MNGS pathogen absolute quantitative detection method based on internal reference conversion method
By using internal reference conversion algorithm and mixed pathogen internal standards in mNGS technology, the problem of difficult to identify responsible pathogens and accurate quantities in the prior art is solved, and the precise quantification of pathogenic microorganisms is achieved, and the accuracy and reliability of quantification are improved.
Patent Information
- Application Number
- CN202411979819.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-30
- Publication Date
- 2025-04-11
AI Technical Summary
The existing mNGS technology is difficult to accurately identify responsible pathogens, especially in open-site infections, and it is difficult to directly judge the content of pathogens in the sample. The existing absolute quantitative technology is complex in operation and high in cost, and the relative quantitative technology cannot directly reflect the pathogen content in the sample.
The absolute quantitative detection method of mNGS pathogen based on the internal reference conversion algorithm is used. By adding mixed internal standards of pathogens to the sample, and homogenizing and de-hosting using the internal reference bacteria Mix, the pathogen copy concentration is calculated and corrected to ensure the accuracy and reliability of the quantitative value.
The precise quantification of pathogenic microorganisms in the sample is achieved, there is no order of magnitude difference between the quantitative value and the theoretical value, and the correlation coefficient R2 of the mNGS quantitative value and the digital PCR quantitative value exceeds 0.9, improving the accuracy and reliability of the quantification.
Smart Images
Figure CN119799928A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of molecular diagnosis technology, and in particular to a metagenomic next generation sequencing (mNGS) pathogen absolute quantitative detection method based on an internal reference conversion algorithm. Background Art
[0002] Metagenome sequencing (mNGS) technology is an advanced molecular diagnostic method that can provide clinicians with rapid, objective and accurate etiological diagnosis results by directly measuring all genetic materials in samples from infected patients. Because it can detect pathogenic microorganisms that cannot be cultured by traditional culture methods, and identify responsible pathogens in mixed infections of multiple pathogens, mNGS technology is becoming increasingly prominent in the diagnosis of critical, difficult, rare and mixed infectious diseases.
[0003] However, there are some challenges in the practical application of mNGS technology. First, due to the existence of conditional pathogens, mNGS is currently unable to identify responsible pathogens in many microbial detections, especially for the diagnosis of open site infections such as the respiratory tract and urinary tract. Secondly, it is currently difficult for mNGS technology to directly determine the pathogen content in the original sample through the detected reads. The main reasons include: 1. The difficulty of nucleic acid extraction of different microorganisms is different: the cell structure and nucleic acid protection mechanism of different microorganisms lead to differences in nucleic acid extraction efficiency, which directly affects the accuracy of subsequent detection. 2. The human content varies greatly between different clinical samples: the human background content in the sample will significantly affect the sensitivity of mNGS for pathogen detection, especially in samples with high human content, the signal of pathogenic microorganisms may be submerged. 3. There are differences in genome size and genome specificity of different microorganisms: the differences in genome size and specific regions of different microorganisms lead to inconsistent numbers of reads generated during sequencing, thereby affecting the accuracy of quantification. 4. Existing absolute and relative quantitative technologies have certain limitations. Although absolute quantitative technologies such as digital PCR (dPCR) can directly obtain the absolute content of the detection target in the sample, they are complex to operate, costly, and cannot detect multiple pathogens at the same time. Although relative quantitative technologies such as qPCR are easy to operate, they can only obtain the ratio of the detection target to the sample's inherent target, and cannot directly reflect the pathogen content in the sample. Therefore, building an mNGS quantitative process is a necessary condition for accurate diagnosis of infectious diseases.
[0004] The accuracy, repeatability and consistency of mNGS quantitative methods with the gold standard method still need to be improved. Therefore, developing a new mNGS pathogen absolute quantitative detection method to improve the accuracy and reliability of quantification has important clinical significance for the accurate diagnosis of infectious diseases. Based on this demand, the present invention provides an mNGS pathogen absolute quantitative detection method based on an internal reference conversion algorithm, which can effectively overcome the limitations of the prior art and achieve accurate quantification of pathogenic microorganisms. Summary of the invention
[0005] The purpose of the present invention is to provide an mNGS pathogen absolute quantitative detection method based on an internal reference conversion algorithm, which does not increase the detection and analysis time, does not change the pathogen detection results of the original sample, can accurately quantify the pathogen content in the sample, and there is no order of magnitude difference between the quantitative value and the theoretical value, as well as the correlation coefficient R between the mNGS quantitative value and the digital PCR (ddPCR) quantitative value. 2 More than 0.9.
