Preeclampsia noninvasive screening method based on deep sequencing of 4-mer terminal motif spectrum characteristics

Abstract

The invention relates to the technical field of noninvasive prenatal diagnosis, and particularly discloses a preeclampsia noninvasive screening method based on deep sequencing 4-mer terminal motif spectrum characteristics, which comprises the following steps: collecting preeclampsia and healthy pregnant woman peripheral blood samples, and extracting free DNA for high-throughput sequencing; 12 key terminal 4-mer motifs are extracted and screened out through bioinformatics; and a multi-layer perceptron (MLP) model is adopted to carry out machine learning analysis, and the preeclampsia risk is predicted. According to the method, 12 terminal 4-mer motif spectrums are specifically screened, and the MLP model is combined, so that the limitation that a traditional screening method is low in specificity and invasive in detection is effectively broken through.

Description

Technical Field

[0001] This invention relates to the field of non-invasive prenatal diagnostic technology, specifically to a non-invasive screening method for preeclampsia based on the characteristics of the 4-mer terminal motif spectrum of deep sequencing. Background Technology

[0002] Preeclampsia (PE), a serious complication specific to pregnancy, typically develops after 20 weeks of gestation. Its core characteristic is new-onset hypertension, often accompanied by proteinuria, maternal liver and kidney damage, or placental dysfunction. Global epidemiological data shows that PE incidence rates are approximately 5%-9% in pregnant women, making it a significant cause of maternal mortality, contributing 10%-15% to the overall rate. Furthermore, PE can lead to adverse pregnancy outcomes such as fetal growth restriction and premature birth, causing over 70,000 maternal deaths and 500,000 fetal deaths annually, posing a major threat to maternal and infant health.

[0003] From a pathogenesis perspective, the pathological process of placental pelvic inflammatory disease (PE) is closely related to placental dysfunction. In early pregnancy (10-12 weeks), placental tissue in PE patients already shows significant gene expression disorders. As pregnancy progresses, problems such as impaired placental spiral artery remodeling and insufficient trophoblast cell infiltration lead to inadequate placental perfusion, resulting in placental ischemia and hypoxia. Ultimately, this induces systemic maternal vasospasm through the release of pro-inflammatory factors and vasoactive substances, manifesting as clinical symptoms of PE. Currently, among the early assessment methods for placental function, chorionic villus sampling can obtain placental tissue for phenotypic analysis, but due to its invasiveness and the risk of miscarriage, it is difficult to widely apply in clinical screening. In addition, with the increasing trend of delayed marriage and childbearing and the rising incidence of metabolic abnormalities during pregnancy, the incidence of PE is showing a year-on-year increasing trend, urgently requiring the development of safe and efficient early screening technologies.

[0004] In current clinical practice, the diagnosis of PE still relies mainly on routine indicators such as blood pressure monitoring and urinary protein quantification. However, these indicators often show significant abnormalities only in the middle and late stages of the disease, resulting in insufficient early predictive sensitivity. In recent years, non-invasive detection technologies based on cell-free peripheral blood DNA (cfDNA) have become a hot research area for PE diagnosis due to their high safety and convenient sampling. Existing studies have explored dimensions such as cfDNA methylation patterns, fragment length distribution, and multi-omics data integration, attempting to discover molecular biomarkers related to PE. However, existing studies mostly focus on macroscopic fragment features or genomic methylation levels, and systematic analysis of the cfDNA terminal base sequence composition (especially short fragment motifs) has not been in-depth. This has led to a lack of efficient and specific non-invasive PE biomarkers based on cfDNA terminal sequence characteristics, making it difficult to meet the clinical need for accurate early screening of PE. Therefore, developing PE screening methods based on cfDNA terminal motif profiles has significant clinical value in improving the efficiency of early PE diagnosis and improving maternal and infant prognosis. Summary of the Invention

[0005] This invention addresses the shortcomings of existing preeclampsia (PE) screening technologies, such as insufficient specificity of non-invasive biomarkers and limited early diagnostic capabilities. It provides a non-invasive preeclampsia screening method based on the motif characteristics of 4-mer terminal sequences from deep sequencing. This method acquires the sequence data of cell-free DNA from the peripheral blood of pregnant women using high-throughput sequencing technology. It focuses on analyzing the motif pattern composed of the last four bases of cfDNA fragments, and combines bioinformatics analysis and machine learning techniques to construct a model, achieving early, non-invasive, and efficient screening for preeclampsia.

