A decision tree model for detecting primer dimer and application thereof

Abstract

The present invention provides a method for detecting primer dimers, and in particular, the new method reduces false positive and false negative results that are typically generated when detecting primer dimers. The method for detecting primer dimers of the present invention accurately detects primer dimers and effectively evaluates primers.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of primer dimer detection, and particularly relates to a decision tree model for detecting primer dimer and application thereof. BACKGROUND

[0002] Primer dimer is a pair of primers or 3' end partial base complementary combination of primers, which is formed into a small molecular weight double-stranded nucleotide fragment under the action of an enzyme. When the 3' end of the PCR primer has significant complementarity, the formation of primer dimer is promoted, thereby reducing the yield of PCR products. In addition, the formation of primer dimer can also lead to more serious problems, such as non-specific amplification due to primer imbalance. Therefore, detecting primer dimer is particularly important for designing and optimizing primers.

[0003] The conventional primer dimer detection software determines whether the primer will produce dimer according to the score or ΔG value of the base complementarity between primers, so as to determine that the primer is a primer dimer as long as the score or ΔG value threshold set by the software is reached, regardless of whether the base complementarity occurs at the 5' end, 3' end or middle part of the primer. Therefore, the 3' end primer dimer is often ignored as the cause of the PCR amplification problem. Therefore, the primer dimer results detected by the conventional primer dimer detection software will inevitably have a large number of false positives and false negatives.

[0004] The more authoritative primer evaluation software oligo 7 cannot directly determine whether the input primer will produce primer dimer, but gives all topological structures of the input primer that may produce dimer and the corresponding ΔG value, which needs to be judged by the user; and the software cannot batch evaluate primer dimer. For researchers who need to evaluate a large number of primers, this undoubtedly greatly increases the time cost of researchers.

[0005] Therefore, how to provide a primer dimer detection method that can not only batch evaluate primer dimer, but also obtain a primer dimer detection result with high accuracy has become a problem to be solved in the industry. SUMMARY

[0006] In view of the shortcomings of the existing primer dimer detection technology, the conventional primer dimer algorithm is upgraded in the present application, which can not only batch evaluate primer dimer for primer sequences, but also has high accuracy of detected primer dimer, further reducing the occurrence of non-specific amplification.

[0007] The method for detecting primer dimer of the present application comprises the following steps:

[0008] 1. Constructing a decision tree model

[0009] The preparation of the decision tree model specifically comprises: selecting real dimer and non-dimer samples in the training set; selecting a certain number of dimer data amplified by real primers as a data set of the training model; calculating the number of mismatches at the 5' end, the number of mismatches at the 3' end, the number of bindings at the 3' end and the number of bases of total binding sites; taking the four variables as response variables; and training the training set to obtain the decision tree model.

[0010] 2. Application of the decision tree model

[0011] Two primer sequences are taken as input, the dimer binding between the two sequences is calculated and screened first, if the number of mismatches at the 5' end is 0, then the number of bindings at the 3' end is observed, if it is greater than or equal to 10, it is considered to be a dimer, if not, the total number of bindings is observed, if it is greater than or equal to 6, the number of bindings at the 3' end is observed again, if it is greater than 6 and the number of mismatches at the 3' end is not greater than 1, it is considered to be a dimer. If the number of bindings at the 3' end is not greater than 6 and there is no mismatch, it is also considered to be a dimer.

[0012] All primer sequences are calculated for the number of mismatches at the 5' end, the number of mismatches at the 3' end, the number of bindings at the 3' end and the number of bases of total binding sites, and then the four variables are taken into the decision tree model to obtain 1 and 0, 1 indicating that there is a dimer between the primers and 0 indicating that there is no dimer between the primers.

[0013] Compared with the prior art, the present application has the beneficial effects that: the present application can batch detect primers for dimers, even if there are thousands of primer sequences, the detection can be completed in a short time, the accuracy of the detected primer dimers based on the present application is high, and the non-specific amplification and amplification failure caused by the primer sequence producing dimers in the PCR amplification process are effectively avoided. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 is a flow chart of the application of the decision tree model;

[0015] Figure 2 is a flow chart of the detection of primer dimers using the decision tree model;

[0016] Figure 3 is the evaluation results of 5 primer dimer detection software, the accuracy = the number of real positive dimers / the number of all dimers; the sensitivity = the number of real positive dimers / the number of all positive dimers; the F value = 2 * (1 / (1 / accuracy + 1 / sensitivity)). DETAILED DESCRIPTION

[0017] In order to deepen the understanding of the present application, the present application will be further described in detail below in combination with examples, which are only used to explain the present application and do not constitute a limitation to the protection scope. The instruments and equipment, consumables and reagents used are commercially available products, except for special instructions.

