Klebsiella pneumoniae KL genotype rapid identification method and system based on infrared spectrum
By combining infrared spectroscopy with a classification model, the complexity and low efficiency of Klebsiella pneumoniae KL genotype identification in existing technologies have been solved, enabling rapid and accurate Klebsiella pneumoniae KL genotype identification, especially the identification of the highly virulent KL1 type, which is applicable to clinical diagnosis and treatment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHUAN-MING (NINGBO) CHEM SCITECH CO LTD
- Filing Date
- 2026-02-28
- Publication Date
- 2026-07-14
AI Technical Summary
There is a lack of rapid and accurate methods for identifying Klebsiella pneumoniae KL genotypes, especially the highly virulent KL1 genotype. Existing methods are complex, time-consuming, costly, and difficult to apply in clinical settings.
Using infrared spectroscopy combined with a classification model, a classification model was constructed through data preparation, preprocessing, feature extraction, and model training to achieve rapid identification of the KL genotype of Klebsiella pneumoniae. Specifically, a neural network model was used for high-precision classification.
It enables high-precision identification of Klebsiella pneumoniae KL genotype within minutes with an accuracy rate of 98%, simplifies operation, reduces costs, and is suitable for rapid clinical diagnosis and precision medication.
Smart Images
Figure CN122392637A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of rapid microbial detection technology, and in particular to a rapid identification method and system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy. Background Technology
[0002] Klebsiella pneumoniae is a common Gram-negative bacillus that can cause pneumonia, sepsis, and urinary tract infections. The KL1 genotype is closely associated with high virulence, often leading to invasive infections and high mortality rates. Currently, existing methods for identifying Klebsiella pneumoniae genotypes mainly include serological testing, molecular biological methods (such as PCR), and gene sequencing. While these methods are accurate, they suffer from drawbacks such as complex operation, long processing time (usually several hours to days), high cost, and the need for specialized equipment and personnel, making them difficult to rapidly apply in clinical settings.
[0003] Infrared spectroscopy is a rapid and non-destructive analytical method that has been widely used in microbial identification and classification. By measuring the infrared spectrum of microorganisms, their chemical composition and structural information can be obtained, enabling rapid clustering and identification. However, there is currently no mature method in the field of infrared spectroscopy for direct and high-precision identification of Klebsiella pneumoniae KL genotypes (especially the highly virulent KL1 genotype).
[0004] Therefore, developing a method and system for rapid identification of Klebsiella pneumoniae KL genotype based on infrared spectroscopy is of great significance for clinical diagnosis and infection control. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to provide a rapid identification method and system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy in order to overcome the problems existing in the prior art mentioned above.
[0006] The technical solution adopted by this invention to solve its technical problem is: a rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy, comprising the following steps:
[0007] S1. Data preparation: Collect several strains of Klebsiella pneumoniae covering multiple KL genotypes, and collect multiple infrared spectra for each strain of Klebsiella pneumoniae, corresponding to one KL genotype tag;
[0008] S2. Data preprocessing and feature extraction: The acquired infrared spectrum is preprocessed, and feature bands are extracted from the preprocessed infrared spectrum;
[0009] S3. Classification Model Construction and Training: The classification model is trained using preprocessed infrared spectra;
[0010] S4. Strain identification: The infrared spectrum of the Klebsiella pneumoniae to be tested after pretreatment is input into the trained classification model, and the classification model outputs the probability of belonging to different KL genotypes.
[0011] Furthermore, step S2 also includes feature analysis: standard linear principal component analysis is used to perform dimensionality reduction and clustering verification on the preprocessed infrared spectrum to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.
[0012] Furthermore, in step S4, the classification model adopts a random forest model, a support vector machine model, a neural network model, or a Lasso logistic regression model.
[0013] Furthermore, the classification model employs a neural network model.
[0014] Furthermore, during the classification model training in step S4, the preprocessed infrared spectrum is divided into a training set and a test set. The training set is used to train the classification model, and the test set is used to verify the trained classification model.
[0015] Furthermore, the preprocessing in step S2 includes normalization, smoothing, and second derivative processing.
[0016] Furthermore, in step S1, there are at least 60 different KL genotypes.
[0017] Furthermore, in step S1, the multiple KL genotypes include at least the KL1 genotype.
[0018] A rapid identification system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy, employing the aforementioned rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy, includes:
[0019] The data acquisition module collects several strains of Klebsiella pneumoniae covering multiple KL genotypes and acquires multiple infrared spectra of each strain.
[0020] The data processing module preprocesses the acquired infrared spectrum and extracts characteristic spectral bands from the preprocessed infrared spectrum.
[0021] The model training module uses preprocessed infrared spectra to train the classification model and optimize model parameters;
[0022] It also includes a strain identification module that outputs the probability that the Klebsiella pneumoniae being tested belongs to different KL genotypes.
[0023] Furthermore, the data processing module includes a feature analysis submodule, which performs dimensionality reduction and clustering verification on the preprocessed infrared spectrum using standard linear principal component analysis to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.