[0006] An mNGS pathogen absolute quantitative detection method based on an internal reference conversion algorithm comprises the following steps:
[0007] Step 1, homogenizing the sample solution to be tested, and performing or not performing a host removal treatment;
[0008] Step 2, adding a mixed pathogen internal standard to the solution obtained in step 1, and then performing mNGS sequencing detection;
[0009] Step 3, calculate the pathogen copy concentration by the following formula:
[0010]
[0011] Where C refers to the copy number, N refers to the number of species-specific reads, L is the genome length, and S is the proportion of genome-specific regions; in the subscripts, IC represents the internal reference, i represents the pathogen, and t represents the type of pathogen and internal reference; the average symbol represents the average value of the results obtained by comparing different versions of genome data;
[0012] The mixed pathogen internal standard contains bacteria, fungi and virus pathogens as internal standards.
[0013] The mixed pathogen internal standard contains pathogens of the types of G+ bacteria, G- bacteria, fungi, DNA viruses and RNA viruses as internal standards.
[0014] The method also includes the step of adding a correction term to the calculation result.
[0015] After adding the correction term, the pathogen copy concentration is calculated by the following formula:
[0016]
[0017] Where w and b are the parameters of the linear model, and the subscript corr refers to the correction value.
[0018] The mixed pathogen internal standard includes halobacterium, halophilic bacillus, Trichoderma citriodora, canine parvovirus, canine adenovirus, mammalian mumps virus type 5 and canine distemper virus. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 : Quantitative Method Evaluation Process
[0020] Figure 2 :Quantitative method to evaluate the DNA extraction amount of samples; the actual nucleic acid concentration of samples with an extraction amount of 1ng is lower than the Qubit detection limit (<0.1ng / μL)
[0021] Figure 3 : Quantitative method to evaluate the library yield of a sample
[0022] Figure 4 : Quantitative methods evaluate the HQ reads of samples
[0023] Figure 5 : Quantitative method to evaluate the comparison of pathogen reads pct of samples
[0024] Figure 6 : Quantitative method to evaluate the detection of external reference pathogens in samples RP20M
[0025] Figure 7 : Quantitative method to evaluate the detection of internal reference microorganisms in samples RP20M
[0026] Figure 8 : Quantitative method to evaluate the detection of internal reference molecules in samples RP20M
[0027] Fig. 9 :Original results of internal reference bacteria quantification
[0028] Fig.10 : Quantitative results of internal reference molecules
[0029] Fig.11 :Comparison of the results of internal reference bacteria quantification before and after correction of bacterial quantification
[0030] Fig.12 :Linear fitting results of the quantitative values of 8 pathogens and theoretical values under internal reference bacteria quantification
[0031] Fig.13 :Linear fitting results of the quantitative values of 8 pathogens and theoretical values under internal reference bacteria quantification DETAILED DESCRIPTION
[0032] The present invention provides an absolute quantitative detection method for common pathogens in mNGS based on an internal reference conversion algorithm, the method comprising the following steps: step one, pre-treating the above-mentioned experimental samples using a homogenization and host removal process, adding an internal reference bacteria Mix after the sample is host removed or before homogenization and crushing; step two, extracting sample nucleic acid; step three, library construction; step four: quantitatively calculating the pathogens detected in the sample.
[0033] The common pathogens described in step 1 include Gram-positive bacteria, Gram-negative bacteria, fungi, DNA viruses and RNA viruses.
[0034] The internal reference bacteria Mix in step 1 includes Halobacterium, Halobacterium, Trichoderma citrinoviride, canine parvovirus, canine adenovirus, mammalian mumps virus type 5, and canine distemper virus. The internal reference bacteria include bacteria, fungi, DNA viruses, and RNA viruses. Because mNGS can detect a wide variety of species, bacteria, fungi, and viruses, their genome sizes vary greatly. The above-mentioned internal reference bacteria can effectively solve the problem of large species differences in mNGS. This combination shows good detection stability.
