Multiplexed pcr-based biological age detection kits, prediction methods, and systems

Abstract

The application discloses a biological age detection kit, a prediction method and a system based on multiplex PCR, belongs to the technical field of bioinformatics, and the kit comprises any one of the following primer pools or a combination thereof: a first primer pool: Seq ID No. 1-6; a second primer pool: Seq ID No. 7-12; a third primer pool: Seq ID No. 13-32; a fourth primer pool: Seq ID No. 33-66; and a fifth primer pool: Seq ID No. 67-86. The methylation rates of corresponding gene sequences are obtained through multiplex PCR detection of the primer pools and primer combinations thereof, and biological age is detected or predicted according to the methylation rates; the detection cost is reduced, and the detection efficiency is improved; the primers in the primer pools are reasonably configured, and mutual interference is not caused.

Description

Technical Field

[0001] This invention relates to the field of bioinformatics technology, specifically to biological age detection kits, prediction methods, and systems based on multiplex PCR. Background Technology

[0002] Currently, biological age detection methods typically include the following approaches: methylation chips are costly and have fixed detection sites; whole-genome methylation sequencing data has low utilization and is expensive; telomere and metabolomics detection methods have poor stability and are difficult to standardize; and low-throughput technologies such as qPCR cannot meet the needs of combined detection of multiple biomarkers.

[0003] These methods struggle to balance cost, throughput, accuracy, flexibility, and clinical applicability, limiting the widespread adoption and application of bio-age testing in large-scale population health management. Multiplex PCR (mPCR), also known as composite PCR, involves adding multiple pairs of primers to the PCR reaction system to amplify multiple sites in the sample, significantly improving amplification efficiency and reducing amplification costs. Summary of the Invention

[0004] To address the aforementioned technical problems in the existing technology, this invention provides a biological age detection kit, prediction method, and system based on multiplex PCR. After multiplex PCR amplification, the methylation rate is detected by next-generation sequencing, and the biological age is predicted.

[0005] This invention discloses a biological age detection kit based on multiplex PCR, comprising any of the following primer pools or combinations thereof: First primer pool: Seq ID No. 1-6; Second primer pool: Seq ID No. 7-12; Third primer pool: Seq ID No. 13-32; Fourth primer pool: Seq ID No. 33-66; Fifth primer pool: Seq ID No. 67-86.

[0006] Preferably, the biological age detection kit also includes bisulfite conversion reagent, PCR premix, and NGS library construction reagent; Each primer pool also includes: Seq ID No. 87 and Seq ID No. 88.

[0007] Preferably, the biological age detection kit also includes purified magnetic beads, a negative control, and a positive control.

[0008] This invention also discloses a method for predicting biological age, characterized by comprising the following steps: Take the primer pool from the above biological age detection kit, mix the primers in the primer pool, and obtain at least one set of primer solutions; The DNA of the sample to be tested is converted with bisulfite conversion reagent to obtain DNA conversion solution; DNA transformation solution and primer solution were added to the corresponding reaction system for amplification to obtain the first amplification product; After purifying the first amplification product with magnetic beads, the purified product was subjected to library amplification to obtain the second amplification product. The second amplification product was purified using magnetic beads to obtain the second purified product; The second purified product was sequenced to obtain sequencing results; Based on the sequencing results, calculate the methylation rate of gene loci / detection indicators; The methylation rate is predicted using a predictive model to obtain the biological age of the sample to be tested.

[0009] Preferably, equal amounts of primers are taken and the primers from each primer pool are mixed separately to obtain multiple sets of primer solutions, with a final concentration of 1 μM. Add 2-20 ng of DNA transformation solution and 1 μL of primer solution to the corresponding reaction system for amplification. The reaction system includes: 10 μL of 2xPCR premix and 9 μL of DNA transformation solution.

[0010] Preferably, the library amplification system includes 25 μL of 2x PCR premix; 20 μL of the first purified product; and 5 μL of library amplification primer premix.