[0018] Example 1

[0019] The present application uses 209 published new coronavirus primer sequences as negative data, 74 primer sequences that produce real dimer sequences in the sequencing process as positive data, and evaluates the primer dimer detection results of a total of 5 software including the present application. The evaluation basis is the accuracy, sensitivity and F value of the detection results, wherein the accuracy is the ratio of the real positive dimer results to all dimer results; the sensitivity is the ratio of the real positive dimer results to all positive dimer results; and the formula of F value is 2*(1 / (1 / accuracy+1 / sensitivity)).

[0020] I. Use data

[0021] ① Negative data: 209 published new coronavirus primer sequences.

[0022] Table 1. Negative data primer sequences

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029] ② Positive data: 74 primer sequences that produce real dimer sequences in the sequencing process.

[0030] Table 2. Positive data primer sequences

[0031]

[0032]

[0033] II. Software evaluation

[0034] Table 3. Primer dimer software used for evaluation

[0035]

[0036]

[0037] Note: "DIMER" is the software of the present application

[0038] III. Evaluation results

[0039] Table 4. Evaluation results of the primer dimer detection software of the fifth aspect (1)

[0040]

[0041] Table 5. Evaluation results of the primer dimer detection software of the fifth aspect (2)

[0042]

[0043] The evaluation results of the primer dimer detection software of the fifth aspect are also shown in Figure 3 . Figure 3 Accuracy = number of true positive dimers / total number of dimers; Sensitivity = number of true positive dimers / total number of positive dimers; F value = 2 * (1 / (1 / accuracy + 1 / sensitivity)).

[0044] The evaluation results show that the accuracy, sensitivity and F value of the primer dimer detection method of the present application are the highest, which is obviously superior to the conventional primer dimer detection software.

[0045] Although the embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1.A method for constructing a decision tree model for detecting primer dimer, the method comprising: selecting real dimer and non-dimer samples in a training set: selecting a certain number of real primer amplified dimer data as a data set of the training model, calculating the number of mismatches at the 5' end of the primer, the number of mismatches at the 3' end, the number of 3' end bindings and the number of bases at the total binding site, taking the four variables as response variables, and training the training set to obtain the decision tree model; The decision tree model is: two primer sequences as input, first calculate the dimer binding between the two sequences, and screen, if the number of 5' end mismatches is 0, then look at the number of 3' end bindings, if it is greater than or equal to 10, it is considered to be a dimer, if not, look at the total binding number, if it is greater than or equal to 6, look at the number of 3' end bindings, if it is greater than 6, and the number of 3' mismatches is not greater than 1, it is considered to be a dimer;If the number of 3' end bindings is not greater than 6, and there is no mismatch, it is also considered to be a dimer; All primer sequences will calculate the number of mismatches at the 5' end and the number of mismatches at the 3' end between each other, the number of 3' end bindings and the number of bases at the total binding site, then bring the four variables into the decision tree model, get 1 and 0, 1 indicates that there is a dimer between the primers, 0 indicates that there is no dimer between the primers. 2.A decision tree model for detecting primer dimer, the decision tree model is constructed by the method of claim 1. 3.The decision tree model of claim 2 is used for detecting primer dimer. 4.A method for detecting primer dimer, comprising the following steps: constructing a decision tree model, the preparation of the decision tree model specifically comprises: selecting real dimer and non-dimer samples in a training set: selecting a certain number of real primer amplified dimer data as a data set of the training model, calculating the number of mismatches at the 5' end of the primer, the number of mismatches at the 3' end, the number of 3' end bindings and the number of bases at the total binding site, taking the four variables as response variables, and training the training set to obtain the decision tree model; The decision tree model is applied, two primer sequences as input, first calculate the dimer binding between the two sequences, and screen, if the number of 5' end mismatches is 0, then look at the number of 3' end bindings, if it is greater than or equal to 10, it is considered to be a dimer, if not, look at the total binding number, if it is greater than or equal to 6, look at the number of 3' end bindings, if it is greater than 6, and the number of 3' mismatches is not greater than 1, it is considered to be a dimer;If the number of 3' end bindings is not greater than 6, and there is no mismatch, it is also considered to be a dimer; All primer sequences will calculate the number of mismatches at the 5' end and the number of mismatches at the 3' end between each other, the number of 3' end bindings and the number of bases at the total binding site, then bring the four variables into the decision tree model, get 1 and 0, 1 indicates that there is a dimer between the primers, 0 indicates that there is no dimer between the primers.

Images