[0035] In step 4, the pathogen copy concentration in the original sample is calculated based on the number of unique reads detected by the internal reference bacteria and the number of unique reads detected by the pathogen; the pathogen copy number calculation method is:
[0036]
[0037] *C: copy number, N: number of species-specific reads, L: genome length (bp), S: proportion of genome-specific regions (%); the subscripts mean: IC: internal reference, i: pathogen, t: pathogen / internal reference type. Since there are multiple reference genome versions for each pathogen, the average value can be calculated based on the results of multiple versions as the length and proportion of specific regions in the denominator. Here t is a label, which can represent fungi, bacteria or viruses; when the target pathogen to be detected is one of these types, the internal reference used also selects the corresponding type of data. For example, when detecting bacteria, the internal reference type included uses the read segment data of Halobacterium and Halobacterium.
[0038] Except for the internal reference which used the default copy number, the pathogen copy number was calculated according to formula (1).
[0039] In step 4, the following formula is used to correct the pathogen copy number:
[0040] De-hosting process and homogenization process:
[0041]
[0042] *w,b: parameters of the linear model, the subscript corr refers to the correction value.
[0043] The host removal process and the homogenization process will correct the pathogen copy number calculated from formula (1) according to formula (2). The linear models of different processes are as follows:
[0044] Bacteria correction model for host removal process: 0.4654*x+1.58
[0045] Homogenization process bacteria correction model: 0.5996*x+0.87
[0046] The x in the model represents log 10 The coefficients represent w and b respectively. Note that no correction is performed for tuberculosis and mycoplasma in the bacterial category. The pathogen correction of the magnetic bead process uses the original read number multiplied by 10, and no correction is performed for other processes.
[0047] In step 4, the following formula is used to calculate the concentration of the detected pathogen:
[0048]
[0049] The final pathogen concentration is calculated according to the obtained pathogen copy number and the sample extraction input volume according to formula (3), and the unit is copies / mL. The default extraction input volume of fresh tissue samples, feces, and paraffin tissue samples is 1 ml. If the input volume unit is not counted in mL, the default is 1 mL.
[0050] The testing and verification process in this patent includes the following parts:
[0051] 1. Internal reference system: The internal reference MIX of 7 non-pathogenic microorganisms was used to systematically evaluate the nucleic acid extraction of different microorganisms, including the difficulty of nucleic acid extraction of different microorganisms, genome size and specificity, and then a quantitative equation was obtained.
[0052] 2. Quantitative equation: The pathogen copy concentration in the original sample is calculated based on the number of unique reads detected by the internal reference bacteria, the number of unique reads detected by the pathogen, the amount of internal reference bacteria added, the genome size of the internal reference bacteria, the genome size of the pathogen, the proportion of specific intervals of the genome of the internal reference bacteria, and the proportion of specific intervals of the genome of the pathogen.
[0053] 3. Quantitative verification: Quantitative verification is performed by mixing pathogens with known copy numbers, including three sample types such as alveolar lavage fluid, blood and cerebrospinal fluid, as well as 15 common pathogens such as G+ bacteria, G- bacteria, fungi, DNA viruses and RNA viruses.
[0054] 4. Correction model: The final quantitative value obtained by using the general model correction is close to the expected value. No correction is performed for special bacteria such as tuberculosis and mycoplasma in the bacterial category.
[0055] 5. Quantitative accuracy verification: Use ddPCR to verify the accuracy of the quantitative method and ensure that the correlation coefficient R2 between the mNGS quantitative value and the ddPCR quantitative value exceeds 0.9.
[0056] Example 1: Quantitative Method Evaluation
[0057] Three negative bronchoalveolar lavage fluid samples were used to mix multiple pathogens in different gradients. The pathogen concentration gradients included gradient A (3000 CFU / mL), gradient B (10,000 CFU / mL) and gradient C (30,000 CFU / mL) (the virus copy number was 10 times that of the corresponding concentration gradient); the external reference pathogens included Mycobacterium tuberculosis, Streptococcus pneumoniae, Staphylococcus aureus, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Aspergillus fumigatus, Candida albicans, human cytomegalovirus and human adenovirus.
