Methods, systems, devices, and media for designing oligonucleotide sequences
By designing oligonucleotide sequences using a large language model, the problems of low efficiency and poor accuracy in existing primer and probe design methods are solved, achieving efficient primer design and probe capture, and improving the specificity and sensitivity of PCR amplification and target detection.
Patent Information
- Application Number
- CN202510591291.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-08
- Publication Date
- 2026-11-17
AI Technical Summary
Existing primer and probe design methods suffer from low efficiency, poor accuracy, and insufficient probe sensitivity in fields such as high-throughput sequencing and target-specific capture, making it difficult to meet the needs of rapid amplification detection and multiple primer set design.
We employ a large language model to design oligonucleotide sequences, generate candidate primer or probe combinations, use a sliding window and preset constraints to screen for the optimal sequence, and combine experimental feedback to optimize and iterate, thereby improving the efficiency of primer design and the sensitivity of probe capture.
It significantly improves the efficiency and accuracy of primer design, enhances the specificity of PCR amplification and the sensitivity of probe capture, and meets the needs of multiplex PCR amplification and target detection.
Smart Images

Figure HDA0005393250240000011 
Figure HDA0005393250240000012 
Figure HDA0005393250240000013
Abstract
Description
Technical Field
[0001] This application relates to the field of bioinformatics, specifically to a method, system, device, and medium for designing oligonucleotide sequences. Background Technology
[0002] Primer and probe design is a crucial step in modern molecular biology, genetic diagnostics, and genomics research. With the rapid development of high-throughput sequencing and molecular diagnostic technologies, primer and probe design has evolved from manual methods to automated and intelligent approaches, supporting scenarios such as efficient PCR amplification and target-specific capture (e.g., exon probes). Primer design technology plays a vital role in the rapid amplification and detection of SARS-CoV-2, extended carrier screening of the human genome (ECS), and the detection of drug resistance mutations in pathogens such as Mycobacterium tuberculosis. This technology is also widely used in viral genome sequencing and enzyme mutation analysis. Traditional methods such as Primer3 and PrimalScheme provide effective solutions, but much room for optimization remains. Summary of the Invention
[0003] This application aims to address at least one of the technical problems existing in the prior art to a certain extent. To this end, this application provides a method, system, apparatus, and medium for designing oligonucleotide sequences. This method develops an efficient and multifunctional primer design approach, overcoming the application bottlenecks of LLM in tNGS, WES, WGS, and mutation design, enabling it to become an expert in primer design; optimizing the probability of primer dimer formation; improving probe hybridization stability; increasing target region coverage and amplification uniformity; improving probe detection sensitivity; accelerating the design and optimization of multiple primer sets; constructing an experimental feedback-driven iterative optimization mechanism, utilizing experimental feedback for multiple rounds of primer design iteration to improve target region coverage and uniformity; and achieving primer specificity optimization in the context of large-scale genomes.
[0004] In a first aspect, this application proposes a method for designing oligonucleotide sequences. According to an embodiment of this application, the method includes: receiving user prompts based on a large language model, the prompts being related to oligonucleotide sequence application methods; determining the application methods of the oligonucleotides in the prompts using the large language model, wherein different application methods have a first mapping relationship with different scenario agents, each scenario agent encapsulating processing logic for the corresponding application method, the processing logic being selected from at least one of the following: generating one or more sets of candidate primers or capture probes for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites; wherein the processing logic includes multiple different types of databases and publicly available online information to be retrieved; generating a retrieval strategy through a scenario agent having a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases, and publicly available online information; retrieving and processing the databases and publicly available online information according to the retrieval keywords to generate nucleic acid sequence information.
