System, method and device for detecting pathogens of infectious diseases
By integrating metagenome, phage group and host transcriptome data, a support vector machine detection model was constructed, which solved the problem of high false positive and false negative rates of infectious disease pathogen detection in the prior art, and achieved more accurate pathogen detection and Better diagnostic strategies.
Patent Information
- Application Number
- CN202411481848.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-23
- Publication Date
- 2025-02-14
AI Technical Summary
The prior art has high false positive and false negative rates in pathogen detection of infectious diseases, and it is difficult to accurately judge the existence of multiple pathogens, which increases the complexity and difficulty of diagnosis.
Integrate metagenomics, phageomics and transcriptomics to build a pathogen detection system for infectious diseases. By integrating metagenomics, phage groups and host transcriptomic data, a support vector machine detection model is established to achieve more accurate pathogen detection.
Improve the accuracy of infectious disease pathogen detection, enables more accurate assessment of whether samples are infected with bacteria, fungi, viruses or parasites, reduce misdiagnosis and misdiagnosis, and provide better diagnostic and therapeutic strategies.
Smart Images
Figure CN119446272A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of molecular diagnosis, and relates to a pathogen detection system, method and device for infectious diseases. Background Art
[0002] Bacteria, viruses, fungi and parasites are the main pathogens that cause infectious diseases. Different types of pathogens cause different ways and characteristics of human immune response. When infected with multiple types of pathogens at the same time, different pathogens may interact with each other, leading to aggravation of the disease and increased difficulty in treatment.
[0003] In recent years, microbiome research has developed rapidly, especially the application of metagenomic sequencing technology in pathogen detection, which has provided great help for disease prevention, diagnosis, treatment, and monitoring and tracking of infectious diseases. The next-generation sequencing technology can simultaneously analyze and identify multiple microbial communities in a sample through a high-throughput sequencing platform.
[0004] Bacteriophages are viruses that infect bacteria and have strong selective specificity for host bacteria. Analysis of bacteriophages in biological samples can reflect the pathogens present in the sample at the species level. Therefore, phage analysis of metagenomic data can further determine whether the patient is infected with a specific bacterial pathogen and avoid misdiagnosis and missed diagnosis.
[0005] However, due to the limitations of existing technologies, there are still certain false positive and false negative rates in pathogen detection results. In addition, the lesion sites of the human body are not all sterile environments, such as the lungs, where there are colonized bacteria. This makes the detection and identification of pathogens more complicated and increases the difficulty of accurate diagnosis. It is difficult for a single detection method to accurately determine the type of pathogen of an infectious disease. Summary of the invention
[0006] Based on the above technical problems, the present invention innovatively integrates metagenomics, phageomics and transcriptomics to construct a pathogen detection system for infectious diseases. The integration of metagenomics, phageomics and host transcriptome data is expected to provide an accurate map of infectious diseases and a deep understanding of the molecular mechanisms in the infection process, which can provide more accurate diagnostic tools and treatment strategies, thereby better preventing and controlling the spread of infectious diseases.
[0007] In a first aspect, the pathogen detection system for infectious diseases of the present invention comprises: The first training data acquisition module is used to analyze the metagenomic data of the sample, screen potential pathogenic microorganisms, and obtain potential pathogenic microorganism training data; The second training data acquisition module is used to analyze the sample macrotranscriptome data, screen immune-related genes, analyze the host transcriptome information, and obtain immune-related gene training data; the macrotranscriptome data contains not only pathogenic microorganism transcriptome information, but also host transcriptome information. Through these host transcriptome data, immune-related genes can be screened; A detection model building module, used to establish a detection model based on a support vector machine, and train the detection model using the potential pathogenic microorganism training data and the immune-related gene training data as training data; A module for acquiring data to be tested, used to obtain macrotranscriptome data of samples to be tested; The data analysis module to be tested is used to analyze the macrotranscriptome data of the sample to be tested, and use the host transcriptome data therein to screen immune-related genes to obtain immune-related gene data to be tested; The pathogen detection module is used to input the immune-related gene data to be detected into the trained detection model to obtain the corresponding pathogen detection results.