[0058] The extraction process and de-hosting operation process are briefly described as follows:
[0059] Take a certain amount of alveolar lavage fluid, add proteinase K, mix well and incubate in a metal bath for 10-20 minutes. Then add Host Depletion Solution, mix well, and incubate at room temperature for 15 minutes; centrifuge to remove the supernatant, add the prepared nuclease Mix, and incubate at 37°C for 5 minutes. Add proteinase K, incubate at 56°C for 10 minutes, add internal reference and PBS, mix well, and extract nucleic acids using the magnetic bead method.
[0060] 1. Experimental results and data analysis
[0061] 1.1. Extraction volume
[0062] like Figure 2 As shown, there was no significant difference in the extraction amount of the three bronchoalveolar lavage fluid samples among the three pathogen gradients, and the extraction amount of the homogenization process was higher than that of the host-free process.
[0063] 1.2. Library yield
[0064] like Figure 3 As shown in the figure, the library amount of all samples in this experiment meets the machine requirement (25ng). Due to the low extraction amount, most samples in the host-free process cannot reach the 10ng library construction input amount, so the total library amount is lower than the corresponding homogenization process samples.
[0065] 1.3.HQ reads
[0066] like Figure 4 As shown in the figure, the HQ reads of all samples in this experiment reached the quality control threshold and analysis requirements (>5Mreads). The data volume of the host removal process was slightly higher than that of the homogenization process, which may be due to different batches of machines. There was no significant difference in HQ reads between different external reference pathogen gradients and different samples.
[0067] 1.4. Pathogen reads pct
[0068] like Figure 5 As shown in the figure, the pathogen reads pct of the three samples in this experiment is between 0.01% and 7% overall, among which the host-free group is higher than the homogenous group (except for gradient B of sample 2). Among the three pathogen gradients, the pathogen reads pct shows a step-up trend, which is consistent with the experimental expectations.
[0069] 1.1. External reference pathogen detection Reads Per 20Million-reads (RP20M)
[0070] Figure 6 This is the detection information of all external reference pathogens in this experiment. The high detection value is marked in red, and the low detection value is marked in blue. Based on the detection information, three rules for mNGS detection of external reference pathogens can be summarized:
[0071] (1) When comparing the detection of the same pathogen in the same sample, there is a significant difference between the homogenization process and the host-free process. The detection level of most pathogens in the host-free process is higher than that in the homogenization process. This shows that the human background content will significantly affect the sensitivity of mNGS for pathogen detection.
[0072] (2) When comparing the same process of the same pathogen gradient (copy number), the detection levels of different pathogens vary greatly, which may be related to the pathogen type and its genome-specific interval (which can only be matched to the genome interval of the species). This also proves that mNGS detection reads or RP20M cannot truly reflect the pathogen content in the sample;
[0073] (3) When comparing the same pathogen in the same process, the detection between different gradients basically conforms to the rule that the higher the copy number, the higher the detection level. This shows that there is a correlation between the detection of RP20M by mNGS and the copy number of the pathogen, which is a necessary prerequisite for absolute quantification.
[0074] 1.1. Internal reference microorganism detection RP20M
[0075] Figure 7This is the detection result of all internal reference microorganisms (hereinafter referred to as internal reference) in this experiment. The added internal reference is a mixture of seven bacteria. The addition amounts of Halobacterium, Halobacterium, Trichoderma citriodora, canine parvovirus, canine adenovirus, mammalian mumps virus type 5, and canine distemper virus are 1000 cells, 1000 cells, 1000 cells, 3,000,000 copies, 15,000 copies, 20,000 copies, and 80,000 copies, respectively. The obtained reads are the normalized values of the actual detection reads of the five DNA internal references: For all detections, the host removal process is higher than the homogenization process. Since the detection reads of some internal references are too low, the quantitative results may be affected. With the increase of the external reference pathogen gradient, there is no obvious change in the detection of the internal reference. The detection ratio of the internal reference is relatively stable, and is positively correlated with the internal addition amount and the size of its genome-specific interval, which shows that the absolute quantification of each other using internal references has a high accuracy.
