevant to pathogenicity prediction in feature data;

[0058] S230, filling missing values ​​in the feature data;

[0059] Specifically, Figure 2 As shown, in the embodiment, the feature data extracted from the CADD database contains a class of features with unknown meanings, such as gene ID, which needs to be deleted; it should be noted that in biology, gene ID is only used as an identifier to facilitate the distinction between different genes, so it is meaningless for the judgment of the pathogenicity of the disease in the embodiment. Therefore, for this type of feature, the embodiment will delete it in step S220. For other feature data containing missing values, such as EncodeH3K27me3-sum, the content recorded in the feature data is used to characterize whether the site is trimethylated. If this feature data is missing, it can be filled with a default value in the embodiment.

[0060] S300, inputting the pathogenicity prediction feature into a pre-generated disease prediction model, and outputting pathogenicity data;

[0061] Specifically, the embodiment inputs the pathogenicity prediction features obtained after binarization in step S200 into a disease prediction model that has been pre-trained and optimized to obtain a set of pathogenicity data.

[0062] S400, obtaining a pathogenicity prediction result according to the pathogenicity data;

[0063] Specifically, the pathogenicity data obtained according to the disease prediction model is a value in the range of 0 to 1, where 0 indicates that the mutation site has no pathogenicity, and 1 indicates that the mutation site is 100% pathogenic. Therefore, the closer the value is to 0, the lower the possibility of pathogenicity, and the closer it is to 1, the higher the possibility of pathogenicity.

[0064] In some optional embodiments, obtaining a pathogenicity prediction result according to pathogenicity data may include steps S410 and S420:

[0065] S410, determining the pathogenicity probability interval through a confusion matrix and a receiver operating characteristic curve;

[0066] S420, obtaining a pathogenicity prediction result according to the pathogenicity data and the pathogenicity probability interval;

[0067] Specifically, the pathogenicity data obtained in the embodiment can determine the optimal threshold range through the confusion matrix and the ROC curve. The final prediction result is: when the returned result is 0-0.02, it means benign, 0.02-0.1 means possibly benign, 0.1-0.8 means possibly pathogenic, and 0.8-1 means pathogenic.

[0068] In some optional embodiments, before the step of inputting the pathogenicity prediction feature into the pre-generated disease prediction model, the process of building the disease prediction model may include S240-S260:

[0069] S240, acquiring a number of data from a gene mutation disease database to form a first data set;

[0070] Among them, the first data set in the embodiment is extracted from a public resource database and can be used as a data set for model training; illustratively, in the embodiment, a data set can be extracted from the ClinVar database, which is a public resource database, and the data records therein are a database of reported gene mutations that cause disease.

[0071] S250, preprocessing the first data set, constructing feature engineering to filter the preprocessed first data set to obtain a training data set and a test data set;

[0072] Specifically, the embodiment first performs necessary data preprocessing on the first data set, deletes some meaningless data and noise data, and fills in some missing values; the obtained low-noise and relatively complete data set is then used to obtain a training data set and a test data set through feature engineering. In the embodiment, the training data set and the test data set contain approximately 330,000 data and 20,000 data, respectively.

[0073] S260, training the disease prediction model using the training data set, and adjusting the parameters of the trained disease prediction model using the test data set;

[0074] Specifically, Figure 3As shown, the embodiment uses the training data set to train the XGBoost algorithm. In the process of XGBoost algorithm training, 196 sets of features are used as X values ​​and the label values ​​are used as y values ​​for function fitting. Then the fitting function is tested using the test data set, that is, the test data set is input into the fitting function, and the obtained predicted value is compared with the actual value. After the disease prediction model training in the embodiment is completed, the model needs to be adjusted according to the test data set to optimize the performance of the model. Among them, the 5-fold cross-validation method is used in the embodiment to optimize the model. The main parameters are max_depth, n_estimators and learning_rate. Given a certain range of values ​​for these three parameters, the embodiment uses the five-fold cross-validation method to select specific parameters so that the prediction model minimizes the average mean difference between the predicted results and the actual results in the test data set. Finally, the embodiment obtains max_depth=5, n_estimators=422, and learning_rate=0.1 under the premise of taking into account the model operation speed and accuracy.