[0006] To achieve the above objectives, this invention proposes a non-invasive screening method for preeclampsia based on the characteristics of 4-mer terminal motifs from deep sequencing, comprising: S1. Sample collection, extraction and sequencing: Peripheral blood samples were collected from pregnant women with preeclampsia (PE) and healthy pregnant women. cfDNA was extracted and high-throughput whole-genome sequencing was performed at a sequencing depth of 30X to obtain raw sequencing data. S2. Data alignment and sorting: The original sequencing reads are aligned with the human reference genome hg38 to generate an alignment result file, which is then sorted and deduplicated to obtain hundreds of billions of cfDNAs. S3. Split the dataset: Divide the preprocessed data into a training set and a test set in a 7:3 ratio. The training set is used for model training and parameter tuning, while the test set is used for independent validation and performance evaluation of the finally trained model. S4. Feature extraction from sequencing data: 4-mer terminal motifs were extracted from the training and test sets respectively, and then the frequency of each 4-mer motif was calculated—representing 4... 4 (256) possible combinations—a proportion relative to the total number of motifs; S5. Feature selection: LASSO regression was used to select features in the training set, and 12 key 4-mer motifs were selected. S6. Build and tune the model: Build an MLP model and tune the parameters using the numerical values ​​of the training set. Select the optimal parameters to obtain a trained prediction model. S7. Model Validation and Performance Evaluation; S8. Calculate and output the predicted probability of preeclampsia.

[0007] Preferably, in S6, the specific process of building the model and tuning parameters is as follows: S61. Label patients with preeclampsia as 1 and healthy pregnant women as 0. S62. Based on the core features of the frequency of 12 4-mer motifs selected, and based on data from preeclampsia samples and healthy pregnant women, an MLP model was constructed. S63. Perform hyperparameter tuning on the above MLP model, select the optimal parameters, and obtain the optimal model. Preferably, in S7, the specific steps for model validation and performance evaluation are as follows: using a strictly isolated test set to evaluate the generalization performance and diagnostic capability of the trained prediction model obtained in step S63, wherein the indicators of generalization performance and diagnostic capability include computational accuracy, sensitivity, specificity, and AUC.

[0008] Preferably, in S8, the specific steps for calculating and outputting the predicted probability of preeclampsia from the screened new pregnant women samples are as follows: S81. Collect peripheral blood samples according to steps S1-S2, extract cfDNA and perform high-throughput sequencing at a depth of 30X, and align to hg38. S82. Following steps S3-S5, quantify the frequency of all 256 possible 4-mer terminal motifs in the cfDNA fragment and screen out 12 key 4-mer motifs. S83. Input the feature vector obtained in S82 into the prediction model trained in S63. S84. Call the model's predict_proba method to obtain the predicted probability P that the sample belongs to the preeclampsia category label 1; S85. Output the predicted probability P as the predicted risk value of the pregnant woman for preeclampsia.

[0009] Therefore, this invention proposes a non-invasive screening method for preeclampsia based on the characteristics of 4-mer terminal motifs from deep sequencing, with the following beneficial effects: (1) This invention establishes the 4-mer terminal motif frequency as a sensitive and specific biomarker for detecting cfDNA cleavage preference. Only 12 terminal motifs are needed to efficiently distinguish preeclampsia, which significantly simplifies the feature complexity and enables early diagnosis of preeclampsia through non-invasive detection, which is helpful for early intervention and treatment.

