Deep analysis system and method for whole genome methylation sequencing

Abstract

The invention discloses a deep analysis system and method for whole genome methylation sequencing, and belongs to the field of biological data processing systems.Data quality control is conducted by means of a quality score calculation formula, a genome library is constructed, and a whole genome methylation sequencing result is obtained. Comparing the filtered methylation sequencing data with a reference genome in a genome library through a deep learning model to obtain methylation sequencing comparison data, constructing a dynamic methylation level evaluation model, and importing the methylation sequencing comparison data into the methylation level evaluation model to perform methylation level evaluation. A methylation map generation model is established, a methylation map is generated from methylation data by using a data enhancement technology strategy, and the processing efficiency and accuracy of genome methylation sequencing data are improved.

Description

Technical Field

[0001] The present invention belongs to the field of biological data processing systems, and specifically is a deep analysis system and method for whole genome methylation sequencing. Background Art

[0002] DNA methylation is one of the important research topics in epigenetics. Whole-genome DNA methylation sequencing is regarded as the "gold standard" for DNA methylation research. It uses T4-DNA ligase to connect adapter sequences at both ends of ultrasonically interrupted genomic DNA fragments. The ligation product is treated with bisulfite to convert unmethylated cytosine C into uracil U, and then uracil U is converted into thymine T through adapter sequence-mediated PCR technology, while methylated C is not affected by bisulfite. Subsequent sequencing of the treated genomic DNA library can analyze the methylation status of the entire genome with single-base resolution, which can accurately evaluate the methylation level of a single C base and has a wide coverage range. However, existing analysis systems and methods face many challenges in processing large amounts of WGBS data, including high data storage and computational burdens, low accuracy of methylation sites, imperfect DNA methylation visualization, and lack of user-friendliness. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention proposes a deep analysis system and method for whole-genome methylation sequencing. The present invention obtains whole-genome methylation sequencing data after sequencing on a machine, uses a quality score calculation formula to perform data quality control, constructs a genome library, compares the filtered methylation sequencing data with a reference genome in the genome library through a deep learning model to obtain methylation sequencing comparison data, constructs a dynamic methylation level evaluation model, imports the methylation sequencing comparison data into the methylation level evaluation model to evaluate the methylation level, builds a methylation map generation model, uses a data enhancement technology strategy to generate a methylation map from the methylation data, and improves the processing efficiency and accuracy of genome methylation sequencing data.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A deep analysis system and method for whole genome methylation sequencing, comprising the following specific steps:

[0006] Obtain the whole genome methylation sequencing data after sequencing, and use the quality score calculation formula to control the data quality;

[0007] Construct a genomic library, and use a deep learning model to compare the filtered methylation sequencing data with the reference genome in the genomic library to obtain methylation sequencing comparison data;

[0008] Construct a dynamic methylation level assessment model, and import the methylation sequencing comparison data into the methylation level assessment model to perform methylation level assessment;

[0009] A methylation map generation model was constructed, and data enhancement technology strategies were used to generate methylation maps from methylation data.

[0010] It should be specifically noted that the acquisition of whole genome methylation sequencing data and the quality control of the data include the following specific steps:

[0011] S11, obtaining the original sequencing data generated by the whole genome methylation sequencing platform after sequencing;

[0012] S12, importing the obtained original sequencing data into a quality score calculation formula for sequencing data to calculate a quality score for the sequencing data, wherein the quality score calculation formula for sequencing data is: Among them, n is the number of sequencing data, x is the sequencing data, and w i is the weight, M(x i ) is the quality index of the sequencing data. If the quality score is greater than or equal to the threshold, it means that the original sequencing data does not need to be filtered and S2 is performed; if the quality score is less than the threshold, it means that the original sequencing data needs to be filtered.

[0013] It should be specifically explained here that the Q value threshold is set to Q≥30, and the coverage threshold is set to ≥10×.