[0011] Preferred methods for constructing predictive models include: Create a filtered dataset; Based on the dataset, a set of methylation sites highly correlated with biological age were selected and used as detection indicators; Establish a training set; The method based on elastic network regression is used to train the training set and its detection indicators to obtain a prediction model.

[0012] Preferably, the detection index is selected from: ELOVL2-AS1, LINC01923, MAP3K5, MN1, STXBP5, TIAL1, ADM, ASPA, CCDC102B, GFPT2, GSE1, TRIM59-IFT80, TRPS1, USP46, ZFAND1, ZNF423, LDAH, LINC01122, LOC100505585, MIR29B2CHG, MKLN1-AS, MMRN2, NFIB, PCDH9, RUNX1T1, TLE1, CXXC5, EPHB3, FHL2, HPSE, LINC00900, LINC02198, NKIRAS2, SKAP1, and SP1; The primers for the detection indicators were selected from Seq ID No. 1-86.

[0013] Preferably, the sample to be tested is collected from: fat, blood, cerebellum, cerebral cortex, heart, kidney, liver, lung, muscle, skin or spleen; Methods for calculating aging rate: Obtain the actual age of the sample to be tested; The aging rate is calculated based on the quotient of biological age and actual age.

[0014] The present invention also provides a prediction system for implementing the above prediction method, comprising a data acquisition module and a prediction module. The acquisition module is used to collect the methylation rate of the detection index; The prediction module is used to predict the methylation rate using a prediction model to obtain the biological age of the sample to be tested.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: multiplex PCR detection is performed through primer pools and primer combinations to obtain the methylation rate of the corresponding gene sequence, and the biological age is detected or predicted based on the methylation rate; a small amount of data can achieve a higher (more than 100x) sequencing depth, reduce detection costs, and improve detection efficiency and accuracy; the primers in the primer pool are reasonably configured and will not cause mutual interference. Attached Figure Description

[0016] Figure 1 This is a flowchart of the biological age prediction method based on multiplex PCR of the present invention; Figure 2 This is a distribution map of organization types in the GEO dataset; Figure 3 It is a distribution density map of DNA methylation rate and methylation entropy in different tissues; Figure 4 These are the methylation rates and methylation entropies of different tissues at different age stages; Figure 5This is a diagnostic graph for the predictive model; Figure 6 This is a correlation graph between actual age and biological age; Figure 7 This is a graph showing a linear relationship between the methylation rate of the ELOVL2-AS1 site and age. Figure 8 It is a graph showing the aging rate of various tissues; Figure 9 This is a quality control image of a multiplex PCR methylation library; Figure 10 This is a logic block diagram of the prediction system of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The present invention will now be described in further detail with reference to the accompanying drawings: The first aspect of this invention provides a biological age detection kit based on multiplex PCR, as shown in Tables 1-5, comprising any of the following primer pools or combinations thereof: First primer pool (pool1): Seq ID No. 1-6; Second primer pool 2: Seq ID No. 7-12; Third primer pool (pool3): Seq ID No. 13-32; Fourth primer pool (pool4): Seq ID No. 33-66; Fifth primer pool (pool5): Seq ID No. 67-86.

[0019] As shown in Table 1, the first primer pool was used to detect the methylation rates of ELOVL2-AS1, LINC01923, and MAP3K5; as shown in Table 2, the second primer pool was used to detect the methylation rates of MN1, STXBP5, and TIAL1; as shown in Table 3, the third primer pool was used to detect the methylation rates of ADM, ASPA, CCDC102B, GFPT2, GSE1, TRIM59-IFT80, TRPS1, USP46, ZFAND1, and ZNF423; as shown in Table 4... 4. The fourth primer pool was used to detect the methylation rates of LDAH, LINC01122, LOC100505585, MIR29B2CHG, MKLN1-AS, MMRN2, NFIB, PCDH9, RUNX1T1, and TLE1. As shown in Table 5, the fifth primer pool was used to detect the methylation rates of CXXC5, EPHB3, FHL2, HPSE, LINC00900, LINC02198, NKIRAS2, SKAP1, and SP1. Each primer pool was used for a multiplex PCR amplification.