[0024] The beneficial effects of this invention are:
[0025] (1) This invention utilizes infrared spectroscopy combined with a classification model to complete the KL genotyping of Klebsiella pneumoniae in a single test within minutes. The operation is simple, greatly shortens the identification time, and has an accuracy rate of up to 98%. No specific primers, antibodies, or sequencing reagents are required during the identification process. The cost of a single test is extremely low, and it does not require the operation of professional personnel.
[0026] (2) This invention achieves the typing of multiple KL genotypes through a classification model, and can directly and quickly identify the KL1 high virulence genotype corresponding to the K1 genotype, providing a key basis for early clinical diagnosis and precision medication. Attached Figure Description
[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0028] Figure 1 This is a flowchart of the rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy in this invention;
[0029] Figure 2 This is a schematic diagram of the rapid identification system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy in this invention;
[0030] Figure 3 This is a two-dimensional clustering diagram of principal component analysis of different KL genotypes of Klebsiella pneumoniae in this invention;
[0031] Figure 4 This is a three-dimensional clustering diagram of principal component analysis of different KL genotypes of Klebsiella pneumoniae in this invention;
[0032] Figure 5 This is the ROC curve of the random forest model in this invention;
[0033] Figure 6 This is the ROC curve of the support vector machine model in this invention;
[0034] Figure 7 This is the ROC curve of the neural network model in this invention;
[0035] Figure 8 This is the ROC curve of the Lasso logistic regression model in this invention;
[0036] Figure 9 This is the distance matrix diagram of the random forest model in this invention;
[0037] Figure 10 This is the distance matrix diagram of the support vector machine model in this invention;
[0038] Figure 11This is the distance matrix diagram of the neural network model in this invention;
[0039] Figure 12 This is the distance matrix diagram of the Lasso logistic regression model in this invention.
[0040] In the diagram: 100, Data Acquisition Module; 200, Data Processing Module; 210, Feature Analysis Submodule; 300, Model Training Module; 400, Strains Identification Module. Detailed Implementation
[0041] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.
[0042] Before describing the embodiments, let me explain why there is no mature method in the prior art for directly and accurately identifying the KL genotype of Klebsiella pneumoniae (especially the highly virulent KL1 type) using infrared spectroscopy.
[0043] It is generally believed in the industry that infrared spectroscopy mainly reflects the overall chemical composition of microbial cells (such as proteins, lipids, and polysaccharides). For different serotypes or genotypes within the same species that are closely related and have highly similar chemical compositions, the spectral differences are extremely subtle, resulting in a low signal-to-noise ratio. Traditional spectral analysis and simple statistical methods cannot achieve stable and reliable differentiation. Therefore, the industry generally limits the application of infrared spectroscopy to higher taxonomic levels (such as genus and species) and has not attempted to apply it to identification scenarios requiring extremely high resolution, such as the KL genotype of Klebsiella pneumoniae. In addition, how to construct a classification model that can accurately map to a specific KL genotype from complex, high-dimensional spectral data has been a long-standing technical challenge in this field. Simple pattern matching or distance clustering methods cannot handle this highly complex and nonlinear classification problem, resulting in low identification accuracy and failing to meet clinical identification requirements.
[0044] Example 1
[0045] like Figure 1 As shown, a rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy includes the following steps:
[0046] S1. Data preparation: Collect several strains of Klebsiella pneumoniae covering multiple KL genotypes, and collect multiple infrared spectra for each strain of Klebsiella pneumoniae, corresponding to one KL genotype tag.
[0047] Specifically, at least 350 strains of Klebsiella pneumoniae were collected. At least 60 different KL genotypes were included, with at least the KL1 genotype, preferably including KL1, KL2, KL3, KL5, ..., KL149. Ten to twelve infrared spectra of each Klebsiella pneumoniae strain were acquired using an FTIR spectrometer to enhance data representativeness.
[0048] S2. Data preprocessing and feature extraction: The acquired infrared spectrum is preprocessed, and feature bands are extracted from the preprocessed infrared spectrum.
[0049] Specifically, preprocessing includes normalization, smoothing, and second-derivative processing to eliminate background interference and enhance the resolution of characteristic spectral bands. Normalization, smoothing, and second-derivative processing are existing techniques and will not be elaborated upon here. Characteristic spectral bands are extracted from the preprocessed infrared spectrum, focusing on characteristic regions associated with the KL1 genotype, such as polysaccharide and lipopolysaccharide-related bands, for example, the 1200-900 cm⁻¹ region. -1 Spectral bands within the range.
[0050] S3. Classification Model Construction and Training: The classification model is trained using preprocessed infrared spectra.
[0051] Specifically, the classification model employs a random forest model, a support vector machine model, a neural network model, or a Lasso logistic regression model, with a neural network model being the preferred choice. During classification model training, the preprocessed infrared spectrum is divided into a training set and a test set. The training set is used to train the classification model, and the test set is used to validate the trained classification model.
