Method for predicting Tm value of dissolution curve

Abstract

The invention provides a solubility curve Tm value prediction method, and relates to the technical field of biological detection, and the method comprises the following steps: constructing a Tm value marking model and a Tm value prediction model; acquiring data information of a to-be-tested sample, and inputting the data information into the Tm value marking model; preprocessing the data information by a Tm value marking model and marking whether the to-be-detected samples have Tm values or not, and forming an effective sample set by the to-be-detected samples with the Tm values; extracting a plurality of peak values from the original dissolution curve to obtain an original Tm value set; encoding the original Tm value set according to the interval temperature; wherein the temperature bits in the original binary vector are connected and arranged according to interval temperatures; and inputting the original dissolution curve in the effective sample set into the Tm value prediction model to obtain a predicted Tm value. The temperature bits are connected and arranged according to the interval temperatures, so that the model learns the correlation between the adjacent temperatures, the influence of the temperature change on the Tm value is more accurately captured, and the prediction accuracy of the Tm value is improved.

Description

Technical Field

[0001] The present invention relates to the field of biological detection technology, and in particular to a method for predicting the Tm value of a melting curve. Background Art

[0002] Polymerase chain reaction (PCR) technology is widely used in molecular biology, medical diagnosis, and genetic research. One of its core steps is melting curve analysis, which evaluates the specificity and purity of amplified products by extracting the melting temperature (Tm value). However, existing methods do not encode the Tm value and do not have While retaining the sequential nature of Tm values, temperature discretization does not preserve the correlation between adjacent intervals, which results in the model being unable to learn the correlation between adjacent temperatures. This makes the model insensitive to small differences and makes it difficult to capture the impact of small temperature changes on Tm values. Traditional methods do not accurately determine Tm values ​​and cannot mark anomalies. Currently, there is an urgent need for a method that can make Tm values ​​more accurate and save more labor time. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for predicting the Tm value of a melting curve. The technical problem to be solved is how to improve the accuracy of Tm value calculation based on the melting curve.

[0004] The present invention is achieved through the following technical solutions:

[0005] A method for predicting the Tm value of a melting curve comprises the following steps:

[0006] Based on the machine learning model, a Tm value labeling model and a Tm value prediction model were constructed;

[0007] Obtaining data information of each test sample in the test sample set, and inputting the data information into a Tm value labeling model; wherein the data information includes an original melting curve;

[0008] The data information is preprocessed by the Tm value labeling model; the Tm value labeling model is used to label the test samples with or without Tm values, and the test samples with Tm values ​​constitute a valid sample set;

[0009] Extracting several peaks from the original melting curve of the test sample in the above valid sample set to obtain an original Tm value set;

[0010] The original Tm value set is encoded according to a preset interval temperature to obtain an original binary vector; wherein the temperature bits in the original binary vector are connected and arranged according to the interval temperature.

[0011] The original melting curves of the samples to be tested in the above-mentioned valid sample set are input into the Tm value prediction model to obtain a prediction binary vector; wherein the temperature bits in the above-mentioned prediction binary vector are connected and arranged according to interval temperatures.

[0012] The original binary vector and the predicted binary vector are compared to determine whether the predicted binary vector is abnormal.

[0013] The original melting curves of the samples to be tested in the valid sample set are divided according to the preset interval temperature to generate a binary vector arranged according to the temperature bits, where the adjacent temperature bits are separated by the preset interval temperature. This fully considers the sequential nature of the temperature and enables the model to learn the correlation between adjacent temperatures, thereby more accurately capturing the impact of temperature changes on Tm values, which helps to improve prediction accuracy.

[0014] The input data information of the test samples (including the original melting curve) is preprocessed through the Tm value marking model to mark whether the test samples have Tm values. The test samples with Tm values ​​are screened out to form a valid sample set, eliminating the interference of samples without Tm values, so that the subsequent prediction model can focus on the analysis of valid samples, reducing the interference of irrelevant data on the model prediction, and thus improving the accuracy of the prediction; by comparing the original binary vector and the predicted binary vector, it is checked whether the Tm value predicted by the Tm value prediction model is normal. If the original binary vector and the predicted binary vector are consistent, the predicted Tm value is normal; otherwise, the predicted Tm value is abnormal, and the original melting curve is marked as an abnormal curve.