[0010] (2) The diagnosis was performed using a machine learning model. The AUC value of the training set and the test set was 0.975, which is highly accurate. It only requires the collection of peripheral blood from pregnant women, does not require invasive operation, is highly safe, is easy to apply in clinical practice, and has greater practicality.

[0011] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0012] Figure 1 This is a flowchart of a non-invasive screening method for preeclampsia based on the characteristics of 4-mer terminal motifs in deep sequencing, according to the present invention. Figure 2This invention provides a frequency correlation diagram of all 256 possible 4-mer terminal motifs based on the spectral characteristics of 4-mer terminal motifs in deep sequencing. Figure 3 This invention provides a LASSO regression screening map based on the characteristics of 4-mer terminal motifs from deep sequencing. Figure 4 This invention relates to a violin diagram based on the characteristic frequencies of the 4-mer terminal motif spectrum of deep sequencing. Figure 5 This is the ROC curve of the MLP model of a non-invasive preeclampsia screening method based on deep sequencing 4-mer terminal motif characteristics in the training set and the test set, where A is the training set and B is the test set. Figure 6 This is a flowchart illustrating an implementation example of a non-invasive preeclampsia screening method based on the 4-mer terminal motif characteristics of deep sequencing according to the present invention. Detailed Implementation

[0013] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0014] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention. Experimental methods in the following embodiments that do not specify specific conditions are generally determined according to national standards. Experimental instruments, equipment, and reagents in the following embodiments that do not specify their sources are all commercially available materials.

[0015] Unless otherwise defined or stated, all technical and scientific terms used in this invention have the same meaning as those skilled in the art. Furthermore, any methods and materials similar to or equivalent to those described herein can be applied to the methods of this invention. It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other.

[0016] like Figure 1 As shown, the non-invasive screening method for preeclampsia based on the 4-mer terminal motif characteristics of deep sequencing provided by the present invention specifically includes the following steps: S1. Sample collection, extraction and sequencing: Peripheral blood samples were collected from pregnant women with preeclampsia (PE) and healthy pregnant women. cfDNA was extracted and high-throughput whole-genome sequencing was performed at a sequencing depth of 30X to obtain raw sequencing data. S2. Data alignment and sorting: The original sequencing reads are aligned with the human reference genome hg38 to generate an alignment result file, which is then sorted and deduplicated to obtain hundreds of billions of cfDNAs. S3. Split the dataset: Divide the preprocessed data into a training set and a test set in a 7:3 ratio. The training set is used for model training and parameter tuning, while the test set is used for independent validation and performance evaluation of the finally trained model. S4. Feature extraction from sequencing data: 4-mer terminal motifs were extracted from the training and test sets respectively, and then the frequency of each 4-mer motif was calculated—representing 4... 4 (256) possible combinations—a proportion relative to the total number of motifs; S5. Feature Selection: LASSO regression was used to select features from the training set, identifying 12 key 4-mer motifs (see...). Figure 3 , 4 ); S6. Build and tune the model: Build the MLP model and tune the parameters using the numerical values ​​from the training set; The specific process of building the model and tuning its parameters is as follows: S61. Label patients with preeclampsia as 1 and healthy pregnant women as 0. S62. Based on the core features of the frequency of 12 4-mer motifs selected, and based on data from preeclampsia samples and healthy pregnant women, an MLP model was constructed. S63. Perform hyperparameter tuning on the above MLP model and select the optimal parameters to obtain the optimal model. S7. Model Validation and Performance Evaluation; The specific steps for model validation and performance evaluation are as follows: Use a strictly isolated test set to evaluate the generalization performance and diagnostic capability of the trained prediction model obtained in step S63. The metrics for generalization performance and diagnostic capability include computational accuracy, sensitivity, specificity, and AUC.

[0017] S8. Calculate and output the predicted probability of preeclampsia.