[0014] It should be specifically noted that the comparison of the filtered methylation sequencing data with the reference genome in the genome library by the deep learning model includes the following specific contents:

[0015] S21. Construct whole genome library, target region library and transcriptome library through experiments;

[0016] S22, converting the methylation sequencing data and the reference genome into a three-dimensional tensor, and encoding the genome sequence using one-hot encoding;

[0017] S23, using Python to construct a convolutional neural network, and importing the filtered data into the convolutional neural network for comparison to identify differences in methylation status;

[0018] S24, importing the compared data into the comparison accuracy and recall calculation formula for model evaluation, wherein the calculation formula for the comparison accuracy is: The calculation formula for recall is: Among them, m is the genomic fragment on the read alignment, N is the total number of genomic fragments, TP is the number of samples correctly predicted as positive, FP is the number of samples incorrectly predicted as positive, and FN is the number of samples incorrectly predicted as negative.

[0019] It should be specifically explained here that the construction of the dynamic methylation level assessment model and the importation of the methylation sequencing comparison data into the methylation level assessment model for methylation level assessment include the following specific steps:

[0020] S31, importing the filtered data into a dynamic weighted methylation level calculation formula to calculate the weighted methylation level, wherein the calculation formula for the dynamic weighted methylation level at time t is:

[0021] Among them, W t is the dynamic weighted methylation level at time t, M i，t is the methylation status (0 or 1) of the i-th site at time t, w i，t is the dynamic weight of the ith site at time t, n is the total number of sites, where the dynamic weight is weighted based on time and biological importance, and its calculation formula is: i，t =w i，0 ×e -λt ×Z, where w i，0 is the initial weight, λ is the decay rate, and Z is the biological importance score of the gene or site; the methylation state can be expressed in binary as:

[0022]

[0023] S32. Construct a feature matrix based on the methylation status, GC content, sequencing depth, and one-hot coding of base sequence features in the database, where the calculation formula for GC content is: Among them, G j is the number of guanine bases in the jth sequence, C j is the number of cytosine bases in the jth sequence, Aj is the number of adenine bases in the jth sequence, T j is the number of thymine bases in the jth sequence. The calculation formula for sequencing depth is: D = T / L, where T is the total number of bases obtained in sequencing, and L is the base length of the target region. The calculation formula for One-Hot encoding is: Similarly, we can get S C , S G , S T . All features are integrated into a feature matrix X, The target variable is

[0024] S33. Perform model training, cross-validation and dynamic evaluation using a support vector machine model based on feature data.

[0025] It should be specifically explained here that the construction of the methylation map generation model and the use of data enhancement technology strategy to generate a methylation map from methylation data include the following specific steps:

[0026] S41. Generate a heat map of the dynamically evaluated methylation data using R language to obtain the methylation levels of different samples and under different conditions;

[0027] S42, set the window size H and step size S of the sliding window, apply the sliding window to each region in the genome, starting from the first base pair, the window range is [kS, kS+H], where k is the number of times the window moves, and calculate the weighted methylation level P of the CpG site in each window, and the calculation formula is: Where W i is the weighted methylation status of the i-th CpG site, and n is the number of CpG sites contained in the window;

[0028] S43, moving the window forward according to the set step size S until the entire genome is traversed, and finally generating a smoothed whole-genome methylation map.

[0029] A deep analysis system for whole genome methylation sequencing, which is implemented based on the above-mentioned deep analysis method for whole genome methylation sequencing, and specifically comprises:

[0030] The quality control module is used to process the quality of raw sequencing data and filter the data;

[0031] A sequence comparison module is used to compare the measured sequence with the reference sequence and generate a comparison file;

[0032] Dynamic methylation level assessment module, which is used to construct a dynamic methylation level assessment model through machine learning to dynamically assess the degree of methylation;

[0033] The methylation map generation module is used to import the dynamically evaluated methylation data into the methylation map generation model to obtain the whole genome methylation map.

[0034] An electronic device comprises: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0035] The processor executes the above-mentioned deep analysis method for whole genome methylation sequencing by calling the computer program stored in the memory.