[0075] In some optional embodiments, the process of constructing feature engineering in step S250 to obtain a training data set and a test data set from the preprocessed first data set may include steps S251-S253:

[0076] S251, performing data cleaning on the first data according to a preset quality value threshold and a credibility threshold;

[0077] S252, completing data with missing values ​​in the first data set after data cleaning;

[0078] S253, binarizing the first data set after the missing values ​​are filled, and dividing the binarized data into a training data set and a test data set;

[0079] Specifically, since there is a lot of data in the ClinVar database, and some data have problems of low quality and low credibility, the collected data needs to be screened and cleaned; then the obtained data features are screened through feature engineering in the embodiment, wherein the operations include deleting certain low-value features, defining default values ​​to fill in missing data, and binarizing certain features, and finally a data set containing 196 features is obtained, that is, there are 196 lists in the data set. Further, the training data set and the test data set obtained through feature engineering contain approximately 330,000 data and 20,000 data, respectively.

[0080] In some optional embodiments, after completing the data with missing values ​​in the first data set after data cleaning, the disease prediction method of the embodiment further includes step S254:

[0081] S254, labeling the cleaned first data to distinguish and obtain second data containing pathogenic mutation sites and third data containing benign mutation sites;

[0082] Specifically, the embodiment can also extract features from the cleaned data through the CADD database and add labels to the data, with the label value of 1 for pathogenic mutation sites and 0 for benign mutation sites.

[0083] In a second aspect, the technical solution of the present invention further provides a disease prediction system, the system comprising:

[0084] An information acquisition module is used to obtain the mutation site information of the chromosome;

[0085] A data processing module is used to extract data features from the chromosome database according to the mutation site information, and perform binary processing on the feature data to obtain pathogenicity prediction features;

[0086] A model prediction module is used to input pathogenicity prediction features into a pre-generated disease prediction model, and output pathogenicity data; and obtain pathogenicity prediction results based on the pathogenicity data;

[0087] The data storage module is used to store chromosome information.

[0088] In a third aspect, an embodiment of the present application provides a computer device for generating disease predictions, comprising: at least one processor; at least one memory for storing at least one program; when the at least one program is executed by at least one processor, the at least one processor runs the method in the first aspect.

[0089] In one embodiment, 20 pieces of mutation data that have been evaluated and confirmed are randomly selected from the ClinVar database, and it is ensured that none of the 20 pieces of data appear in the training data set. Among them, 10 pieces are pathogenic mutations. 10 pieces are benign mutations, and then predictions are made using the prediction model described in the above embodiments; Table 1 shows the pathogenicity results obtained by prediction of a certain embodiment of the technical solution of this application:

[0090] Table 1

[0091]

[0092]

[0093] Among them, #Chrom, Pos, Ref, Alt are listed as input data, Label is the pathogenicity result given by the ClinVar database, and Prediction and Score are the results given by the prediction model of the embodiment.

[0094] #Chrom: represents the chromosome number

[0095] Pos: indicates the position on the chromosome

[0096] Ref: indicates the base before mutation

[0097] Alt: indicates the base after mutation

[0098] Label: The values ​​in this column are only 1.0 and 0.0, where 1.0 means the mutation is pathogenic and 0.0 means the mutation is benign.

[0099] Prediction: This column is the result of the prediction returned by the embodiment model. There are four results, namely Pathogenic, Likely Pathogenic, Likely Benign and Benign. The corresponding Chinese meanings are pathogenic, possibly pathogenic, possibly benign, and benign.

[0100] Score: It is the score given by the prediction model based on the input data. The score ranges from 0 to 1. The closer to 1, the higher the possibility of pathogenicity, and the closer to 0, the higher the possibility of benign mutation.

[0101] Among the 10 pathogenicity results in Table 1, the results given by the model in the embodiment are consistent with the evaluated results and are highly reliable.

[0102] The results of benign mutations are shown in Table 2:

[0103] Table 2

[0104]

[0105]

[0106] In the data in Table 2, the results predicted by the model of the embodiment are basically consistent with the identification and have a high degree of credibility. However, in the second row of data, the prediction results are biased and inconsistent with the results evaluated on ClinVar. This is due to the prediction accuracy of the model, that is, the model of the embodiment can achieve an accuracy of more than 95%, but cannot achieve 100% accuracy.