[0015] Furthermore, based on the machine learning model, the specific steps of constructing the Tm value labeling model include:

[0016] Obtaining historical data information of each first historical sample in the first historical sample set; wherein the historical data information includes an original melting curve and an original mark with or without a Tm value mark;

[0017] Preprocessing the historical data information of each first historical sample in the first historical sample set;

[0018] Dividing the preprocessed first historical sample set into a first training set and a first test set;

[0019] Inputting the historical data information in the first training set into the machine learning model for training to obtain a Tm value labeling model;

[0020] Inputting the original melting curve of a first historical sample in the first test set into the Tm value marker model to obtain a test marker indicating whether the first historical sample has a Tm value marker;

[0021] The original mark and the test mark of the first historical sample are matched. If the match is successful, the mark of the first historical sample is normal; if the match fails, the mark of the first historical sample is abnormal.

[0022] The machine learning model learns the mapping relationship from melting curve features to the presence or absence of Tm value markers on the training set and verifies it on the test set to test the learning effect and generalization ability of the Tm value marker model. The original melting curve of the first historical sample in the first test set is input into the Tm value marker model to test the Tm value marker model. The original marker of the first historical sample and the test marker are then matched to verify the accuracy of the Tm value marker model. Problems and deficiencies in the Tm value marker model are promptly identified to provide direction for subsequent model optimization, ensuring that the final constructed Tm value marker model can accurately distinguish samples with and without Tm values, screening valid sample data for subsequent steps, reducing the interference of invalid data on the prediction model, and thus improving the accuracy of Tm value prediction.

[0023] Furthermore, based on the machine learning model, the specific steps of constructing a Tm value prediction model include:

[0024] Obtaining historical data information of each second historical sample in the second historical sample set; wherein the historical data information includes an original melting curve and an original Tm value;

[0025] Preprocessing the historical data information of each second historical sample in the second historical sample set;

[0026] Dividing the preprocessed second historical sample set into a second training set and a second test set;

[0027] Input the historical data information in the second training set into the machine learning model for training to obtain a Tm value prediction model;

[0028] Inputting the original melting curve of a second historical sample in the second test set into the Tm value prediction model to obtain a predicted Tm value of the second historical sample;

[0029] The original Tm value and the predicted Tm value of the second historical sample are matched. If the match is successful, the prediction of the second historical sample is successful; if the match fails, the prediction of the second historical sample is unsuccessful.

[0030] The original melting curve of the second historical sample is input into the Tm prediction model to obtain a predicted Tm value. The original and predicted Tm values ​​are then matched to measure the Tm prediction model's performance on the test set and determine whether the model can accurately predict the Tm value corresponding to the new melting curve. If the match fails, it indicates that the model failed to predict the Tm value for that sample, which helps identify problems and deficiencies in the Tm prediction model, provides direction for subsequent model optimization, and encourages researchers to adjust and improve the model to improve its prediction accuracy.

[0031] Furthermore, the second historical sample set is composed of the first historical samples in the first historical sample set that are originally marked with Tm values.

[0032] The Tm value labeling model and the Tm value prediction model are coordinated. In the actual use stage, the Tm value labeling model is used to identify the samples to be tested with Tm values ​​to obtain valid samples; and then the Tm value prediction model is used to predict the Tm value.

[0033] Furthermore, preprocessing the historical data information of each first historical sample in the first historical sample set includes removing first historical samples without Ct values.

[0034] Samples without Tm values ​​indicate that they are not infected with the virus; samples without Ct values ​​indicate that the experiment has failed and the sample has no reference value. The model marked at the Tm value is cleared.

[0035] Furthermore, the data information is preprocessed using the Tm value labeling model, and the specific steps include:

[0036] The above data information also includes data source information and experimental process information of the samples to be tested;

[0037] Clean the above data information to remove the data source information and experimental process information of the samples to be tested;

[0038] The original melting curve in the above data information was normalized.

[0039] While information about the data source and experimental process of the test sample is helpful in understanding the sample background and experimental conditions, it does not directly contribute to the determination of the Tm value and may even become a confounding factor. Removing this information during the data cleaning phase allows the model to focus more on the raw melt curve data, which is closely related to the Tm value. The data in the raw melt curve may have different dimensions and numerical ranges. This makes it difficult to balance the importance of each feature during model training, resulting in slower or even impossible convergence. Normalizing the raw melt curve and mapping all data to a specific interval (such as [0, 1] or [-1, 1]) eliminates the influence of dimension and makes the features numerically comparable. This allows the model to adjust parameters more efficiently during training, accelerating convergence and improving training efficiency.