[0005] Therefore, the method of this application systematically designs highly efficient oligonucleotide sequences to achieve multiplex PCR amplification or probe capture of multiple target regions. Specifically, for primer design, multiple candidate primers are first generated based on one or more gene target regions. Then, multiple candidate primer pair combinations are determined by combining these candidate primers. Finally, a large language model is used to screen the optimal combination from the candidate primer pair combinations, thereby obtaining a primer set capable of efficiently amplifying multiple target regions. For probe design, candidate probe sets are first generated based on the target region. A large language model is used to screen the optimal combination from the candidate probe sets, thereby constructing a probe set suitable for target detection. This process not only improves the efficiency and accuracy of primer or probe design but also significantly enhances the specificity and success rate of subsequent PCR amplification, or the sensitivity and specificity of probe capture.
[0006] According to embodiments of this application, the target region includes coding regions, regions containing drug-resistant mutations, species-specific regions for identification, methylation modification regions, and tumor hotspot mutation regions, etc. The oligonucleotide sequence can be used for various applications such as target gene or target region amplification, whole-genome amplification or capture, site-directed gene mutagenesis, and protein mutant design. According to embodiments of this application, a large language model can receive user prompts multiple times during the completion of a design task. The large language model can comprehensively analyze the user's application method for designing nucleic acid sequences based on multiple prompts.
[0007] According to embodiments of this application, the system intelligently analyzes user needs or application scenarios, maps the scenarios to intelligent agents, and then selects appropriate databases to retrieve and generate target areas.
[0008] According to an embodiment of this application, a sliding window is used to generate a nucleic acid sequence based on the target region, and the nucleic acid sequence that meets the preset constraints is selected as the candidate nucleic acid sequence.
[0009] According to an embodiment of this application, if no nucleic acid sequence meets the constraint condition within 20 consecutive sliding windows, the constraint condition is relaxed.
[0010] According to embodiments of this application, the constraints include at least one of GC content, repeatability tolerance, nucleotide diversity, Tm value, and secondary structure.
[0011] According to an embodiment of this application, the method further includes: performing a multiplex PCR amplification experiment using the candidate nucleic acid sequence, and evaluating the performance of the multiplex PCR amplification experiment; and adjusting a first weighting constant and a second weighting constant based on the performance evaluation.
[0012] In a second aspect, this application proposes an oligonucleotide design system based on a large language model. According to an embodiment of this application, the system includes: a receiving module for receiving user prompts based on a large language model, the prompts being related to oligonucleotide sequence application methods; a generation module for determining the application methods of the oligonucleotides in the prompts through the large language model, wherein different application methods have a first mapping relationship with different scenario agents, each scenario agent encapsulating processing logic for the corresponding application method, the processing logic being selected from at least one of the following: generating one or more sets of candidate primers for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites; wherein the processing logic includes multiple different types of databases and publicly available online information to be retrieved; and a retrieval module for generating a retrieval strategy through scenario agents having a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases, and publicly available online information; and retrieving and processing the databases and publicly available online information according to the retrieval keywords to generate nucleic acid sequence information.
[0013] In a third aspect, this application proposes an electronic device. According to an embodiment of this application, the electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the method described in the first aspect.
[0014] In a fourth aspect, this application provides a readable storage medium. According to an embodiment of this application, the readable storage medium stores computer-readable instructions for causing a computer to perform the method described in the first aspect.