[0018] The specific steps for calculating and outputting the predicted probability of preeclampsia from newly screened pregnant women are as follows: S81. Collect peripheral blood samples according to steps S1-S2, extract cfDNA and perform high-throughput sequencing at a depth of 30X, and align to hg38. S82. Following steps S3-S5, quantify the frequency of all 256 possible 4-mer terminal motifs in the cfDNA fragment, and screen out 12 key 4-mer motifs (see...). Figure 2 ); S83. Input the feature vector obtained in S82 into the prediction model trained in S63. S84. Call the model's predict_proba method to obtain the predicted probability P that the sample belongs to the preeclampsia category label 1; S85. Output the predicted probability P as the predicted risk value of the pregnant woman for preeclampsia.

[0019] In S6, the MLP model is specifically as follows: The model uses a multilayer perceptron: Built with Keras and encapsulated using KerasClassifier for integration with GridSearchCV, enabling hyperparameter tuning. Grid search employs 5-fold cross-validation, with accuracy as the scoring metric. The final parameters are determined as follows: number of hidden layer neurons (units=32), dropout rate (dropout_rate=0.01), L2 regularization strength (l2_reg=0.01), optimizer (Adam), number of training epochs (16), and batch size (batch_size=8).

[0020] The multilayer perceptron architecture consists of three hidden layers with ReLU activation functions. Each layer applies L2 regularization and Dropout to prevent overfitting. The output layer is a single neuron, using a sigmoid activation function to handle binary classification. The model uses binary cross-entropy as the loss function for optimization.

[0021] In S3, dividing the samples into a 7:3 ratio of training and test sets is a key step in achieving model reliability and generalization ability. 1) Training set (used in model building): Used for parameter learning, hyperparameter tuning, and 5-fold cross-validation of the MLP model. Through learning 12 key 4-mer motif frequency features, the model grasps the mapping relationship between "features → preeclampsia risk" (see...). Figure 5 (A)

[0022] 2) Test set (completely independent of the training process): Used to objectively evaluate the predictive performance of the model on unseen data after training. On 57 test samples (25 cases of preeclampsia and 32 normal pregnancies), the MLP model AUC = 0.975 (see...). Figure 5 The model performed excellently in metrics such as accuracy (0.89) and precision (0.83), proving that it does not merely "memorize" the training data but has the ability to generalize to real clinical samples.

[0023] Therefore, sample partitioning is not only a common best practice in the field of machine learning, but also ensures that the predictive model of this invention is both effective and reliable in real clinical applications: the training set is responsible for "learning" the classification rules, and the test set is responsible for "testing" whether the rules can be generalized and applied.

[0024] like Figure 6As shown in the embodiment of the present invention, a non-invasive screening method for preeclampsia based on deep sequencing 4-mer terminal motif characteristics is provided. This method uses an MLP model based on the frequencies of 12 4-mer terminal motifs in the subject of screening to predict the probability of preeclampsia. The method consists of six steps: Step 1: Obtain peripheral blood samples from pregnant women, extract cfDNA and perform quantification, library construction and sequencing; Step 2: Preprocessing the sequencing data by comparing, sorting, and deduplicating it with the human reference genome hg38; Step 3: Extract the 4-mer terminal motif spectra from the training and test sets respectively, and then calculate the frequency of each 4-mer motif; Step 4: Select 12 key 4-mer terminal motifs; Step 5: Input the frequency values ​​of the 12 motifs as features into the MLP model; Step 6: Calculate the probability of preeclampsia in the individuals to be screened using the model.

[0025] As shown in Table 1, the diagnostic performance of the MLP model was validated on the training and test sets. The results showed that the area under the curve (AUC) of the model on the test set was 0.975, which can accurately identify patients and healthy people.

