Ai-assisted primer design system and ai-assisted primer design method

TWI946328BActive Publication Date: 2026-11-11NATIONAL DEFENSIVE MEDICAL CENTER
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
TW115114698
Authority / Receiving Office
TW · TW
Patent Type
Patents
Current Assignee / Owner
Filing Date
2026-04-15
Publication Date
2026-11-11
Estimated Expiration
2046-04-14

Smart Images

  • Figure TWG2TB001921139_001
    Figure TWG2TB001921139_001
  • Figure TWG2TB001921139_002
    Figure TWG2TB001921139_002
  • Figure TWG2TB001921139_003
    Figure TWG2TB001921139_003
Patent Text Reader

Abstract

This disclosure presents an AI-assisted primer design system, comprising: an input module configured to receive the chromosomal coordinates of a target gene; a data acquisition module configured to extract nucleotide sequences from a reference genome based on the chromosomal coordinates; a primer generation module configured to generate multiple candidate primer pairs; a feature generation module configured to generate feature vectors for each candidate primer pair; a machine learning module configured to input the feature vectors into a machine learning model to generate machine learning prediction scores; a tile-based overlapping module configured to establish multiple amplification segments; a primer selection and optimization module configured to select primer combinations based on the machine learning prediction scores and amplification segments of each candidate primer pair; an output module configured to output primer combinations as primer sets; and a processor connected to the aforementioned modules and configured to control the operation of the aforementioned modules. This disclosure also discloses an AI-assisted primer design method.
Need to check novelty before this filing date? Find Prior Art

Claims

1. An AI-assisted primer design system, comprising: an input module configured to receive chromosomal coordinates of a target gene; a data extraction module configured to extract a corresponding nucleotide sequence from a reference genome based on the chromosomal coordinates; a primer generation module configured to generate multiple candidate primer pairs based on the nucleotide sequence; a feature generation module configured to generate a feature vector for each candidate primer pair; a machine learning module configured to input the feature vector of each candidate primer pair into a machine learning model to generate a machine learning prediction score for each candidate primer pair; a tile-style overlapping coverage module configured to establish a plurality of overlapping amplification segments covering the target gene; and a primer selection and optimization module configured to select a primer combination from the multiple candidate primer pairs based on the machine learning prediction score of each candidate primer pair and the plurality of amplification segments. An output module configured to output the primer combination as a primer set; and a processor connected to the input module, the data acquisition module, the primer generation module, the feature generation module, the machine learning module, the tile-based overlapping coverage module, the primer selection and optimization module, and the output module, configured to control the operation of the input module, the data acquisition module, the primer generation module, the feature generation module, the machine learning module, the tile-based overlapping coverage module, the primer selection and optimization module, and the output module.

2. The AI-assisted primer design system as described in claim 1, wherein the chromosome coordinates include a chromosome number and a coordinate range.

3. The AI-assisted primer design system as described in claim 1, wherein the primer generation module is configured to generate the plurality of candidate primer pairs based on an primer length and / or a target amplification segment length.

4. The AI-assisted primer design system as described in claim 1, wherein the feature vector includes a thermodynamic feature and / or a structural feature.

5. The AI-assisted primer design system as described in claim 1, wherein the feature vector comprises one or more of the group consisting of GC content, ΔTm, ΔGC, 3' GC clamp, consecutive identical nucleotide lengths, hairpin structure energy, self-dimerization energy, and heterodimerization energy.

6. The AI-assisted primer design system as described in claim 1, wherein the machine learning model is trained on success and failure data from a next-generation ordered augmented library.

7. The AI-assisted primer design system as described in claim 6, wherein the success and failure data of the next-generation sequencing amplification library includes multiplex PCR experimental results.

8. The AI-assisted primer design system as described in claim 1, wherein the machine learning model is selected from a group consisting of XGBoost, LightGBM, and Random Forest.

9. The AI-assisted primer design system as described in claim 1, wherein the machine learning model is built based on the feature vectors of each candidate primer pair and a corresponding success or failure flag.

10. The AI-assisted primer design system as described in claim 1, wherein the plurality of amplification segments are continuous and uniform amplification segments.

11. The AI-assisted primer design system as described in claim 1, wherein the tile-type overlapping overlay module is further configured to establish the plurality of amplified segments according to an overlap length.

12. The AI-assisted primer design system as described in claim 1, wherein adjacent plurality of amplification segments have a fixed overlap length with each other.

13. The AI-assisted primer design system as described in claim 1, wherein the primer selection and optimization module is further configured to select the primer combination from the plurality of candidate primer pairs based on a primer compatibility parameter.

14. The AI-assisted primer design system as described in claim 13, wherein the primer compatibility parameter is selected from the group consisting of ΔTm, ΔGC, and cross-interference energy.

15. The AI-assisted primer design system as described in claim 1, wherein the primer selection and optimization module is further configured to select the primer combination from the plurality of candidate primer pairs based on the coverage integrity of the target gene by the plurality of amplification segments.

16. The AI-assisted primer design system as described in claim 1, wherein the primer selection and optimization module is further configured to minimize the ΔTm difference and ΔGC difference in the primer combination.

17. An AI-assisted primer design method, comprising: receiving chromosomal coordinates of a target gene using an input module; extracting a corresponding nucleotide sequence from a reference genome using a data extraction module based on the chromosomal coordinates; generating multiple candidate primer pairs using a primer generation module based on the nucleotide sequence; generating a feature vector for each candidate primer pair using a feature generation module; inputting the feature vector of each candidate primer pair into a machine learning model using a machine learning module to generate a machine learning prediction score for each candidate primer pair; establishing a plurality of overlapping amplification segments covering the target gene using a tile-based overlapping module; selecting a primer combination from the multiple candidate primer pairs using a primer selection and optimization module based on the machine learning prediction scores of each candidate primer pair and the plurality of amplification segments; and outputting the primer combination as a primer set using an output module.

18. The AI-assisted primer design method as described in claim 17, wherein training the machine learning model comprises: collecting data from a first-generation ordinal amplification library, wherein the data contains amplification results of a plurality of primer pairs; a feature vector corresponding to each primer pair; training the machine learning model using a machine learning algorithm based on the amplification results and the feature vectors to establish a predictive model; and updating the machine learning model using newly accumulated second-generation ordinal amplification data.

19. The AI-assisted primer design method as described in claim 18, wherein the amplification result includes an amplification success marker and an amplification failure marker.

20. The AI-assisted induction design method as described in claim 18, wherein the machine learning algorithm is a supervised machine learning algorithm.

Citation Information

Patent Citations

  • Bird identification universal primer design method based on machine learning optimization and application

    CN121709031A

  • System for finding design region of nucleic acid detection and method thereof

    TW202601670A