[0020] Multiplex PCR detection using primer pools and their combinations yields the methylation rate of corresponding gene sequences, and the biological age can be detected or predicted based on the methylation rate; this reduces detection costs and improves detection efficiency; primer pools are grouped according to the Tm value of each primer, amplicon length, primer dimer and cross-complementarity, sequence homology, methylation sequence characteristics, and experimental verification results; the primer configuration within the primer pool is reasonable and will not cause mutual interference.

[0021] Table 1

[0022] Table 2

[0023] Table 3

[0024] Table 4

[0025] Table 5

[0026] One approach is to introduce a partial adapter sequence at the 5' end during primer synthesis, allowing direct library primer amplification during the second round of library amplification without the need for an additional adapter ligation process.

[0027] The adapter sequence Seq ID No. 87 introduced by the forward primer sequence (odd rows in Tables 1-5) is as follows: 5'-ACACTCTTTCCCTACACGACGCTCTTCCGATCT-forward primer-3' The adapter sequence introduced by the reverse primer sequence (even-numbered rows in Tables 1-5) is Seq ID No. 88, and the specific sequence is as follows: 5'-GACTGGAGTTCAGACGTGTGCTCTTCCGATCT-reverse primer-3' Some cgID sites use the same primer pairs, such as cg14870509, cg01085196, cg08987848, cg16862911, cg22655206, and cg03516596. This increases the concentration of the primer pair in the primer pool and allows for the detection of the methylation rate at these sites.

[0028] The kit also includes bisulfite conversion reagent, PCR premix, NGS library construction reagent, and negative and positive controls. The PCR premix includes DNA polymerase, dNTPs, buffer, and Mg2+. 2+ For example, conventional PCR premixes can be used, which will not be elaborated upon in this invention. Library construction reagents include adapters, ligases, and purification magnetic beads.

[0029] A second aspect of this invention provides a method for predicting biological age, such as... Figure 1 As shown, it includes the following steps: Step S1: Take the primers from the primer pool and mix them to obtain the primer solution. The final concentration of the primer pool is 1 μM.

[0030] Step S2: The DNA of the sample to be tested is converted with bisulfite conversion reagent. The DNA conversion solution and primer solution are added to the corresponding reaction system for amplification to obtain the first amplification product. The amount of DNA added to each reaction system is 2-20 ng.

[0031] The test sample consists of lysed red blood cells, digested with protease, and genomic DNA extracted. The concentration and purity of the genomic DNA are then determined. The specific method for DNA extraction is existing technology and will not be described further in this invention. A bisulfite conversion reagent can convert unmethylated C to U, while methylated C remains unchanged.

[0032] The PCR reaction system includes: 10 μL of 2x PCR premix; 1 μL of primer pool; and 9 μL of DNA transformation buffer.

[0033] Step S3: After purifying the first amplification product with 1.3X magnetic beads, the first purified product is subjected to library amplification to obtain the second amplification product.

[0034] The library amplification system consisted of: 25 μL of 2x PCR premix; 20 μL of the first purified product; and 5 μL of library amplification primer premix. The library was constructed by introducing adapters through library amplification and PCR enrichment. The library amplification primer premix used an existing NGS library preparation kit, which will not be described in detail here. During library amplification, quality control images were collected. Figure 9 A specific library amplification quality control chart is shown.

[0035] Step S4: After library amplification, the second amplification product is purified by using 0.8X magnetic beads to obtain the second purified product.

[0036] The library concentration can be determined using a double-stranded DNA fluorescence quantitative assay kit (dsDNA HS Assay Kit). The distribution of nucleic acid fragments in the library was quality controlled using an Agilent 4200 bioanalyzer. The second purified product was then subjected to PE150 sequencing.

[0037] Step S5: Sequencing the second purified product to obtain sequencing results, including the methylation rate of gene loci.

[0038] After the data is processed, it undergoes quality control, comparison, and deduplication. The methylation rate β of each target CpG site is then calculated. The methylation rate reflects the methylation ratio, and its calculation method is existing technology, which will not be elaborated upon in this invention.