[0052] It should be noted that the neural network model was preferred for classification because, after cross-validation, it achieved an accuracy of 98%, an F1 score of 0.99, and an expected area under the ROC curve close to 1. Furthermore, a clear diagonal structure is observed in the inter-class distance matrix output by the model, indicating high clustering of samples within each class and good separability between classes, further validating the model's discriminative ability. Performance evaluations of other models can be found in [link to performance evaluation]. Figures 5-12 And the following table:
[0053]
[0054] S4. Strain identification: The infrared spectrum of the Klebsiella pneumoniae to be tested after pretreatment is input into the trained classification model, and the classification model outputs the probability of belonging to different KL genotypes.
[0055] Specifically, if the classification model outputs the KL1 genotype, then the Klebsiella pneumoniae is identified as a highly virulent Klebsiella pneumoniae, providing a key basis for early clinical diagnosis and precise medication.
[0056] By using infrared spectroscopy combined with a classification model, the KL genotype of the strain can be completed in a single test within minutes. The operation is simple, significantly shortening the identification time, with an accuracy rate of up to 98%. No specific primers, antibodies, or sequencing reagents are required during the identification process, resulting in extremely low cost per test and no reliance on professional personnel.
[0057] In some embodiments, step S2 further includes feature analysis: using standard linear principal component analysis (PCA) to perform dimensionality reduction and clustering verification on the preprocessed infrared spectra in order to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.
[0058] like Figure 3 and Figure 4 As shown in the figure, the infrared spectra of KL1 genotype Klebsiella pneumoniae can be clearly clustered in unsupervised mode, indicating that its spectral characteristics are unique.
[0059] Example 2
[0060] like Figure 2 As shown, a rapid identification system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy is used, employing the rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy in Example 1, including:
[0061] The data acquisition module 100 collects several strains of Klebsiella pneumoniae covering multiple KL genotypes and acquires multiple infrared spectra of each strain of Klebsiella pneumoniae.
[0062] The data processing module 200 preprocesses all the collected infrared spectra and extracts characteristic spectral bands from the preprocessed infrared spectra.
[0063] The model training module 300 uses preprocessed infrared spectra to train the classification model and optimize model parameters;
[0064] And the strain identification module 400 outputs the probability that the Klebsiella pneumoniae to be tested belongs to different KL genotypes.
[0065] In some embodiments, the data processing module 200 includes a feature analysis submodule 210, which performs dimensionality reduction and clustering verification on the preprocessed infrared spectrum using standard linear principal component analysis to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.
[0066] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.
Claims
1. A rapid method for identifying the KL genotype of Klebsiella pneumoniae based on infrared spectroscopy, characterized in that, Includes the following steps: S1. Data preparation: Collect several strains of Klebsiella pneumoniae covering multiple KL genotypes, and collect multiple infrared spectra for each strain of Klebsiella pneumoniae, corresponding to one KL genotype tag; S2. Data preprocessing and feature extraction: The acquired infrared spectrum is preprocessed, and feature bands are extracted from the preprocessed infrared spectrum; S3. Classification Model Construction and Training: The classification model is trained using preprocessed infrared spectra; S4. Strain identification: The infrared spectrum of the Klebsiella pneumoniae to be tested after pretreatment is input into the trained classification model, and the classification model outputs the probability of belonging to different KL genotypes.
2. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1, characterized in that, Step S2 also includes feature analysis: standard linear principal component analysis is used to perform dimensionality reduction and clustering verification on the preprocessed infrared spectrum to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.
3. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1, characterized in that, In step S4, the classification model can be a random forest model, a support vector machine model, a neural network model, or a Lasso logistic regression model.
4. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 3, characterized in that, The classification model uses a neural network model.
5. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1, characterized in that, In step S4, during the classification model training, the preprocessed infrared spectrum is divided into a training set and a test set. The training set is used to train the classification model, and the test set is used to verify the trained classification model.
6. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1, characterized in that, The preprocessing in step S2 includes normalization, smoothing, and second derivative processing.
7. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1, characterized in that, In step S1, there are at least 60 different KL genotypes.
8. The rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 1 or 7, characterized in that, In step S1, the multiple KL genotypes include at least the KL1 genotype.
9. A rapid identification system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy, employing the rapid identification method for Klebsiella pneumoniae KL genotype based on infrared spectroscopy as described in any one of claims 1-8, characterized in that, include: The data acquisition module (100) collects several strains of Klebsiella pneumoniae covering multiple KL genotypes and acquires multiple infrared spectra of each strain of Klebsiella pneumoniae. The data processing module (200) preprocesses the acquired infrared spectrum and extracts characteristic spectral bands from the preprocessed infrared spectrum. The model training module (300) uses preprocessed infrared spectra to train the classification model and optimize the model parameters; And a strain identification module (400) that outputs the probability that the Klebsiella pneumoniae to be tested belongs to different KL genotypes.
10. The rapid identification system for Klebsiella pneumoniae KL genotype based on infrared spectroscopy according to claim 9, characterized in that, The data processing module (200) includes a feature analysis submodule (210), which performs dimensionality reduction and clustering verification on the preprocessed infrared spectrum using standard linear principal component analysis to verify the independent clustering characteristics of KL1 genotype Klebsiella pneumoniae.