[0076] 1.1. Internal reference molecule detection RP20M
[0077] Figure 8 The RP20M results for the detection of three dsDNA internal reference molecules in this experiment were added in amounts of 0.4, 0.04, and 0.004 ng, respectively. The detection of internal reference molecules in the host-free process was also 1-2 orders of magnitude higher than that in the homogenization process, with small fluctuations in detection between samples and no significant correlation with pathogen gradients. In addition, the detection ratio of the three molecules X, Y, and Z was also basically close to the added mass ratio, indicating that the relative quantification between internal reference molecules also had a high degree of accuracy.
[0078] 1.8. Original Quantification of Internal Reference Bacteria
[0079] The theoretical calculation formula for the internal reference bacteria quantification is as follows. The first step is to calculate the pathogen content based on the number of unique reads detected by the internal reference, the number of unique reads detected by the pathogen, the amount of internal reference bacteria added, the size of the unique interval of the internal reference bacteria genome, and the size of the unique interval of the pathogen genome. The second step is to obtain the pathogen copy concentration in the original sample based on the pathogen content and the extraction input volume.
[0080] The quantitative calculation formula of internal reference bacteria is as follows:
[0081]
[0082] N ic : Number of unique reads detected by internal reference bacteria
[0083] N mic : Number of unique reads detected for pathogens
[0084] Cic :Amount of internal reference bacteria added (copies)
[0085] C mic :Pathogen content (copies)
[0086] L ic : Unique interval size of internal reference bacterial genome (bp)
[0087] L mic : Unique interval size of pathogen genome (bp)
[0088] C: Pathogen copy concentration in the original sample (copies / mL)
[0089] V: Extraction input volume (mL)
[0090] When calculating, select the internal reference type according to different pathogen types.
[0091] The results of the original quantitative analysis of the bacteria are as follows Fig. 9 As shown, the horizontal axis is the 17 external reference pathogenic microorganisms, and the vertical axis is the ratio of the quantitative value to the actual value, displayed in log10. The data in the red box comes from the host removal process, and the data in the blue box comes from the homogenization process. When the quantitative value is close to the actual value, the vertical axis value will be close to 0. When the vertical axis is negative, it means that the quantitative value is too low, otherwise it is too high. Fig. 9 It can be seen that the quantitative values of most G+ and G- bacteria are high, which is speculated to be because the actual addition amount of Halobacterium and Halobacterium as internal references may be lower than the theoretical addition amount (from the supplier). The quantitative values of Mycoplasma and Human Cytomegalovirus in the host removal process are low, which may be due to the high loss of these pathogens in the host removal process. The detection and quantitative results of Candida glabrata are seriously low, which may be because the actual addition amount of the bacteria is not accurate.
[0092] 1.9. Quantification of internal reference molecules
[0093] Fig.10 is the theoretical calculation formula for the internal reference molecule. The first step is based on the assumption that the microbial detection reads ratio is equal to the nucleic acid mass ratio. The unique molecular mass of the pathogen is calculated by the number of molecular internal reference detection reads, the number of unique reads detected by the pathogen, and the amount of internal reference molecule added. The second step is to calculate the number of pathogen copies added during library construction by the unique molecular mass of the pathogen and the mass of the unique interval of a single pathogen genome. The third step is to calculate the pathogen copy concentration in the original sample by the extraction volume per mL of the original sample, the library entry volume, and the number of pathogen copies added during library construction.
[0094]
[0095] Nic : Number of reads detected by molecular internal reference
[0096] M ic : Molecular internal reference addition amount (ng)
[0097] N mic : Number of unique reads detected for pathogens
[0098] M mic :Pathogen unique: Molecular mass (ng)
[0099] C mic : The number of pathogen copies added when building the library (copies)
[0100] GU: Quality of unique intervals in a single pathogen genome (ng / copy)
[0101] C: Pathogen copy concentration in the original sample (copies / mL)
[0102] DNA yield per mL: Extraction amount per mL of original sample (ng)
[0103] DNA input: the amount of DNA input for library construction (ng)
[0104] like Fig.10 As shown in the figure, the quantitative results of the internal reference molecule are generally low, and the box range is too long in the host removal process, indicating that the quantitative repeatability is poor. Therefore, the quantitative results using the internal reference molecule are inferior to those of the internal reference bacteria, and the internal reference bacteria are preferred as the quantitative method for mNGS.