[0039] Step S6: Predict the methylation rate using a prediction model to obtain the biological age of the sample to be tested.

[0040] The method for constructing the prediction model includes the following steps: Step 301: Create a filtered dataset.

[0041] The selected datasets included methylation microarrays from public databases, WGBS data, and large biological age model data from local laboratories using WGBS. The primary public dataset was GSE223748 from the Gene Expression Omnibus (GEO), provided by Steve Horvath's team. This dataset contains 900 human samples covering 11 tissue types, including adipose tissue, blood, cerebellum, cortex, heart, kidney, liver, lung, muscle, skin, and spleen, with sample ages ranging from 26.2 to 101 years. Figure 2 This is a distribution diagram of GEO public data across various organizations based on the t-SNE (t-distributed Stochastic Neighbor Embedding) algorithm, where the horizontal axis represents the first dimension t-SNE1 and the vertical axis represents the second dimension t-SNE2; for example... Figure 3 and Figure 4The DNA methylation rates and methylation entropies of various organizations in the public data are similar. Figure 4 In the figures, the horizontal axis represents age group, and the vertical axis represents methylation rate and methylation entropy. Methylation entropy... Entropy The calculation formula is: ; in, MRi It is the first i methylation rate at each site N It is the total number of sites.

[0042] Raw methylation data were preprocessed using the ChAMP package (version 2.30.0), filtering low-quality probes and sites based on detection p-values ​​to retain high-quality data; missing values ​​were estimated using the k-nearest neighbor (KNN) imputation method; probes associated with single nucleotide polymorphisms (SNPs) and multi-site probes were removed. The methylation level at each site was quantified by a β value, calculated as β = M / (M+U), where M is the methylation signal intensity and U is the unmethylation signal intensity, with β values ​​ranging from 0 (completely unmethylated) to 1 (completely methylated).

[0043] Step 302: Based on the dataset, select a set of methylation sites that are highly correlated with biological age and use them as detection indicators.

[0044] A linear model was fitted to the β value of each CpG site using the lmFit function in the limma R package (version 3.56.2), with age as a continuous variable and tissue type as a covariate. The p-value was corrected using the Benjamin-Hochberg method, with selection criteria of a corrected p-value < 0.01 and an absolute change in methylation per year |ΔBeta / year| > 0.002. A set of CpG sites highly correlated with biological age was ultimately obtained as detection indicators. This site combination, after large-sample screening and model validation, demonstrated high sensitivity and specificity for biological age.

[0045] Specifically, it consists of 43 methylation sites. Multiplex PCR primers were designed based on these methylation sites. The primers were optimized and grouped according to different primer annealing temperatures, amplifying the entire target region in different single-tube reaction systems. The amplification efficiency was uniform, with low non-specific amplification and few primer dimers, making it suitable for low starting sample volumes.

[0046] Step 303: Establish training and test sets.

[0047] The training set was derived from methylation data of 173 blood samples from the GSE223748 dataset. The 173 blood samples from GSE223748 were randomly divided into a training set (approximately 70%) and a test set (approximately 30%).

[0048] Step 304: Based on the elastic network regression method, train the training set and its detection indicators to obtain the prediction model.

[0049] The ElasticNet algorithm was employed, and the model was trained using the glmnet package in R. ElasticNet combines the advantages of Ridge Regression and LASSO Regression, is suitable for high-dimensional data (i.e., fewer samples than features), effectively handles collinearity issues between sites, and automatically performs feature selection. However, the training algorithm is not limited to this.

[0050] Specifically, the model was trained using the `cv.glmnet` function from the `glmnet` package (version 4.1.8) in R. Ten-fold cross-validation was used for internal validation. The alpha value was set within the range of 0.1 to 0.9, and the optimal lambda value was selected through ten-fold cross-validation. The model with the lowest mean absolute error (MAE) on the validation set was used as the final model. The final model can be applied to blood samples to obtain the DNA methylation age (i.e., biological age) of each sample.