[0036] A computer-readable storage medium, characterized in that it stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned deep analysis method for whole genome methylation sequencing.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention obtains whole-genome methylation sequencing data after sequencing on a machine, performs quality control of the data using a quality score calculation formula, constructs a genome library, compares the filtered methylation sequencing data with a reference genome in the genome library through a deep learning model to obtain methylation sequencing comparison data, constructs a dynamic methylation level assessment model, imports the methylation sequencing comparison data into the methylation level assessment model to perform methylation level assessment, builds a methylation map generation model, and uses a data enhancement technology strategy to generate a methylation map from the methylation data, thereby improving the processing efficiency and accuracy of genome methylation sequencing data. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a schematic diagram of the overall process of a deep analysis method for whole genome methylation sequencing of the present invention;

[0040] Figure 2 This is a schematic diagram of the overall framework of a deep analysis system for whole genome methylation sequencing according to the present invention.

[0041] Figure 3 This is a schematic diagram of an experimental strategy for whole genome methylation sequencing according to the present invention. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0043] Example 1

[0044] See also Figure 1 , an embodiment provided by the present invention: Figure 1 As shown, a deep analysis method for whole genome methylation sequencing comprises the following specific steps:

[0045] like Figure 3This is an experimental strategy for whole-genome methylation sequencing. The whole-genome methylation sequencing technology uses T4-DNA ligase to connect the adapter sequences at both ends of the ultrasonically interrupted genomic DNA fragments. The ligated products are treated with bisulfite to convert the unmethylated cytosine C into uracil U, and then the uracil U is converted into thymine T through the PCR technology mediated by the adapter sequence. The PCR amplified fragments are sequenced using NGS sequencing technology to obtain sequencing data;

[0046] Obtain the whole genome methylation sequencing data after sequencing, and use the quality score calculation formula to control the data quality;

[0047] Construct a genomic library, and use a deep learning model to compare the filtered methylation sequencing data with the reference genome in the genomic library to obtain methylation sequencing comparison data;

[0048] Construct a dynamic methylation level assessment model, and import the methylation sequencing comparison data into the methylation level assessment model to perform methylation level assessment;

[0049] A methylation map generation model was constructed, and data enhancement technology strategies were used to generate methylation maps from methylation data.

[0050] It should be specifically noted in this embodiment that obtaining whole genome methylation sequencing data and performing data quality control include the following specific steps:

[0051] S11, obtaining the original sequencing data generated by the whole genome methylation sequencing platform after sequencing;

[0052] S12, importing the obtained raw sequencing data into a quality score calculation formula for sequencing data to calculate a quality score for the sequencing data, wherein the quality score calculation formula for sequencing data is: Among them, n is the number of sequencing data, x is the sequencing data, and w i is the weight, M(x i ) is the quality index of the sequencing data. If the quality score is greater than or equal to the threshold, it means that the original sequencing data does not need to be filtered and S2 is performed; if the quality score is less than the threshold, it means that the original sequencing data needs to be filtered.

[0053] It should be specifically explained here that the Q value threshold is set to Q≥30, and the coverage threshold is set to ≥10×.

[0054] It should be specifically explained here that the comparison of the filtered methylation sequencing data with the reference genome in the genome library by the deep learning model includes the following specific contents:

[0055] S21. Construct whole genome library, target region library and transcriptome library through experiments;

[0056] S22, converting the methylation sequencing data and the reference genome into a three-dimensional tensor, and encoding the genome sequence using one-hot encoding;

[0057] S23, using Python to construct a convolutional neural network, and importing the filtered data into the convolutional neural network for comparison to identify differences in methylation status;

[0058] In this embodiment, the specific steps of S23 include:

[0059] S231, import the data processed by S22 into the deep learning library, define the model structure, including the input layer, convolution layer, activation function, pooling layer, flattening layer and fully connected layer, select cross entropy as the loss function, use Adam optimizer, and set accuracy as the evaluation indicator;

[0060] S232, dividing the data into a training set and a validation set, using the training set to train the model and monitoring the performance through the validation set;

[0061] S233. Use an independent test set to evaluate the trained model and calculate the loss and accuracy;.

[0062] S234. Predict new methylation data to identify differences in methylation status between different samples.