[0015] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description
[0016] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:
[0017] Figure 1 This is a flowchart of the primer design process in Embodiment 1 of this application;
[0018] Figure 2 This is a schematic diagram of the LLM optimization iteration in Embodiment 1 of this application;
[0019] Figure 3 This is a comparison chart of the present application with other optimizers in Embodiment 1 of this application;
[0020] Figure 4 This is an electrophoresis diagram of the SARS-CoV-2 virus sequencing and detection experimental results in Example 2 of this application;
[0021] Figure 5 This is a graph showing the coverage results of the SARS-CoV-2 virus primer set in Example 2 of this application;
[0022] Figure 6 This is a graph showing the results of the Extended Carrier Screening (ECS) in Embodiment 3 of this application, wherein, Figure 6 (a) is a screening sequencing electrophoresis image; Figure 6 (b) shows the sequencing analysis results;
[0023] Figure 7 This is a graph showing the detection results of Mycobacterium tuberculosis in Example 4 of this application, wherein... Figure 7 (a) is a graph showing the results of uniformity, specificity, and dimerization rate of the designed primer set; Figure 7 (b) is a graph of standardized sequencing reads; Figure 7 (c) Read the ratio result map for a specific target area;
[0024] Figure 8 This is a diagram showing the enzyme mutation recognition sequencing results in Example 5 of this application, wherein... Figure 8 (a) is a graph showing the first round of sequencing results for the enzyme TdT; Figure 8 (b) shows the results of the second round of sequencing of the enzyme TdT; Figure 8(c) shows the results of primer set uniformity, specificity, and dimerization rate for four important enzymes. Detailed Implementation
[0025] The embodiments of this application are described in detail below. The embodiments described below are exemplary and are only used to explain this application, and should not be construed as limiting this application.
[0026] It should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more.
[0027] The endpoints and any values of the ranges disclosed herein are not limited to the precise ranges or values, and these ranges or values should be understood to include values close to these ranges or values. For numerical ranges, the endpoint values of the various ranges, the endpoint values of the various ranges and individual point values, and individual point values can be combined with each other to obtain one or more new numerical ranges, which should be considered as specifically disclosed herein.
[0028] In this document, the terms “comprising” or “including” are open-ended expressions, meaning that they include the contents specified in this application but do not exclude other contents.
[0029] In this document, the terms “optionally,” “optionally,” or “optionally” generally refer to an event or condition that may, but may not, occur, and the description includes both cases in which the event or condition occurs and cases in which the event or condition does not occur.
[0030] Terms and Definitions
[0031] In this paper, the term "primer dimer rate" is the ratio of reads shorter than 70 bp to all reads after low-quality control. A threshold for the primer dimer rate can be set to customize redesign conditions.
[0032] In this paper, the term "target region ratio" refers to the ratio of the number of reads matched to the target region to the total number of reads after low-quality control processing. The threshold of the target region ratio can be set to customize the redesign conditions.
[0033] In this paper, the term "0.1×uniformity" refers to the percentage of regions whose read count exceeds 0.1 times the average read count of all target regions.
[0034] In this paper, the term "large language model" refers to a large-scale natural language processing model trained using deep learning techniques. These models typically have billions or even hundreds of billions of parameters and are capable of understanding and generating natural language text. Through unsupervised learning on large-scale text datasets, they learn the syntax, semantics, and context of a language, thus possessing powerful language understanding and generation capabilities. Examples include Qwen, GLM, and ChatGPT.
[0035] In this paper, the term "Saddle algorithm" refers to an algorithm for designing highly multiplex PCR primer sets that optimizes primer design by using simulated annealing combined with primer dimer formation probability.
[0036] Methods for designing oligonucleic acids
[0037] In a first aspect, this application proposes a method for designing oligonucleotide sequences. According to an embodiment of this application, the method includes: receiving user prompts based on a large language model, the prompts being related to oligonucleotide sequence application methods; determining the application methods of the oligonucleotides in the prompts using the large language model, wherein different application methods have a first mapping relationship with different scenario agents, each scenario agent encapsulating processing logic for the corresponding application method, the processing logic being selected from at least one of the following: generating one or more sets of candidate primers or capture probes for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites; wherein the processing logic includes multiple different types of databases and publicly available online information to be retrieved; generating a retrieval strategy through a scenario agent having a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases, and publicly available online information; retrieving and processing the databases and publicly available online information according to the retrieval keywords to generate nucleic acid sequence information.