[0051] Step 305: Model validation and performance evaluation.

[0052] On the validation set, the predictive model achieved a mean age of effect (MAE) of 3.36 years, with a correlation coefficient greater than 0.93 between predicted and actual ages. The model's generalization ability was validated on independent external datasets (such as the 450K dataset containing disease samples), where the MAE remained low. Comparison of aging rates across different tissue types revealed that the aging rate of blood samples effectively reflects systemic aging.

[0053] To further characterize the heterogeneity of age-related methylation changes, methylation entropy analysis was performed, and sliding window analysis using the DEswan package was used to identify peak regions of age-related methylation changes. These analyses contribute to a comprehensive understanding of the dynamic characteristics of target sites.

[0054] The following diagnostic analysis was performed on the features of the prediction model: Residual analysis: such as Figure 5 The residual plot of the predicted values ​​and actual ages of the prediction model shows a random distribution, indicating that the model prediction results are reliable and there is no systematic bias.

[0055] Weight coefficient analysis: The model's weight coefficients approximate a normal distribution, with most coefficients being negative. This is highly consistent with the previously observed phenomenon that "overall methylation levels decrease with age," validating the biological rationale of the prediction model from an algorithmic perspective.

[0056] Key feature verification: such as Figure 7In the prediction model, CpG sites associated with the ELOVL2 gene were given high weights, and the methylation level of this gene showed a high correlation with age, confirming the representativeness of the features selected in the prediction model.

[0057] Internal validation: Validated on a test set containing 32 independent samples, the model's mean absolute error (MAE) between predicted biological age and actual age was 3.36 years. Figure 6 This demonstrates extremely high prediction accuracy.

[0058] Cross-organizational validation and application: The trained prediction model was applied to predict the biological age of other organs and tissues (such as fat, muscle, heart, liver, nervous system, etc.), and the aging rate (i.e., biological age / chronological age) of each tissue was calculated. Figure 8 The results showed that the aging rates of muscles, heart, and nervous system (cerebellum and cortex) are relatively stable throughout the lifespan. However, tissues such as the liver exhibit dynamic changes: the aging rate declines rapidly in early adulthood and then stabilizes in old age. This cross-tissue application not only validates the robustness of the predictive model but also reveals tissue-specific aging patterns.

[0059] External dataset validation: To further validate the model's generalization ability, it was applied to an independent external dataset containing disease samples detected using the Illumina 450K chip.

[0060] The prediction model maintains good predictive performance on these datasets and can effectively identify accelerated aging phenomena (i.e., aging rate > 1.05) in disease samples, demonstrating the application potential of this method and model in assessing health risks.

[0061] A third aspect of the present invention provides a prediction system for implementing the above-described prediction method, such as... Figure 10 As shown, it includes a data acquisition module 1 and a prediction module 2. The acquisition module 1 is used to acquire the methylation rate of gene loci; The prediction module 2 is used to predict the methylation rate through a prediction model to obtain the biological age of the sample to be tested.

[0062] This invention has the following characteristics: 1. Significantly reduced costs. Utilizing multiplex PCR targeted amplification, sequencing only the target sites results in smaller data volumes and lower sequencing costs, far lower than methylation microarrays and whole-genome methylation sequencing (WGBS), making it suitable for large-scale population screening.

[0063] 2. High throughput and high efficiency. A single tube can simultaneously amplify dozens of target sites, allowing for the parallel detection of a large number of samples at once. The experimental procedure is simplified, the cycle is short, and standardized and automated detection can be achieved.

[0064] 3. High accuracy and good stability. Targeted amplification combined with high-depth sequencing provides accurate quantification of methylation levels with good reproducibility; multi-site joint modeling is employed, and the prediction accuracy is close to the gold standard, significantly superior to methods such as telomere, qPCR, and phenotypic models.

[0065] 4. Flexible and scalable site selection. Unlike chips with fixed sites, this invention can flexibly update the site combination based on new aging biomarkers and characteristics of the Chinese population, offering strong scalability and facilitating continuous model iteration.