[0063] The following example code describes this step:

[0064] from tensorflow.keras.models import Sequential

[0065] from tensorflow.keras.layers import Conv1D,MaxPooling1D,Flatten,Dense

[0066] model=Sequential()

[0067] model.add(Conv1D(64,kernel_size=3,activation='relu',input_shape=(sequence_length,num_features)))

[0068] model.add(MaxPooling1D(pool_size=2))

[0069] model.add(Flatten())

[0070] model.add(Dense(32,activation='relu'))

[0071] model.add(Dense(1,activation='sigmoid'))

[0072] model.compile(optimizer='adam',loss='binary_crossentropy',metrics=['accuracy'])

[0073] S24, importing the compared data into the comparison accuracy and recall calculation formula for model evaluation, wherein the calculation formula for the comparison accuracy is: The calculation formula for recall is: Among them, m is the genomic fragment on the read alignment, N is the total number of genomic fragments, TP is the number of samples correctly predicted as positive, FP is the number of samples incorrectly predicted as positive, and F is the number of samples incorrectly predicted as negative.

[0074] It should be specifically explained here that the construction of the dynamic methylation level assessment model and the importation of the methylation sequencing comparison data into the methylation level assessment model for methylation level assessment include the following specific steps:

[0075] S31, importing the filtered data into a dynamic weighted methylation level calculation formula to calculate the weighted methylation level, wherein the calculation formula for the dynamic weighted methylation level at time t is:

[0076] Where Wt is the dynamic weighted methylation level at time t, Mi,t is the methylation state of the i-th site at time t (0 or 1), and w i , t is the dynamic weight of the ith site at time t, n is the total number of sites, where the dynamic weight is weighted based on time and biological importance, and its calculation formula is: w i，t =w i，0 ×e -λt ×Z, where w i,0 is the initial weight, λ is the decay rate, and Z is the biological importance score of the gene or site; the methylation state can be expressed in binary as:

[0077]

[0078] S32. Construct a feature matrix based on the methylation status, GC content, sequencing depth, and one-hot coding of base sequence features in the database, where the calculation formula for GC content is: Where Gj is the number of guanine bases in the jth sequence, C j is the number of cytosine bases in the jth sequence, A j is the number of adenine bases in the jth sequence, T j is the number of thymine bases in the jth sequence. The calculation formula for sequencing depth is: D = T / L, where T is the total number of bases obtained in sequencing, and L is the base length of the target region. The calculation formula for One-Hot encoding is: Similarly, we can get S C , S G , S T . All features are integrated into a feature matrix X, The target variable is

[0079] S33. Perform model training, cross-validation and dynamic evaluation using a support vector machine model based on feature data.

[0080] In this embodiment, the specific steps of S23 include:

[0081] S331, split the data processed by S32: divide the data set into a training set and a test set, 70%-80% for training and 20%-30% for testing.

[0082] S332. Use the SVM module in Scikit-learn to create a model instance and use the training set to train the model.

[0083] S333. Use the cross-validation tool to evaluate the stability of the model and calculate the average accuracy.

[0084] S334. Make predictions on the test set and calculate evaluation indicators such as accuracy, F1 score, and recall.

[0085] The following example code describes this step:

[0086] import numpy as np

[0087] import pandas as pd

[0088] from sklearn.model_selection import train_test_split

[0089] from sklearn.preprocessing import StandardScaler

[0090] from sklearn.svm import SVR

[0091] from sklearn.metrics import mean_squared_error,r2_score

[0092] import matplotlib.pyplot as plt

[0093] Methylation_Status

[0094] data=pd.read_csv('methylation_data.csv')

[0095] X=data[['M_1','M_2','M_3']]

[0096] y=data['Methylation_Status']

[0097] X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.2,random_state=42)

[0098] scaler=StandardScaler()

[0099] X_train=scaler.fit_transform(X_train)

[0100] X_test=scaler.transform(X_test)

[0101] model=SVR(kernel='rbf')

[0102] model.fit(X_train,y_train)

[0103] y_pred=model.predict(X_test)

[0104] mse=mean_squared_error(y_test,y_pred)

[0105] r2=r2_score(y_test,y_pred)

[0106] print(f'Mean Squared Error:{mse:.4f}')

[0107] print(f'R-squared:{r2:.4f}')

[0108] plt.scatter(y_test,y_pred)

[0109] plt.plot([y_test.min(),y_test.max()],[y_test.min(),y_test.max()],'r--')