[0038] Therefore, the method of this application systematically designs efficient oligonucleotide sequences to achieve multiplex PCR amplification or probe capture of multiple target regions. Specifically, for primer design, multiple candidate primers are first generated based on one or more gene target regions. Then, multiple candidate primer pair combinations are determined by combining these candidate primers. Finally, a large language model is used to screen the optimal combination from the candidate primer pair combinations, thereby obtaining a primer set capable of efficiently amplifying multiple target regions. For probe design, candidate probe sets are first generated based on the target region. A large language model is used to screen the optimal combination from the candidate probe sets, thereby constructing a probe set suitable for target detection. This process not only improves the efficiency and accuracy of primer or probe design, but also significantly enhances the specificity and success rate of subsequent PCR amplification, or the sensitivity and specificity of probe capture. The target regions include coding regions, regions containing drug-resistant mutations, species-specific regions, methylation modification regions, tumor hotspot mutation regions, etc. The oligonucleotide sequences can be used for various applications such as target gene or target region amplification, whole-genome amplification or capture, site-directed gene mutagenesis, and protein mutant design. Target sites include nucleotides and / or amino acids.
[0039] According to embodiments of this application, the above method may also have at least one of the following additional technical features:
[0040] According to an embodiment of this application, a large language model can receive user prompts multiple times during the completion of a design task, wherein the large language model can comprehensively analyze the application method of the user's designed nucleic acid sequence based on multiple prompts.
[0041] According to embodiments of this application, the system intelligently analyzes user needs or application scenarios, maps the scenarios to intelligent agents, and then selects appropriate databases to retrieve and generate target areas.
[0042] According to embodiments of this application, a sliding window is used to generate nucleic acid sequences based on the target region, and nucleic acid sequences that meet preset constraints are selected as candidate nucleic acid sequences. Thus, by using a sliding window to generate primer sequences on the target region and filtering out sequences that meet the preset constraints, multiple candidate nucleic acid sequences are determined, providing a foundation for subsequent nucleic acid sequence optimization.
[0043] According to some specific embodiments of this application, for primer design, a sliding window is used to generate primer sequences in the upstream and downstream sequences respectively, and primer sequences that meet preset constraints are selected as candidate upstream primers and multiple candidate downstream primers. For probe design, a sliding window is used to generate probe sequences from the target sequence, and probe sequences that meet preset constraints are selected as candidate probes.
[0044] According to embodiments of this application, when no primer sequence satisfying the constraints is found within 20 consecutive sliding windows, the constraints are relaxed. Thus, when no primer sequence satisfying the preset constraints is found within 20 consecutive sliding windows, the diversity of primer sequences is increased by gradually relaxing the GC content range, increasing repetition tolerance, reducing nucleotide diversity requirements, adjusting the sliding window step size, and expanding the primer length range. This increases the likelihood of finding suitable primer sequences, ensuring the flexibility and success rate of primer design, while also guaranteeing primer quality and performance to a certain extent.
[0045] According to some specific embodiments of this application, for probe design, the constraints are relaxed in high GC regions and complex structural regions, including reducing the step size, where the complex structural regions include tumor hotspots.
[0046] According to embodiments of this application, the constraints include at least one of GC content, repeatability tolerance, nucleotide diversity, Tm value, secondary structure, and specificity loss. According to some specific embodiments of this application, the constraints for primer design include at least one of GC content, repeatability tolerance, nucleotide diversity, Tm value, and secondary structure; the constraints for probe design include GC content, Tm value, secondary structure, and specificity loss. Thus, by setting constraints such as GC content (0.4–0.62), repeatability tolerance, and nucleotide diversity (0.1–0.4), the generated nucleic acid sequences are screened to ensure that the selected candidate nucleic acid sequences are physicochemically stable and have good amplification specificity and efficiency.
[0047] According to embodiments of this application, the optimal primer pair combination is obtained by using dimer loss as the optimization criterion, or the optimal probe combination is obtained by using probe specificity loss as the optimization criterion, and by iterative optimization using the large language model. Thus, the optimal primer pair combination is selected from multiple primer pairs to obtain a primer set for multiplex PCR amplification of multiple target regions. Through the optimization capability of the large language model, the probability of primer dimer formation can be effectively reduced, the coverage of target regions and amplification uniformity can be improved, thereby significantly improving the quality of primer design and the efficiency of subsequent PCR amplification.