[0066] 5. High sample adaptability. It can maintain high efficiency in amplification of trace and degraded samples (such as oral swabs, trace amounts of blood, and old samples), making it applicable to a wider range of scenarios and more practical for clinical and health management.

[0067] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A biological age detection kit based on multiplex PCR, characterized in that, Includes any of the following primer pools or combinations thereof: First primer pool: Seq ID No. 1-6; Second primer pool: Seq ID No. 7-12; Third primer pool: Seq ID No. 13-32; Fourth primer pool: Seq ID No. 33-66; Fifth primer pool: Seq ID No. 67-86.

2. The biological age detection kit according to claim 1, characterized in that, It also includes PCR premix, bisulfite conversion reagent and NGS library construction reagent; Each primer pool also includes: Seq ID No. 87 and Seq ID No.

88.

3. The biological age detection kit according to claim 2, characterized in that, It also includes purified magnetic beads, negative controls, and positive controls.

4. A biological age prediction method, characterized in that, Includes the following steps: Take the primer pool of the biological age detection kit as described in any one of claims 1-3, mix the primers in the primer pool, and obtain at least one set of primer solutions; The DNA of the sample to be tested is converted with bisulfite conversion reagent to obtain DNA conversion solution; DNA transformation solution and primer solution were added to the corresponding reaction system for amplification to obtain the first amplification product; After purifying the first amplification product with magnetic beads, the purified product was subjected to library amplification to obtain the second amplification product. The second amplification product was purified using magnetic beads to obtain the second purified product; The second purified product was sequenced to obtain sequencing results; Based on the sequencing results, calculate the methylation rate of gene loci / detection indicators; The methylation rate is predicted using a predictive model to obtain the biological age of the sample to be tested.

5. The biological age prediction method according to claim 4, characterized in that, Take equal amounts of primers and mix the primers from each primer pool separately to obtain multiple sets of primer solutions. The final concentration of the primer solutions is 1 μM. Add 2-20 ng of DNA transformation solution and 1 μL of primer solution to the corresponding reaction system for amplification. The reaction system includes: 10 μL of 2x PCR premix. 9ul of DNA transformation solution.

6. The biological age prediction method according to claim 4, characterized in that, The library amplification system includes 25 μL of 2x PCR premix; 20 μL of the first purified product; and 5 μL of library amplification primer premix.

7. The biological age prediction method according to claim 4, characterized in that, Methods for building predictive models include: Create a filtered dataset; Based on the dataset, a set of methylation sites highly correlated with biological age were selected and used as detection indicators; Establish a training set; The method based on elastic network regression is used to train the training set and its detection indicators to obtain a prediction model.

8. The biological age prediction method according to claim 7, characterized in that, The detection indicators are selected from: ELOVL2-AS1, LINC01923, MAP3K5, MN1, STXBP5, TIAL1, ADM, ASPA, CCDC102B, GFPT2, GSE1, TRIM59-IFT80, TRPS1, USP46, ZFAND1, ZNF423, LDAH, LINC01122, LOC100505585, MIR29B2CHG, MKLN1-AS, MMRN2, NFIB, PCDH9, RUNX1T1, TLE1, CXXC5, EPHB3, FHL2, HPSE, LINC00900, LINC02198, NKIRAS2, SKAP1, and SP1; The primers for the detection indicators were selected from Seq ID No. 1-86.

9. The biological age prediction method according to claim 7, characterized in that, The samples to be tested were collected from: fat, blood, cerebellum, cerebral cortex, heart, kidney, liver, lung, muscle, skin, or spleen; Methods for calculating aging rate: Obtain the actual age of the sample to be tested; The aging rate is calculated based on the quotient of biological age and actual age.

10. A prediction system for implementing the biological age prediction method as described in any one of claims 4-9, characterized in that, Includes a data acquisition module and a prediction module. The acquisition module is used to collect the methylation rate of the detection index; The prediction module is used to predict the methylation rate using a prediction model to obtain the biological age of the sample to be tested.

Images