[0008] Specifically, the first training data acquisition module includes: A primary screening unit, which removes DNA sequences of low quality (≤Q30), low complexity (the percentage of bases different from the next base <30%) and small fragments (<35 bp) in the original data; A human sequence screening unit, which screens out human sequences aligned to the Genome Research Consortium human build 38 (GRCh38) human reference genome; A comparison unit compares all remaining sequences to a database on the website of the National Center for Biotechnology Information (NCBI) to screen out pathogenic microorganisms contained in the sample.
[0009] The phage analysis module is used to compare all sequences except human sequences to the phage GPD database (Gut Phage Database), identify the phages contained in the sample, clarify the bacteria corresponding to the phages in the sample, correct the metagenome detection results, and determine the pathogenic microorganisms in the biological sample.
[0010] Specifically, the second training data acquisition module includes: The primary screening unit is used to filter low-quality reads and remove low-quality bases at the 3′ end; The alignment unit is used to align the short reads that have passed the initial screening to the GRCh38 human reference genome and screen the short reads that are uniquely aligned; The calculation unit is used to calculate the gene expression level, and the normalized value of the expression amount is represented by TPM; The output unit is used to retain protein-coding genes according to Gencode annotation and remove low-expression genes with TPM < 3 in more than 99% of samples.
[0011] Specifically, the detection model building module includes: A determination unit for determining the expression of potential pathogenic microorganisms and immune-related genes in biological samples of subjects known to have bacterial infectious diseases, fungal infectious diseases, viral infectious diseases, and parasitic infectious diseases; The training unit is used to construct a detection model for detecting pathogenic microorganisms of infectious diseases according to the measurement results of the measurement unit.
[0012] It should be noted that the types of samples include but are not limited to blood, cerebrospinal fluid, and alveolar lavage fluid. The pathogenic microorganisms include at least one of bacteria, fungi, viruses, and parasites.
[0013] In a second aspect, the present invention also provides a detection method for pathogens of infectious diseases for the purpose of non-disease diagnosis and treatment, the pathogen detection method comprising: Analyze the metagenomic data of the samples, screen potential pathogenic microorganisms, and obtain training data for potential pathogenic microorganisms; Analyze the sample macrotranscriptome data, screen immune-related genes through the host transcriptome information, and obtain immune-related gene training data; Establishing a detection model based on a support vector machine, and using the potential pathogenic microorganism training data and the immune-related gene training data as training data to train the detection model; Obtaining macrotranscriptome data of samples to be tested; Analyzing the macrotranscriptome data of the sample to be detected, screening immune-related genes through the host transcriptome information therein, and obtaining immune-related gene data to be detected; The immune-related gene data to be detected is input into the trained detection model to obtain corresponding pathogen detection results.
[0014] In a third aspect, the present invention further provides a computer storage medium, wherein the storage medium stores a plurality of computer instructions, and the plurality of computer instructions are used to enable a computer to execute the above method.
[0015] In a fourth aspect, the present invention further provides a device, which adopts the above-mentioned detection system or detection method.
[0016] Compared with the prior art, the present invention "system, method and device for detecting pathogens of infectious diseases" has the following beneficial effects: The present invention utilizes a series of complex immune responses and activated signal pathways triggered by the interaction between pathogens and host cells during pathogen infection, resulting in corresponding physiological responses in host cells, including reverse tracing of changes in gene expression levels. The present invention utilizes high-throughput sequencing technology, especially RNA sequencing technology, to make host transcriptome sequencing an effective means of revealing host cell responses to pathogen infection. The present invention's sequencing of the host transcriptome can clarify the expression levels of multiple genes in host cells, thereby revealing the pattern of gene expression changes in host cells after infection, identifying signal pathways and biological processes related to infection, and understanding the response differences of different cell types. Therefore, by studying infection-related genes and signal pathways in the host transcriptome, the patient's infection status can be determined, and even the type of pathogen infected by the patient can be known, thereby providing important evidence for the diagnosis of pathogens for infectious diseases, and helping clinical decision-making and the formulation of treatment strategies.