[0040] Furthermore, when the traceability signal is received, the data source information and experimental process information of the sample to be tested are obtained from the above data information.

[0041] Patient information, experimental parameters (well number, reagent batch, PCR instrument channel) and operation records (detection time, operator ID, original curve file) are associated with each sample. Through database indexing, reverse tracking of "diagnostic result → Tm value → experimental conditions" is achieved, supporting rapid problem location (such as temperature drift caused by a certain batch of reagents).

[0042] Furthermore, the preset interval temperature is 0.05°C to 0.3°C, preferably 0.3°C, which improves the model's ability to capture small temperature differences.

[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0044] The original melting curves of the samples to be tested in the valid sample set are divided according to preset interval temperatures to generate binary vectors of temperature bits, and the binary vectors of adjacent temperature bits are arranged continuously, which fully considers the sequential nature of temperature and enables the model to learn the correlation between adjacent temperatures, thereby more accurately capturing the impact of temperature changes on Tm values, which helps to improve prediction accuracy.

[0045] The Tm value marking model is used to preprocess the input data information of the test samples (including the original melting curve), mark whether the test samples have Tm values, and screen out the test samples with Tm values ​​to form a valid sample set, eliminating the interference of samples without Tm values, so that the subsequent prediction model can focus on the analysis of valid samples, reducing the interference of irrelevant data on the model prediction, and thus improving the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings:

[0047] Figure 1 As the main flow chart;

[0048] Figure 2 Schematic diagram of the corresponding relationship between binary vector and temperature. DETAILED DESCRIPTION

[0049] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0050] First embodiment:

[0051] A method for predicting Tm value of melting curve, combined with Figure 1 , including the following steps:

[0052] Based on the machine learning model (which can be any of the XGBoost model, RNN convolutional neural network, MLPRegressor, etc., the XGBoost model is used as an example below), a Tm value labeling model and a Tm value prediction model are constructed;

[0053] Obtaining data information of each test sample in the test sample set, and inputting the data information into a Tm value labeling model; wherein the data information includes an original melting curve;

[0054] The data information is preprocessed by the Tm value labeling model; the Tm value labeling model is used to label the test samples with or without Tm values, and the test samples with Tm values ​​constitute a valid sample set;

[0055] Extracting several peaks (up to 5 peaks) from the original melting curves of the test samples in the above valid sample set to obtain an original Tm value set;

[0056] The original Tm value set is encoded according to the preset interval temperature to obtain a 151-bit original binary vector; wherein the temperature bits in the original binary vector are arranged in a continuous manner according to the interval temperature; for example: if the interval temperature is 0.3°C, the 0th bit of the binary vector is 40.0°C, the 1st bit is 40.3°C, the 2nd bit is 40.6°C, the 3rd bit is 40.9°C, ..., the 49th bit is 54.7°C, the 50th bit is 55°C, the 51st bit is 55.3°C, the 52nd bit is 55.6°C, ..., the 150th bit is 85°C. If two peaks are extracted, namely 55°C and 55.3°C, the 50th bit and the 51st bit of the binary vector are 1 (see Figure 2 ).

[0057] The original melting curves of the samples in the valid sample set are input into the Tm value prediction model to generate a predicted binary vector (i.e., the encoded binary vector of the predicted Tm value). This vector is then passed through a convolutional layer to capture local temperature dependencies. The temperature bits in the predicted binary vector are arranged in intervals.

[0058] The original binary vector and the predicted binary vector are compared to determine whether the predicted binary vector is abnormal.

[0059] The original melting curves of the samples to be tested in the valid sample set are divided according to preset interval temperatures to generate binary vectors of temperature bits, and the binary vectors of adjacent temperature bits are arranged continuously, which fully considers the sequential nature of temperature and enables the model to learn the correlation between adjacent temperatures, thereby more accurately capturing the impact of temperature changes on Tm values, which helps to improve prediction accuracy.

[0060] The Tm value marking model is used to preprocess the input data information of the test samples (including the original melting curve), mark whether the test samples have Tm values, and screen out the test samples with Tm values ​​to form a valid sample set, eliminating the interference of samples without Tm values, so that the subsequent prediction model can focus on the analysis of valid samples, reducing the interference of irrelevant data on the model prediction, and thus improving the accuracy of the prediction.