[0105] 1.10. Quantitative calibration of internal reference bacteria
[0106] The accuracy and repeatability of quantification using internal reference bacteria are better than quantification using internal reference molecules, so internal reference bacteria are considered to be used for quantification. However, the amount of internal reference bacteria added may be lower than the theoretical value, resulting in a serious underestimation of the quantitative results of bacteria. Therefore, when using internal reference bacteria for quantification, the quantitative results of bacteria need to be corrected. The specific correction method is shown in the table below. By introducing the two quantitative parameters w and b, the quantitative values of all external reference bacteria are closest to the actual values.
[0107]
[0108]
[0109] The linear fitting formula is as follows: log10(C 理论浓度 )=w log 10(C 实际浓度 ) + b
[0110] like Fig.12As shown, when w = 0.4654, b = 1.58 in the host-free process, and w = 0.5996, b = 0.87 in the homogenization process, the quantitative ratio of all bacteria can be achieved within ±1 order of magnitude.
[0111] The results of the Pearson correlation analysis of mNGS for the quantification of some pathogens in Example 1 are as follows: Fig.12 As shown in the figure, points of different colors represent the quantitative results of external reference pathogens in different samples. The correlation coefficients between the calculated values and the theoretical values are all over 0.8, and the p value is significant.
[0112] Differences in pathogen quantification between Example 2 and ddPRC method
[0113] For the same positive clinical sample, mNGS and ddPCR methods were used for pathogen quantification, including bacteria (Baumannii, aeruginosa and tuberculosis), fungi (Aspergillus fumigatus) and viruses (CMV). Fig.13 , log-linear fitting of ddPCR quantification and mNGS quantification results, R 2 It can reach 0.9131. With the ddPCR result as a reference, there is no order of magnitude difference between the quantitative value and the theoretical value (the ratio is between 0.1-10 times). The quantitative accuracy of the mNGS of the present invention is high.
Claims
1. An mNGS pathogen absolute quantitative detection method based on an internal reference conversion algorithm, characterized in that: The steps include: Step 1, homogenizing the sample solution to be tested, and performing or not performing a host removal treatment; Step 2, adding a mixed pathogen internal standard to the solution obtained in step 1, and then performing mNGS sequencing detection; Step 3, calculate the pathogen copy concentration by the following formula: Where C refers to the copy number, N refers to the number of species-specific reads, L is the genome length, and S is the proportion of genome-specific regions; in the subscripts, IC represents the internal reference, i represents the pathogen, and t represents the type of pathogen and internal reference; the average symbol represents the average value of the results obtained by comparing different versions of genome data; The mixed pathogen internal standard contains bacteria, fungi and virus pathogens as internal standards.
2. The mNGS pathogen absolute quantitative detection method based on the internal reference conversion algorithm according to claim 1, characterized in that: The mixed pathogen internal standard contains pathogens of the types of G+ bacteria, G- bacteria, fungi, DNA viruses and RNA viruses as internal standards.
3. The mNGS pathogen absolute quantitative detection method based on the internal reference conversion algorithm according to claim 1, characterized in that: The method also includes the step of adding a correction term to the calculation result.
4. The mNGS pathogen absolute quantitative detection method based on the internal reference conversion algorithm according to claim 3, characterized in that: After adding the correction term, the pathogen copy concentration is calculated by the following formula: Where w and b are the parameters of the linear model, and the subscript corr refers to the correction value.
5. The mNGS pathogen absolute quantitative detection method based on the internal reference conversion algorithm according to claim 1, characterized in that: The mixed pathogen internal standard includes halobacterium, halophilic bacillus, Trichoderma citriodora, canine parvovirus, canine adenovirus, mammalian mumps virus type 5 and canine distemper virus.