[0017] The present invention integrates the metagenome, metatranscriptome, pathogenic microorganism information, and host immunity-related information to effectively improve the accuracy of infectious disease pathogen detection.
[0018] The present invention adopts a method of comparing a phage database. Since phages have strong selection specificity for their hosts, the accuracy of pathogen detection results can be further improved based on the phage information in the sample.
[0019] The pathogen detection system provided by the present invention has reliable results and is easy to operate. After a given biological sample, it can accurately evaluate whether the sample is infected with bacteria, fungi, viruses, parasites, and whether there is a mixed infection. It can quickly trace the source and block the transmission chain outside the clinic. In the clinic, it can clarify the diagnosis, guide precise treatment, effectively avoid problems such as incorrect medication and excessive medication, and improve the clinical treatment effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a workflow diagram of the pathogen detection system.
[0021] Figure 2 This is the workflow diagram of the measurement unit. Metagenomic and / or metatranscriptomic data are input into the measurement unit to determine whether there are potential pathogenic microorganisms, including bacteria, fungi, viruses, and parasites, in the biological samples of the subjects. Finally, the diagnosis results of the pathogens infected by the subjects are output.
[0022] Figure 3 The pathogen detection system established by the present invention uses the ROC curve of the clinical sample macrotranscriptome data information to predict the patient's infection status.
[0023] Figure 4 It is the improved wet experimental process of the present invention, which uses the formaldehyde cross-linking method to remove the host and the non-host to perform mNGS detection with two wet experimental methods, the host ratio in the original data. DETAILED DESCRIPTION
[0024] The technical solution of the present invention is clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0025] Embodiment 1: The present application provides a pathogen detection system for infectious diseases, such as Figure 1 As shown, the pathogen detection system comprises: The first training data acquisition module is used to analyze the metagenomic data of the sample, screen potential pathogenic microorganisms, and obtain potential pathogenic microorganism training data; The second training data acquisition module is used to analyze the sample macrotranscriptome data, use the host transcriptome information therein to screen immune-related genes, and obtain immune-related gene training data; A detection model building module, wherein the model building module uses the potential pathogenic microorganism screening results and the immune-related gene expression as model inputs to establish a detection model based on a support vector machine, and uses the potential pathogenic microorganism training data and the immune-related gene training data as training data to train the detection model; A module for acquiring data to be tested, used to obtain macrotranscriptome data of samples to be tested; The data analysis module to be tested is used to analyze the macrotranscriptome data of the sample to be tested, screen immune-related genes using the host transcriptome information, and obtain immune-related gene data to be tested; The pathogen detection module is used to input the immune-related gene data to be detected into the trained detection model to obtain the corresponding pathogen detection results.
[0026] Specifically, when training the detection model, the immune-related gene training data is input into the detection model, and its output is compared with the corresponding potential pathogenic microorganism training data. According to the comparison results, the back propagation algorithm is used to adjust the parameters in the detection model to achieve the purpose of training. After the training is completed, the trained detection model can also be verified.
[0027] Specifically, the first training data acquisition module includes: A primary screening unit, which removes DNA sequences of low quality (≤Q30), low complexity (the percentage of bases different from the next base <30%) and small fragments (<35 bp) in the original data; A human sequence screening unit, which screens out human sequences aligned to the Genome Research Consortium human build 38 (GRCh38) human reference genome; A comparison unit compares all remaining sequences to a database on the website of the National Center for Biotechnology Information (NCBI) to screen out pathogenic microorganisms contained in the sample.
[0028] The phage analysis module compares all sequences except human sequences to the Gu Phage Database (GPD) to identify the phages contained in the sample, clarify the bacteria corresponding to the phages in the sample, correct the metagenomic detection results, and determine the pathogenic microorganisms in the biological sample.