[0061] A reference scenario involves inputting a raw melting curve containing 186,138 noisy data points into the Tm value labeling model. The model then predicts whether each of these 186,138 noisy data points contains a Tm value and labels it, automatically flagging abnormal samples (i.e., no Tm value, multiple peaks, or low signal-to-noise ratio). 150,000 high-quality, non-abnormal data points labeled with Tm values ​​are then fed into the Tm value prediction model, which then predicts the Tm value for each data point.

[0062] The encoding process is as follows: Given a temperature bit value ∈[40,85], encoding step s=0.3, encoding center point (Explanation: From Start looping until End of loop); then Encoded into 151 binary bits, place 1 at the distance Recent Position, all others are 0; calculate distance The nearest center point code position The specific mathematical formula is: ,in, Indicates the starting point of encoding, here .

[0063] For example: =55.7, s=0.3, passed The result of the center point code position closest to 55.7 is: the 52nd bit is 1, and the others are 0.

[0064] Second embodiment:

[0065] Based on the first embodiment, the training and testing of the Tm value marker model. Based on the XGBoost model, the specific steps of constructing the Tm value marker model include:

[0066] Obtaining historical data information of each first historical sample in the first historical sample set; wherein the historical data information includes an original melting curve and an original mark with or without a Tm value mark;

[0067] Preprocessing the historical data information of each first historical sample in the first historical sample set, the preprocessing including data cleaning (removing irrelevant data, i.e., data source information and experimental process information) and normalization;

[0068] Dividing the preprocessed first historical sample set into a first training set and a first test set;

[0069] The historical data information in the first training set is input into the XGBoost model for training to obtain a Tm value labeling model;

[0070] Inputting the original melting curve of a first historical sample in the first test set into the Tm value marker model to obtain a test marker indicating whether the first historical sample has a Tm value marker;

[0071] The original mark and the test mark of the first historical sample are matched. If the match is successful, the mark of the first historical sample is normal; if the match fails, the mark of the first historical sample is abnormal.

[0072] The XGBoost model learns the mapping relationship from melting curve features to the presence or absence of Tm value markers on the training set and verifies it on the test set to test the learning effect and generalization ability of the Tm value marker model. The original melting curve of the first historical sample in the first test set is input into the Tm value marker model to test the Tm value marker model. The original marker of the first historical sample and the test marker are then matched to verify the accuracy of the Tm value marker model. Problems and deficiencies in the Tm value marker model are promptly identified to provide direction for subsequent model optimization, ensuring that the final constructed Tm value marker model can accurately distinguish samples with and without Tm values, filter out valid sample data for subsequent steps, reduce the interference of invalid data on the prediction model, and thus improve the accuracy of Tm value prediction.

[0073] From the beginning of obtaining historical data information, through preprocessing, data set division, model training, test evaluation and error marking, each link is related and influenced by each other. By continuously circulating this process, the model performance can be gradually optimized to ensure that the Tm value marking model can screen out effective samples, which helps to improve the accuracy and stability of the entire Tm value prediction process, and provides high-quality input data for the subsequent Tm value prediction model. High-quality input data means that the subsequent prediction model can learn based on more relevant and accurate information, thereby improving the accuracy of Tm value prediction; through the synergistic effect of before and after, the entire Tm value prediction process is optimized at each link, and the accuracy of Tm value prediction is ultimately improved.

[0074] A referenceable use case will use 158,672 of the 186,138 data as training data and 27,466 as test data; Each data contains 151 dissolution peak curve data and whether it contains Tm value marking, and data source tracing information. In the binary classification task, XGBoost has the following advantages compared with other models (such as logistic regression, random forest, support vector machine, etc.): high efficiency and scalability, regularization to prevent overfitting, strong flexibility and parameter adjustment, optimization in structural risk minimization, etc.

[0075] In the test, it is found that the hyperparameters of the XGBoost model are set as follows: the maximum number of trees used in the ensemble is 5000, the maximum depth of each tree is 10, the learning step of updating the weight in each iteration is 0.05, and the objective function to be optimized is the probability value between 0 and 1. The log loss function is used to evaluate the model effect. The test results are as follows: among the 27,466 data in the test set, there are 3 data without Tm value predicted out and 30 data with Tm value predicted out. The comprehensive prediction accuracy of the test set is 99.88%.