[0048] According to an embodiment of this application, the dimer loss is determined based on the following formula: In this model, Dimer Rate is calculated using the Saddle algorithm; BLAST Score is the result of BLAST alignment of each candidate primer pair with the background database; w1 is the first weighting constant; w2 is the second weighting constant; P is the panel; and Ai is the amplicon. Therefore, primer pair design is evaluated and optimized by calculating the dimer loss, which is obtained by weighted summation of the primer dimer rate (DimerRate) and BLAST score (BLAST Score). The primer dimer rate, calculated using the Saddle algorithm, reflects the probability of dimer formation between primers, while the BLAST score, obtained by BLAST alignment of each candidate primer pair with the background database, reflects the similarity between the primer and non-target sequences. The weighting constants w1 and w2 are used to balance the influence of both in the loss function. Custom design conditions can be defined, and by minimizing this loss function, the optimal primer pair combination can be selected.
[0049] According to embodiments of this application, the method further includes: performing multiplex PCR amplification experiments using the candidate nucleic acid sequence and evaluating the performance of the multiplex PCR amplification experiments; and adjusting a first weighting constant and a second weighting constant based on the performance evaluation. The probe set is used to perform the probe capture experiment, and the performance of the probe capture experiment is evaluated. Thus, through an experimental feedback mechanism, the performance of the primer set in practical applications is evaluated, such as primer dimerization rate, target region ratio, 0.1× uniformity, probe specificity, and other indicators are analyzed. Then, based on these performance evaluation results, the weighting constants in the loss function are adjusted to optimize the parameters of primer or probe design, thereby improving the quality and performance of the primer or probe set in subsequent primer or probe design, and achieving more efficient multiplex PCR amplification or probe capture.
[0050] Oligonucleotide Design System Based on Large Language Model
[0051] In a second aspect, this application proposes an oligonucleotide design system based on a large language model. According to an embodiment of this application, the system includes: a receiving module for receiving user prompts based on a large language model, the prompts being related to oligonucleotide sequence application methods; a generation module for determining the application methods of the oligonucleotides in the prompts through the large language model, wherein different application methods have a first mapping relationship with different scenario agents, each scenario agent encapsulating processing logic for the corresponding application method, the processing logic being selected from at least one of the following: generating one or more sets of candidate primers for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites; wherein the processing logic includes multiple different types of databases and publicly available online information to be retrieved; and a retrieval module for generating a retrieval strategy through scenario agents having a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases, and publicly available online information; and retrieving and processing the databases and publicly available online information according to the retrieval keywords to generate nucleic acid sequence information.
[0052] Those skilled in the art will understand that the features and advantages described above for the methods used to design oligonucleotides also apply to this system, and will not be repeated here.
[0053] In a third aspect, this application proposes an electronic device. According to an embodiment of this application, the electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the method described in the first aspect.
[0054] Those skilled in the art will understand that the features and advantages described above for the methods used to design oligonucleotides also apply to this electronic device, and will not be repeated here.
[0055] In a fourth aspect, this application provides a readable storage medium. According to an embodiment of this application, the readable storage medium stores computer-readable instructions for causing a computer to perform the method described in the first aspect.
[0056] Those skilled in the art will understand that the features and advantages described above for the methods used to design oligonucleotides also apply to this readable storage medium, and will not be repeated here.
[0057] The following will explain the solution of this application with reference to embodiments. Those skilled in the art will understand that the following embodiments are for illustrative purposes only and should not be considered as limiting the scope of this application. Where specific techniques or conditions are not specified in the embodiments, they are performed according to the techniques or conditions described in the literature in the art or according to the product instructions. Reagents or instruments whose manufacturers are not specified are all conventional products that can be obtained commercially.