[0110] plt.xlabel('Actual Methylation Status')

[0111] plt.ylabel('Predicted Methylation Status')

[0112] plt.title('Actual vs Predicted Methylation Status')

[0113] plt.show()

[0114] It should be specifically explained here that the construction of the methylation map generation model and the use of data enhancement technology strategy to generate a methylation map from methylation data include the following specific steps:

[0115] S41. Generate a heat map of the dynamically evaluated methylation data using R language to obtain the methylation levels of different samples and under different conditions;

[0116] The following example code describes this step:

[0117] rownames(methylation_data)<-row_names

[0118] pheatmap(methylation_data,

[0119] cluster_rows=TRUE,

[0120] cluster_cols=TRUE,

[0121] display_numbers=TRUE,

[0122] main="Methylation Levels Heatmap",

[0123] fontsize_number=8,

[0124] annotation_col=data.frame(Condition=conditions))

[0125] S42, set the window size H and step size S of the sliding window, apply the sliding window to each region in the genome, starting from the first base pair, the window range is [kS, kS+H], where k is the number of times the window moves, and calculate the weighted methylation level P of the CpG site in each window, and the calculation formula is: Where W i is the weighted methylation status of the i-th CpG site, and n is the number of CpG sites contained in the window;

[0126] S43, moving the window forward according to the set step size S until the entire genome is traversed, and finally generating a smoothed whole-genome methylation map.

[0127] Example 2

[0128] like Figure 2 As shown, a deep analysis system for whole genome methylation sequencing is implemented based on the above-mentioned deep analysis method for whole genome methylation sequencing, which specifically includes a quality control module, a sequence alignment module, a dynamic methylation level assessment module, a methylation map generation module and a network analysis strategy verification module. The quality control module is used to perform quality processing on the original sequencing data and filter the data. The sequence alignment module is used to align the measured sequence with the reference sequence and generate an alignment file. The dynamic methylation level assessment module is used to construct a dynamic methylation level assessment model through machine learning to dynamically assess the methylation degree. The methylation map generation module is used to import the dynamically assessed methylation data into the methylation map generation model to obtain a whole genome methylation map.

[0129] Example 3

[0130] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0131] The processor executes the above-mentioned deep analysis method for whole genome methylation sequencing by calling the computer program stored in the memory.

[0132] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors and one or more memories, wherein at least one computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement a method for in-depth analysis of whole genome methylation sequencing provided by the above method embodiment. The electronic device may also include other components for implementing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input and output interface, so as to input and output data. This embodiment will not be described in detail here.

[0133] Example 4

[0134] This embodiment provides a computer-readable storage medium having a rewritable computer program stored thereon;

[0135] When the computer program runs on a computer device, the computer device executes the above-mentioned deep analysis method for whole genome methylation sequencing.

[0136] For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0137] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0138] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0139] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through a wired network or / and a wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0140] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. 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 can be combined in any one or more embodiments or examples in a suitable manner.

[0141] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A deep analysis system and method for whole genome methylation sequencing, characterized in that: It includes the following specific steps: Obtain the whole genome methylation sequencing data after sequencing, and use the quality score calculation formula to control the data quality; Construct a genomic library, and use a deep learning model to compare the filtered methylation sequencing data with the reference genome in the genomic library to obtain methylation sequencing comparison data; Construct a dynamic methylation level assessment model, and import the methylation sequencing comparison data into the methylation level assessment model to perform methylation level assessment; A methylation map generation model was constructed, and data enhancement technology strategies were used to generate methylation maps from methylation data.

2. A deep analysis system and method for whole genome methylation sequencing according to claim 1, characterized in that: The step of obtaining whole genome methylation sequencing data and performing data quality control comprises the following specific steps: S11, obtaining the original sequencing data generated by the whole genome methylation sequencing platform after sequencing; S12, importing the obtained raw sequencing data into a quality score calculation formula for sequencing data to calculate a quality score for the sequencing data, wherein the quality score calculation formula for sequencing data is: Among them, n is the number of sequencing data, x is the sequencing data, and w i is the weight, M(x i ) is the quality index of the sequencing data. If the quality score is greater than or equal to the threshold, it means that the original sequencing data does not need to be filtered and S2 is performed; if the quality score is less than the threshold, it means that the original sequencing data needs to be filtered.