[0107] From the above specific implementation process, it can be concluded that the technical solution provided by the present invention has the following advantages or strengths compared with the prior art:

[0108] The technical solution of the present application obtains an efficient and accurate disease prediction model through feature screening and data processing for fitting training, integrates biological knowledge and statistical knowledge, and greatly mines the relationship between mutation point information and target diseases, which is conducive to improving the accuracy of disease prediction. The constructed prediction model can be put into use directly without adaptation and adjustment. It has stronger applicability and scalability, and can provide doctors and patients with accurate reference for judging the pathogenicity of a mutation site.

[0109] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.

[0110] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features can be integrated into a single physical device and / or software module, or one or more functions and / or features can be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0111] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, which can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0112] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0113] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0114] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A disease prediction system, characterized in that: include: An information acquisition module is used to obtain the mutation site information of the chromosome; A data processing module, used to extract data features from the chromosome database according to the mutation site information, and perform binarization processing on the feature data to obtain pathogenicity prediction features; A model prediction module, used to input the pathogenicity prediction features into a pre-generated disease prediction model and output pathogenicity data; and obtaining a pathogenicity prediction result according to the pathogenicity data; The step of obtaining a pathogenicity prediction result according to the pathogenicity data includes: determining a pathogenicity probability interval through a confusion matrix and a receiver operating characteristic curve; obtaining the pathogenicity prediction result according to the pathogenicity data and the pathogenicity probability interval; A data storage module, used for storing chromosome information; Before the step of inputting the pathogenicity prediction feature into the pre-generated disease prediction model, the disease prediction system is further used to implement: Acquire a number of data from a gene mutation disease database to form a first data set; Preprocess the first data set, construct feature engineering to filter the preprocessed first data set to obtain a training data set and a test data set; The disease prediction model is trained through the training data set, and the parameters of the trained disease prediction model are adjusted through the test data set; The feature engineering step of filtering the preprocessed first data set to obtain a training data set and a test data set includes the following steps: Performing data cleaning on the first data according to a preset quality value threshold and a credibility threshold; Completing the data with missing values ​​in the first data set after data cleaning; After completing the data with missing values ​​in the first data set after data cleaning, the disease prediction system is further used to implement the following steps: The first data after cleaning is labeled to distinguish and obtain second data containing pathogenic mutation sites and third data containing benign mutation sites.

2. A disease prediction system according to claim 1, characterized in that: Before binarizing the feature data to obtain pathogenicity prediction features, the disease prediction system is further used to implement at least one of the following steps: Removing data irrelevant to pathogenicity prediction from the feature data; Filling the missing values ​​in the feature data.

3. A disease prediction system according to claim 1, characterized in that: The feature engineering step of filtering the preprocessed first data set to obtain a training data set and a test data set further includes the following steps: The first data set after missing values ​​are filled is binarized, and the binarized data is divided into the training data set and the test data set.

4. A disease prediction system according to claim 1, characterized in that: The step of extracting data features from the chromosome database according to the mutation site information includes: Generate an index indicator based on the mutation site information, and obtain target gene information through whole genome comparison; The ratio of guanine to cytosine is obtained according to the base information upstream and downstream of the mutation site, and the epigenetic characteristics are obtained by matching the target gene information as the data characteristics.