[0058] Example 1
[0059] Primer design process as follows Figure 1 As shown, the user first obtains the target region for PCR amplification and whether it contains a set of SNPs (single nucleotide polymorphisms). A sliding window of 40 base pairs is then used to move within this region to generate primer candidate sequences that conform to the "candidate generation strategy." This process is called "candidate generation." If no valid candidate sequence is found within 20 sliding positions, the design constraints (e.g., GC content, repeatability tolerance, nucleotide diversity, etc.) are gradually relaxed. The termination position of each primer is limited by the maximum and minimum primer length. Once the start and termination positions are determined and forward and reverse primer candidate sequences are stored, the generation process continues until the target region is completely covered. Primer candidate sequences are generated within a 40-base-pair sliding window. Sequences with lengths between the minimum and maximum design lengths are first screened, and then filtered using Gibbs free energy. Sequences with reverse complementarity exceeding 5 base pairs are then filtered out. The remaining candidate sequences are selected based on GC content (0.4–0.62) and nucleotide diversity (0.1–0.4), and sequences with repetitive patterns are discarded. When the 40bp sliding window stops sliding (i.e., when candidate primers that meet the requirements have been generated), k candidate primers will be generated. The LLM optimization iteration diagram is shown below. Figure 2 As shown. For each amplicon (plex), there are two sliding windows, generating k f One positive primer and k r Using a reverse primer, the result will be k. f ·k r There are n candidate pairs. Therefore, given n amplicones in a panel, the search space for the optimization problem is as follows: Where ai is the total number of candidates for amplicon i, a randomly selected panel can be represented as: P = {Ai | Ai ∈ N, 1 <= Ai <= ai, 1 <= i <= n}, where Ai is the index of the candidate selected from amplicon i, and n is the number of amplicons in the panel. The dimer loss is expressed as: Dimer Rate is calculated using the Saddle algorithm, and BLAST Score is the result of BLASTing each candidate (indexed by Ai) against the background database. Each panel P is represented as a list of Ai. In the LLM optimization method, through designed hints, including guidance for LLM to optimize the panel design based on known input-function value pairs (such as historical panel vectors), the lowest loss function value is found, and new panel candidates are output. LLM iteratively generates 10 new panels P, which satisfy the candidate index constraint 1 ≤ Ai ≤ ai, and then evaluates them using the loss function. The optimization objective is to minimize the loss function for a given search space. In each round of optimization, the panel with the smallest loss among the 10 new panels generated by LLM is added to the "panel history vector" [P1, P2, ..., Pm]. In each round of optimization, LLM allows modification of one to three variables in each panel. The panel history vector continues to collect a list of gradually decreasing losses and removes those with higher losses.
[0060] Taking greedy, Adalead, LLM, and GA as examples to optimize tasks with 12 and 78 amplicon counts, the optimization results are as follows: Figure 3 As shown.
[0061] Example 2
[0062] In this embodiment, a 131-key primer set was designed to cover the entire SARS-CoV-2 viral genome, and the results were validated by comparison with traditional methods (ARTIC v5.3.2, PrimeGen-Primer3). Virus sequencing and detection experimental results are as follows: Figure 4 As shown, the primer set designed using the method in Example 1 achieved a coverage rate of 98.9%. Figure 5 As shown, it is significantly superior to other methods.
[0063] Example 3
[0064] In this embodiment, extended carrier screening (ECS) was performed. Primer sets were designed for 35 major clinical diseases using the method described in Example 1, covering the exon regions of the target genes. The screening sequencing results are as follows: Figure 6 (a) shows a depth coverage of 97.85%, and the designed primer set exhibits good uniformity, specificity, and dimerization rate. Figure 6 As shown in (b), it meets the requirements for subsequent bioinformatics analysis.