[0076] Specific embodiments, training and testing of Tm value prediction model. Based on the XGBoost model, the specific steps of constructing the Tm value prediction model include:

[0077] Obtain the historical data information of each second historical sample in the second historical sample set; wherein the historical data information includes the original dissolution curve and the original Tm value;

[0078] Preprocess the historical data information of each second historical sample in the above-mentioned second historical sample set, which includes data cleaning (removing irrelevant data, i.e. data source information and experimental process information) and normalization processing;

[0079] Divide the preprocessed second historical sample set into a second training set and a second test set;

[0080] The historical data information in the second training set is input into the XGBoost model for training to obtain a Tm value prediction model;

[0081] Inputting the original melting curve of a second historical sample in the second test set into the Tm value prediction model to obtain a predicted Tm value of the second historical sample;

[0082] The original Tm value and the predicted Tm value of the second historical sample are matched. If the match is successful, the second historical sample is predicted successfully; if the match fails, the second historical sample is predicted failed, and the second historical sample that fails to match is marked as abnormal data.

[0083] The original melting curve of the second historical sample is input into the Tm prediction model to obtain a predicted Tm value. The original and predicted Tm values ​​are then matched to measure the Tm prediction model's performance on the test set and determine whether the model can accurately predict the Tm value corresponding to the new melting curve. If the match fails, it indicates that the model failed to predict the Tm value for that sample, which helps identify problems and deficiencies in the Tm prediction model, provides direction for subsequent model optimization, and encourages researchers to adjust and improve the model to improve its prediction accuracy.

[0084] Combined with the previously constructed Tm value tagging model, valid samples with Tm values ​​were screened for prediction. This synergistic effect enabled the subsequent prediction model to focus on processing valid data related to Tm values, reducing interference from invalid data. Furthermore, both the tagging model and the prediction model utilized the XGBoost algorithm, ensuring consistency and coherence in data processing and model building. This further improved the accuracy and reliability of the entire prediction process, resulting in more precise Tm value predictions.

[0085] A reference scenario uses 152,339 of the 186,138 data points as training data and 26,307 as test data. Each data point contains 151 melting peak curves and encoded Tm values, along with data source traceability information. In a multi-classification task, the number of encoded data points is used as the number of classifiers to train, based on the Tm encoding data. This means the number of encoders is N, and the multi-classification model consists of N binary classification models.

[0086] During testing, we found that the XGBoost model's hyperparameter settings, including a maximum number of 5,000 trees in the ensemble, a maximum depth of 10 per tree, a learning step of 0.05 for updating weights in each iteration, an optimization objective with a probability between 0 and 1, and a logarithmic loss function for model evaluation, performed well. The test results showed that of the 27,466 data points in the test set, 46 had incorrectly predicted Tm values, resulting in an overall prediction accuracy of 99.2%.

[0087] Third embodiment:

[0088] The difference from the second embodiment is that, during the training and testing phases, the second historical sample set is composed of the first historical samples in the first historical sample set that are originally marked with Tm values.

[0089] In the actual use stage, the data information of the sample to be tested does not contain the original mark of whether or not the Tm value is present. The valid sample input into the Tm value prediction model is the sample to be tested with a Tm value identified by the Tm value marking model, thereby realizing the synergy between the Tm value marking model and the Tm value prediction model.

[0090] In a specific embodiment, preprocessing the historical data information of each first historical sample in the first historical sample set includes removing the first historical samples without Ct values.

[0091] Samples without Tm values ​​indicate that they are not infected with the virus; samples without Ct values ​​indicate that the experiment has failed and the sample has no reference value. The model marked at the Tm value is cleared.

[0092] Fourth embodiment:

[0093] Based on any of the above embodiments, the data information is preprocessed using the above Tm value labeling model, and the specific steps include:

[0094] The above data information also includes data source information and experimental process information of the samples to be tested;

[0095] Clean the above data information to remove the data source information and experimental process information of the samples to be tested;

[0096] The original melting curve in the above data information was normalized.

[0097] While information about the data source and experimental process of the test sample is helpful in understanding the sample background and experimental conditions, it does not directly contribute to the determination of the Tm value and may even become a confounding factor. Removing this information during the data cleaning phase allows the model to focus more on the raw melt curve data, which is closely related to the Tm value. The data in the raw melt curve may have different dimensions and numerical ranges. This makes it difficult to balance the importance of each feature during model training, resulting in slower or even impossible convergence. Normalizing the raw melt curve and mapping all data to a specific interval (such as [0, 1] or [-1, 1]) eliminates the influence of dimension and makes the features numerically comparable. This allows the model to adjust parameters more efficiently during training, accelerating convergence and improving training efficiency.

[0098] Fifth embodiment:

[0099] Based on any of the above embodiments, when the tracing signal is received, the data source information and the experimental process information of the sample to be tested are obtained from the above data information.