[0029] Specifically, the second training data acquisition module includes: A primary screening unit, wherein the primary screening unit filters low-quality reads and removes bases with low quality (≤Q30) at the 3′ end; An alignment unit, which aligns the reads that have passed the initial screening to the GRCh38 human reference genome and selects the reads that are uniquely aligned; A calculation unit, wherein the calculation unit calculates the gene expression level, and the normalized value of the expression amount is represented by TPM; An output unit, wherein the output unit retains protein-coding genes according to Gencode annotation and removes low-expression genes with TPM<3 in more than 99% of samples.
[0030] Specifically, the detection model building module includes: The measuring unit is used to measure the expression of potential pathogenic microorganisms and immune-related genes in a sample of a subject known to have a bacterial infection, a fungal infection, a viral infection, or a parasitic infection, such as Figure 2 As shown; A training unit is used to construct a detection model for infectious disease pathogen detection according to the measurement results of the measurement unit.
[0031] Embodiment 2: The present invention also provides a method for detecting pathogens of infectious diseases for the purpose of non-disease diagnosis and treatment, the method comprising the following steps: Analyze the metagenomic data of the samples, screen potential pathogenic microorganisms, and obtain training data for potential pathogenic microorganisms; Analyze the sample macrotranscriptome data, use the host transcriptome information to screen immune-related genes, and obtain immune-related gene training data; Establishing a detection model based on a support vector machine, and using the potential pathogenic microorganism training data and the immune-related gene training data as training data to train the detection model; Obtaining macrotranscriptome data of samples to be tested; Analyzing the macrotranscriptome data of the sample to be tested, analyzing the host transcriptome data therein, screening immune-related genes, and obtaining immune-related gene data to be tested; The immune-related gene data to be detected is input into the trained detection model to obtain corresponding pathogen detection results.
[0032] Embodiment 3: An embodiment of the present invention also provides a computer storage medium and a device, wherein the storage medium stores a plurality of computer instructions, and the plurality of computer instructions are used to enable a computer to execute the method of embodiment 2; the device adopts the detection system of embodiment 1 or the detection method of embodiment 2.
[0033] Embodiment 4: The embodiment of the present invention also verified the performance of pathogen detection of the system established in Example 1. A total of 1,176 clinical sample macrotranscriptome data were included. After excluding 21 unqualified data, a total of 1,155 clinical sample macrotranscriptome data were analyzed, including 355 blood samples, 390 bronchoalveolar lavage fluid (BALF) samples, and 410 cerebrospinal fluid (CSF) samples. The system established by the present invention uses blood sample, BALF sample and CSF sample macrotranscriptome data to predict the patient's viral infection status, and its AUC reached 0.836, 0.906 and 0.921 ( Figure 3 ), which is higher than the model constructed by Bodkin et al. in 2022 for predicting viral infection using host gene expression characteristics (AUCs were 0.82, 0.89, and 0.91, respectively). This shows that the model constructed by the present invention has a high accuracy in predicting the infection status of patients using clinical sample macrotranscriptome data.
[0034] Embodiment 5: The present invention also improves the previous wet experiment method for the system established in Example 1. A total of 41 clinical BALF samples were included, and the free nucleic acid extraction method was improved by differential centrifugation, and the method of cross-linking chromatin by formaldehyde was used to reduce the host nucleic acid in the extract. Through this improved method, the proportion of pathogenic microorganism nucleic acid in the extract was increased (the host proportion before and after the improvement was 98% and 84%, respectively) ( Figure 4 ). Ultimately, the system established in Example 1 can provide a basis for accurate detection of pathogens.
[0035] This example also verifies the detection of bacteria and fungi using other samples.
[0036] The embodiments described above are only some embodiments of the present invention, not all embodiments. The detailed description of the embodiments of the present invention is not intended to limit the scope of the invention claimed for protection, but only represents selected embodiments of the present invention. All other embodiments obtained without creative work and related deductions and substitutions made by ordinary technicians in the field under the conditions of the concept of the present invention belong to the scope of protection of the present invention.