3. A deep analysis system and method for whole genome methylation sequencing according to claim 2, characterized in that: The method of comparing the filtered methylation sequencing data with the reference genome in the genome library by using a deep learning model includes the following specific contents: S21. Construct whole genome library, target region library and transcriptome library through experiments; S22, converting the methylation sequencing data and the reference genome into a three-dimensional tensor, and encoding the genome sequence using one-hot encoding; S23, constructing a convolutional neural network, and importing the filtered data into the convolutional neural network for comparison to identify differences in methylation status; S24, importing the compared data into the comparison accuracy and recall calculation formula for model evaluation, wherein the calculation formula for the comparison accuracy is: The calculation formula for recall is: Among them, m is the genomic fragment on the read alignment, N is the total number of genomic fragments, TP is the number of samples correctly predicted as positive, FP is the number of samples incorrectly predicted as positive, and FN is the number of samples incorrectly predicted as negative.

4. A deep analysis system and method for whole genome methylation sequencing according to claim 3, characterized in that: The construction of the dynamic methylation level evaluation model includes the following specific contents: S31, importing the filtered data into a dynamic weighted methylation level calculation formula to calculate the weighted methylation level, wherein the calculation formula for the dynamic weighted methylation level at time t is: Among them, W t is the dynamic weighted methylation level at time t, M i,t is the methylation status of the i-th site at time t, w i,t is the dynamic weight of the ith site at time t, n is the total number of sites, where the dynamic weight is weighted based on time and biological importance, and its calculation formula is: i，t =w i，0 ×e -λt ×Z, where w i,0 is the initial weight, λ is the decay rate, and Z is the biological importance score of the gene or site; S32. Construct a feature matrix based on the methylation status, GC content, sequencing depth, and one-hot coding of base sequence features in the database, where the calculation formula for GC content is: Among them, G j is the number of guanine bases in the jth sequence, C j is the number of cytosine bases in the jth sequence, A j is the number of adenine bases in the jth sequence, T j is the number of thymine bases in the jth sequence. The calculation formula for sequencing depth is: D = T / L, where T is the total number of bases obtained in sequencing, and L is the base length of the target region. The calculation formula for One-Hot encoding is: Similarly, we can get S C , S G , S T . All features are integrated into a feature matrix X, The target variable is S33. Perform model training, cross-validation and dynamic evaluation using a support vector machine model based on feature data.

5. A deep analysis system and method for whole genome methylation sequencing according to claim 4, characterized in that: The generating of the methylation profile includes the following specific contents: S41. Generate a heat map of the dynamically evaluated methylation data using R language to obtain the methylation levels of different samples and under different conditions; S42, set the window size H and step size S of the sliding window, apply the sliding window to each region in the genome, starting from the first base pair, the window range is [kS, kS+H], where k is the number of times the window moves, and calculate the weighted methylation level P of the CpG site in each window, and the calculation formula is: Where W i is the weighted methylation status of the i-th CpG site, and n is the number of CpG sites contained in the window; S43, moving the window forward according to the set step size S until the entire genome is traversed, and finally generating a smoothed whole-genome methylation map.

6. A deep analysis system for whole genome methylation sequencing, which is implemented based on the deep analysis method for whole genome methylation sequencing according to any one of claims 1 to 6, characterized in that: Specifically include: The quality control module is used to process the quality of raw sequencing data and filter the data; A sequence comparison module is used to compare the measured sequence with the reference sequence and generate a comparison file; Dynamic methylation level assessment module, which is used to construct a dynamic methylation level assessment model through machine learning to dynamically assess the degree of methylation; The methylation map generation module is used to import the dynamically evaluated methylation data into the methylation map generation model to obtain the whole genome methylation map.

7. An electronic device comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; It is characterized in that the processor executes a deep analysis method for whole genome methylation sequencing as described in any one of claims 1 to 5 by calling a computer program stored in the memory.

8. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes a deep analysis method for whole genome methylation sequencing as described in any one of claims 1 to 5.

Images