[0100] Patient information, experimental parameters (well number, reagent batch, PCR instrument channel) and operation records (test time, operator ID, original curve file) are associated with each sample. Through database indexing, reverse tracking of "diagnostic result → Tm value → experimental conditions" is achieved, supporting rapid location of problems (such as temperature drift caused by a certain batch of reagents), and solving the defect of existing technologies that cannot associate Tm values ​​with original experimental parameters (well position, channel, reagent batch).

[0101] Sixth embodiment:

[0102] Based on any of the above embodiments, the preset interval temperature is 0.05°C to 0.3°C, preferably 0.3°C, and the temperature range (40.0°C to 85.0°C) is divided into intervals of 0.3°C to generate a 151-bit binary vector; thereby improving the model's ability to capture subtle temperature differences.

[0103] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting the Tm value of a melting curve, characterized in that: The following steps are involved: Constructing Tm value labeling model and Tm value prediction model; Obtaining data information of each test sample in the test sample set, and inputting the data information into a Tm value labeling model; wherein the data information includes an original melting curve; The data information is preprocessed by the Tm value marking model; the Tm value marking model is used to mark whether the test sample has a Tm value, and the test samples with Tm values ​​constitute a valid sample set; Extracting a number of peaks from the original melting curves of the samples to be tested in the valid sample set to obtain an original Tm value set; Encoding the original Tm value set according to a preset interval temperature to obtain an original binary vector; wherein the temperature bits in the original binary vector are connected and arranged according to the interval temperature; Inputting the original melting curve of the sample to be tested in the valid sample set into the Tm value prediction model to obtain a prediction binary vector; wherein the temperature bits in the prediction binary vector are connected and arranged according to interval temperatures; The original binary vector and the predicted binary vector are compared to determine whether the predicted binary vector is abnormal.

2. The method for predicting the Tm value of a melting curve according to claim 1, wherein The specific steps of constructing the Tm value labeling model include: Obtaining historical data information of each first historical sample in the first historical sample set; wherein the historical data information includes an original melting curve and an original mark indicating whether a Tm value mark is present; preprocessing the historical data information of each first historical sample in the first historical sample set; Dividing the preprocessed first historical sample set into a first training set and a first test set; Inputting the historical data information in the first training set into the machine learning model for training to obtain a Tm value labeling model; Inputting the original melting curve of a first historical sample in the first test set into the Tm value marker model to obtain a test marker indicating whether the first historical sample has a Tm value marker; The original mark and the test mark of the first historical sample are matched. If the match is successful, the mark of the first historical sample is normal; if the match fails, the mark of the first historical sample is abnormal.

3. The method for predicting the Tm value of a melting curve according to claim 1 or 2, wherein: The specific steps of constructing the Tm value prediction model include: Obtaining historical data information of each second historical sample in the second historical sample set; wherein the historical data information includes an original melting curve and an original Tm value; preprocessing the historical data information of each second historical sample in the second historical sample set; Dividing the preprocessed second historical sample set into a second training set and a second test set; Inputting the historical data information in the second training set into the machine learning model for training to obtain a Tm value prediction model; Inputting the original melting curve of a second historical sample in the second test set into the Tm value prediction model to obtain a predicted Tm value of the second historical sample; The original Tm value and the predicted Tm value of the second historical sample are matched. If the match is successful, the prediction of the second historical sample is successful; if the match fails, the prediction of the second historical sample is failed.

4. The method for predicting the Tm value of a melting curve according to claim 3, wherein The second historical sample set is composed of the first historical samples in the first historical sample set that are originally marked with Tm values.

5. The method for predicting the Tm value of a melting curve according to claim 3, wherein Preprocessing the historical data information of each first historical sample in the first historical sample set includes removing the first historical samples without Ct values.

6. The method for predicting the Tm value of a melting curve according to claim 1, wherein The data information is preprocessed by the Tm value labeling model, and the specific steps include: The data information also includes data source information and experimental process information of the sample to be tested; Cleaning the data information to remove data source information and experimental process information of the sample to be tested; The original melting curve in the data information is normalized.

7. The method for predicting the Tm value of a melting curve according to claim 6, wherein: When the tracing signal is received, the data source information and the experimental process information of the sample to be tested are obtained from the data information.

8. The method for predicting the Tm value of a melting curve according to claim 1, wherein The preset interval temperature is 0.05°C to 0.3°C.

Images