[0026] Table 1 Evaluation metrics of MLP models on the training / test sets

[0027] Therefore, this invention provides a non-invasive screening method for preeclampsia based on the 4-mer terminal motif characteristics of deep sequencing. By analyzing specific 4-mer terminal motif characteristics in cell-free DNA from the peripheral blood of pregnant women, it achieves high specificity and accuracy for early diagnosis. This method only requires peripheral blood collection, involves no invasive procedures, and is highly safe. Furthermore, the diagnostic accuracy is significantly improved through machine learning models, with an area under the curve exceeding 0.975 on the test set. In addition, this method simplifies feature complexity, requiring only 12 key motifs for efficient differentiation, providing strong support for early intervention and treatment. Simultaneously, its strong practicality lays the foundation for the fabrication of targeted detection chips and has broad clinical application prospects.

[0028] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A non-invasive screening method for preeclampsia based on deep sequencing 4-mer terminal motif profile characteristics, characterized in that, include: S1. Sample collection, extraction and sequencing: Peripheral blood samples were collected from pregnant women with preeclampsia (PE) and healthy pregnant women. cfDNA was extracted and high-throughput whole-genome sequencing was performed at a sequencing depth of 30X to obtain raw sequencing data. S2. Data alignment and sorting: The original sequencing reads are aligned with the human reference genome hg38 to generate an alignment result file, which is then sorted and deduplicated to obtain billions of cfDNAs. S3. Split the dataset: Divide the preprocessed data into a training set and a test set in a 7:3 ratio. The training set is used for model training and parameter tuning, while the test set is used for independent validation and performance evaluation of the finally trained model. S4. Feature extraction from sequencing data: 4-mer terminal motifs were extracted from the training and test sets respectively, and then the frequency of each 4-mer motif was calculated—representing 4... 4 (256) possible combinations—a proportion relative to the total number of motifs; S5. Feature selection: LASSO regression was used to select features in the training set, and 12 key 4-mer motifs were selected. S6. Build and tune the model: Build an MLP model and tune the parameters using the numerical values ​​of the training set. Select the optimal parameters to obtain a trained prediction model. S7. Model Validation and Performance Evaluation; S8. Calculate and output the predicted probability of preeclampsia.

2. The non-invasive screening method for preeclampsia based on deep sequencing 4-mer terminal motif characteristics according to claim 1, characterized in that, In S6, the specific process of building the model and tuning parameters is as follows: S61. Label patients with preeclampsia as 1 and healthy pregnant women as 0. S62. Based on the core features of the frequency of 12 4-mer motifs selected, and based on data from preeclampsia samples and healthy pregnant women, an MLP model was constructed. S63. Perform hyperparameter tuning on the above MLP model, select the optimal parameters, and obtain the trained prediction model.

3. The non-invasive screening method for preeclampsia based on deep sequencing 4-mer terminal motif characteristics according to claim 1, characterized in that, In S7, the specific steps for model validation and performance evaluation are as follows: Use a strictly isolated test set to evaluate the generalization performance and diagnostic capability of the trained prediction model obtained in S6. The metrics for generalization performance and diagnostic capability include computational accuracy, sensitivity, specificity, and AUC.

4. A non-invasive screening method for preeclampsia based on deep sequencing 4-mer terminal motif profiles as described in any one of claims 1-3, characterized in that, In S8, the specific steps for calculating and outputting the predicted probability of preeclampsia from newly screened pregnant women are as follows: S81. Collect peripheral blood samples according to steps S1-S2, extract cfDNA and perform high-throughput sequencing at a depth of 30X, and align to hg38. S82. Following steps S3-S5, quantify the frequency of all 256 possible 4-mer terminal motifs in the cfDNA fragment and screen out 12 key 4-mer motifs. S83. Input the feature vector obtained in S82 into the prediction model trained in S63. S84. Call the model's predict_proba method to obtain the predicted probability P that the sample belongs to the preeclampsia category label 1; S85. Output the predicted probability P as the predicted risk value of the pregnant woman for preeclampsia.

Images