[0065] Example 4
[0066] In this embodiment, a multiple primer set was designed using the method of Example 1 to target drug resistance sites in Mycobacterium tuberculosis, covering 1200 drug resistance mutation regions. The results of the uniformity, specificity, and dimerization rate of the designed primer set are as follows: Figure 7As shown in (a), the MTR value reached 88.8%, and the standardized sequencing reads are as follows. Figure 7 As shown in (b), the read ratio results for a specific target area are as follows: Figure 7 As shown in (c), it demonstrates excellent detection capabilities.
[0067] Example 5
[0068] In this embodiment, primer sets targeting four important enzymes (Luc, KODm, Cdi1, TdT) were designed using the method of Example 1. The two rounds of sequencing results for the TdT enzyme are as follows: Figure 8 (a) and Figure 8 As shown in (b), the results of primer set uniformity, specificity, and dimerization rate for the four important enzymes are as follows: Figure 8 As shown in (c), it demonstrates excellent detection capabilities.
[0069] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0070] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.
Claims
1. A method for designing oligonucleotide sequences, characterized in that, The method includes: User prompts are received based on a large language model, and these prompts are related to the application methods of oligonucleotide sequences. The application method of oligonucleotides in the prompt words is determined by a large language model. Different application methods have a first mapping relationship with different scene agents. Each scene agent encapsulates the processing logic of the corresponding application method. The processing logic is selected from at least one of the following: generating one or more sets of candidate primers or capture probes for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, and generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites. The processing logic described therein includes multiple different types of databases and publicly available online information to be retrieved; A retrieval strategy is generated by a scene agent that has a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases, and publicly available online information; The database and publicly available online information are retrieved and processed based on the search keywords to generate nucleic acid sequence information.
2. The method according to claim 1, characterized in that, Based on the large language model, multiple user prompts can be received during the completion of a design task. The large language model can comprehensively analyze the application method of the user's designed nucleic acid sequence based on multiple prompts.
3. The method according to claim 2, characterized in that, The system intelligently analyzes user needs or application scenarios, maps these scenarios to intelligent agents, and then selects appropriate databases to retrieve and generate target areas.
4. The method according to claim 1 or claim 3, characterized in that, A sliding window is used to generate nucleic acid sequences based on the target region, and the nucleic acid sequences that meet preset constraints are selected as candidate nucleic acid sequences.
5. The method according to claim 4, characterized in that, If no nucleic acid sequence meets the constraints within 20 consecutive sliding windows, the constraints are relaxed.
6. The method according to claim 4, characterized in that, The constraints include at least one of GC content, repeatability tolerance, nucleotide diversity, Tm value, and secondary structure.
7. The method according to claim 6, characterized in that, Further includes: The candidate nucleic acid sequence was used to perform multiplex PCR amplification experiments, and the performance of the multiplex PCR amplification experiments was evaluated. and The first and second weighting constants are adjusted based on the performance evaluation.
8. An oligonucleotide design system based on a large language model, characterized in that, The system includes: Receiving module: used to receive user prompts based on a large language model, wherein the prompts are related to the application method of oligonucleotide sequences; The generation module is used to determine the application method of the oligonucleotide in the prompt word through a large language model. Different application methods have a first mapping relationship with different scene agents. Each scene agent encapsulates the processing logic of the corresponding application method. The processing logic is selected from at least one of the following: generating one or more sets of candidate primers for one or more gene target regions, generating one or more sets of candidate primers or capture probes for one or more target regions, generating multiple sets of candidate primers or capture probes for multiple genes, and generating one or more sets of candidate primers or capture probes carrying target mutations for one or more target sites. The processing logic described therein includes multiple different types of databases and publicly available online information to be retrieved; Retrieval module: used to generate a retrieval strategy through a scene agent that has a first mapping relationship with the application method, wherein the retrieval strategy includes keywords, databases and publicly available network information; The database and publicly available online information are retrieved and processed based on the search keywords to generate nucleic acid sequence information.
9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method of claims 1 to 7.
10. A readable storage medium, characterized in that, The readable storage medium stores computer-readable instructions for causing a computer to perform the method of any one of claims 1 to 7.