Claims
1. A pathogen detection system for infectious diseases, characterized in that: The pathogen detection system comprises: The first training data acquisition module is used to analyze the metagenomic data of the sample, screen potential pathogenic microorganisms, and obtain potential pathogenic microorganism training data; The second training data acquisition module is used to analyze the sample macrotranscriptome data, screen immune-related genes, and obtain immune-related gene training data; A detection model building module, used to establish a detection model based on a support vector machine, and train the detection model using the potential pathogenic microorganism training data and the immune-related gene training data as training data; A module for acquiring data to be tested, used to obtain macrotranscriptome data of samples to be tested; The data analysis module to be tested is used to analyze the macrotranscriptome data of the sample to be tested, and use the host transcriptome data therein to screen immune-related genes to obtain immune-related gene data to be tested; The pathogen detection module is used to input the immune-related gene data to be detected into the trained detection model to obtain the corresponding pathogen detection results.
2. The infectious disease pathogen detection system according to claim 1, characterized in that: The first training data acquisition module includes: The primary screening unit is used to remove low-quality, low-complexity and small-fragment DNA sequences from the original data; Human sequence screening unit, used to screen out human sequences aligned to the GRCh38 human reference genome; The alignment unit is used to align all remaining sequences to the database of the NCBI website to screen out pathogenic microorganisms contained in the sample.
3. The infectious disease pathogen detection system according to claim 1, characterized in that: The first training data acquisition module also includes: The phage analysis module is used to align all sequences without human sequences to the phage GPD database, identify the phages contained in the sample, clarify the bacteria corresponding to the phages in the sample, correct the metagenomic detection results, and determine the pathogenic microorganisms in the biological sample.
4. The infectious disease pathogen detection system according to claim 1, characterized in that: The second training data acquisition module includes: The primary screening unit is used to filter low-quality reads and remove low-quality bases at the 3′ end; The alignment unit is used to align the short reads that have passed the initial screening to the GRCh38 human reference genome and screen the short reads that are uniquely aligned; The calculation unit is used to calculate the gene expression level, and the normalized value of the expression amount is represented by TPM; The output unit is used to retain protein-coding genes according to Gencode annotation and remove low-expression genes with TPM < 3 in more than 99% of samples.
5. The infectious disease pathogen detection system according to claim 1, characterized in that: The detection model building module includes: A determination unit for determining the expression of potential pathogenic microorganisms and immune-related genes in biological samples of subjects known to have bacterial infectious diseases, fungal infectious diseases, viral infectious diseases, and parasitic infectious diseases; The training unit is used to construct a detection model for detecting pathogenic microorganisms of infectious diseases according to the measurement results of the measurement unit.
6. The infectious disease pathogen detection system according to claim 1, characterized in that: The types of samples include but are not limited to blood, cerebrospinal fluid, and bronchoalveolar lavage fluid.
7. The infectious disease pathogen detection system according to claim 1, characterized in that: The pathogenic microorganisms include at least one of bacteria, fungi, viruses and parasites.
8. A method for detecting pathogens of infectious diseases for the purpose of non-disease diagnosis and treatment, characterized in that: The pathogen detection method comprises: Analyze the metagenomic data of the samples, screen potential pathogenic microorganisms, and obtain training data for potential pathogenic microorganisms; Analyze the sample macrotranscriptome data, analyze the host transcriptome information, screen immune-related genes, and obtain immune-related gene training data; Establishing a detection model based on a support vector machine, and using the potential pathogenic microorganism training data and the immune-related gene training data as training data to train the detection model; Obtaining macrotranscriptome data of samples to be tested; Analyzing the macrotranscriptome data of the sample to be tested, analyzing the host transcriptome information in the macrotranscriptome, screening immune-related genes, and obtaining immune-related gene data to be tested; The immune-related gene data to be detected is input into the trained detection model to obtain corresponding pathogen detection results.
9. A computer storage medium, characterized in that The storage medium stores a plurality of computer instructions, and the plurality of computer instructions are used to enable a computer to execute the method of claim 8.
10. A device, characterized in that: The device adopts the detection system according to any one of claims 1 to 7 or the detection method according to claim 8.