5. A disease prediction computer device, characterized in that: include: at least one processor; At least one memory, for storing at least one program; when the at least one program is executed by the at least one processor, the at least one processor runs the following method: Obtain the mutation site information of chromosomes; Extracting data features from a chromosome database according to the mutation site information, and binarizing the feature data to obtain pathogenicity prediction features; Inputting the pathogenicity prediction feature into a pre-generated disease prediction model, and outputting pathogenicity data; and Obtaining a pathogenicity prediction result according to the pathogenicity data; The step of obtaining a pathogenicity prediction result according to the pathogenicity data includes: determining a pathogenicity probability interval through a confusion matrix and a receiver operating characteristic curve; Obtaining the pathogenicity prediction result according to the pathogenicity data and the pathogenicity probability interval; Before the step of inputting the pathogenicity prediction feature into a pre-generated disease prediction model, the method further includes: Acquire a number of data from a gene mutation disease database to form a first data set; Preprocess the first data set, construct feature engineering to filter the preprocessed first data set to obtain a training data set and a test data set; The disease prediction model is trained through the training data set, and the parameters of the trained disease prediction model are adjusted through the test data set; The feature engineering step is to filter the preprocessed first data set to obtain a training data set and a test data set. The following steps are involved: Performing data cleaning on the first data according to a preset quality value threshold and a credibility threshold; Completing the data with missing values ​​in the first data set after data cleaning; After completing the data with missing values ​​in the first data set after data cleaning, the method further includes: The first data after cleaning is labeled to distinguish and obtain second data containing pathogenic mutation sites and third data containing benign mutation sites.

6. A disease prediction computer device according to claim 5, characterized in that: Before binarizing the feature data to obtain pathogenicity prediction features, the method further includes: Removing data irrelevant to pathogenicity prediction from the feature data; Filling the missing values ​​in the feature data.

7. A disease prediction computer device according to claim 5, characterized in that: The feature engineering step of filtering the preprocessed first data set to obtain a training data set and a test data set further includes the following steps: The first data set after missing values ​​are filled is binarized, and the binarized data is divided into the training data set and the test data set.

8. A disease prediction computer device according to claim 5, characterized in that: The step of extracting data features from the chromosome database according to the mutation site information includes: Generate an index indicator based on the mutation site information, and obtain target gene information through whole genome comparison; The ratio of guanine to cytosine is obtained according to the base information upstream and downstream of the mutation site, and the epigenetic characteristics are obtained by matching the target gene information as the data characteristics.

9. A storage medium storing a program executable by a processor, characterized in that: The processor executable program is used to execute the following method when executed by the processor: Obtain the mutation site information of chromosomes; Extracting data features from a chromosome database according to the mutation site information, and binarizing the feature data to obtain pathogenicity prediction features; Inputting the pathogenicity prediction feature into a pre-generated disease prediction model, and outputting pathogenicity data; and obtaining a pathogenicity prediction result according to the pathogenicity data; The step of obtaining a pathogenicity prediction result according to the pathogenicity data includes: determining a pathogenicity probability interval through a confusion matrix and a receiver operating characteristic curve; Obtaining the pathogenicity prediction result according to the pathogenicity data and the pathogenicity probability interval; Before the step of inputting the pathogenicity prediction feature into a pre-generated disease prediction model, the method further includes: Acquire a number of data from a gene mutation disease database to form a first data set; Preprocess the first data set, construct feature engineering to filter the preprocessed first data set to obtain a training data set and a test data set; The disease prediction model is trained through the training data set, and the parameters of the trained disease prediction model are adjusted through the test data set; The feature engineering step of filtering the preprocessed first data set to obtain a training data set and a test data set includes the following steps: Performing data cleaning on the first data according to a preset quality value threshold and a credibility threshold; Completing the data with missing values ​​in the first data set after data cleaning; After completing the data with missing values ​​in the first data set after data cleaning, the method further includes: The first data after cleaning is labeled to distinguish and obtain second data containing pathogenic mutation sites and third data containing benign mutation sites.

10. A storage medium according to claim 9, characterized in that: Before binarizing the feature data to obtain pathogenicity prediction features, the method further includes: Removing data irrelevant to pathogenicity prediction from the feature data; Filling the missing values ​​in the feature data.

11. A storage medium according to claim 9, characterized in that: The feature engineering step of filtering the preprocessed first data set to obtain a training data set and a test data set further includes the following steps: The first data set after missing values ​​are filled is binarized, and the binarized data is divided into the training data set and the test data set.

12. A storage medium according to claim 9, characterized in that: The step of extracting data features from the chromosome database according to the mutation site information includes: Generate an index indicator based on the mutation site information, and obtain target gene information through whole genome comparison; The ratio of guanine to cytosine is obtained according to the base information upstream and downstream of the mutation site, and the epigenetic characteristics are obtained by matching the target gene information as the data characteristics.

Images