Microorganism rapid detection and identification method and application thereof

Abstract

The invention provides a rapid microorganism detection and identification method and application thereof, and relates to the field of food and medicine quality control, the method comprises the following steps: (1) preparing a surface enhanced Raman scattering (SERS) substrate material; (2) preparing a microbial sample; (3) collecting an SERS spectrum; (4) construction and optimization of a microorganism identification model: the construction and application of the microorganism identification model are to pre-process SERS spectral data, and a machine learning algorithm is used to establish and train a microorganism classification model; calculating a model performance evaluation index according to a real result and a prediction result, determining an optimal model and applying the optimal model to the microorganism sample; the model performance evaluation index is accurate, simultaneous identification of multiple microorganisms in a mixed microorganism sample can be achieved, the detection time is short, the accuracy is high, a large amount of time cost and labor cost needed by testing are saved, the analysis efficiency is improved, and the detection cost is reduced.

Description

Technical Field

[0001] The invention belongs to the field of food and drug quality control, and specifically relates to a microbial rapid detection and identification method and application thereof. Background Art

[0002] Foodborne diseases are one of the most serious public health problems in the world, which puts heavy pressure on healthcare systems and hinders the development of economies and trade. About one-third of foodborne diseases are caused by contaminated food or drugs, and bacterial contamination, as the main contamination factor, may occur at all stages of the food and drug supply chain from production to distribution. Therefore, rapid monitoring of foodborne pathogens is crucial to preventing foodborne diseases.

[0003] At present, the microbial detection methods used by food and drug manufacturers and inspection agencies are mainly traditional culture methods, such as the plate method, membrane filtration method and the most probable number method. This type of method has mature technology and simple equipment, but the operation process is cumbersome, the detection cycle is long, the sensitivity is low, and the accuracy is poor. Surface-enhanced Raman scattering (SERS) refers to the adsorption of the molecules of the test object on the rough surface of metal nanomaterials, which can enhance the Raman signal of the test object by 10 6 -10 14 As an emerging Raman spectroscopy detection technology, SERS inherits the fingerprint specificity of Raman spectroscopy; it solves the problem of low sensitivity of ordinary Raman spectroscopy and is one of the most sensitive analytical technologies today; it requires very little sample volume and no pretreatment or only simple pretreatment; it is multiplexed and can detect multiple targets at the same time under a single excitation wavelength. It has been greatly developed and applied in biomedical imaging, disease diagnosis and screening, environmental monitoring, food safety and other aspects.

[0004] However, the actual application of this technology in the real world still faces obstacles such as interference from complex background signals in food or drug matrices and difficulty in distinguishing characteristic peaks. For example, the prior art CN117491339A provides a SERS detection method and system for rapid identification of pathogenic bacteria, and establishes a rapid and accurate pathogen detection method that can detect 8×10 7 CFU / mL of pathogenic bacteria, but the safety value stipulated in food safety is much less than 8×10 7 CFU / mL. There is an urgent need to develop a more rapid, sensitive, accurate and efficient microbial detection method for on-site monitoring. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a method for rapid detection and identification of microorganisms and its application.

[0006] In one aspect, the present invention provides a method for rapid detection and identification of microorganisms.

[0007] Specifically, the method comprises the following steps:

[0008] (1) Preparation of SERS substrate materials:

[0009] Take the fifth generation polyamide-amine dendrimer aqueous solution and add it to the gold nanoparticle solution A. After stirring in the dark, add the gold nanoparticle solution B dropwise to the mixture. Continue stirring until the mixture has been stirred for 10 minutes. -9 -10 -6 M fifth generation polyamide-amine dendrimer aqueous solution pre-treated silicon wafer to obtain SERS substrate material;

[0010] (2) Preparation of microbial sample solution:

[0011] The microorganisms are inoculated into a tryptic soytone liquid medium for overnight culture to obtain a microorganism stock solution, which is then treated and diluted to obtain a single bacterial sample solution; the single bacterial sample solutions are respectively mixed according to a proportion to obtain a mixed bacterial sample solution;

[0012] (3) Collection of SERS spectra;

[0013] (4) Construction and optimization of microbial identification model:

[0014] The SERS spectral data of the microbial samples obtained in step (3) are preprocessed, and a microbial classification model is established and trained using a machine learning algorithm; the model parameters are optimized, and the model performance evaluation index is calculated based on the actual results and the predicted results to determine the optimal model S A ; The model performance evaluation indicator is accuracy (ACC).

[0015] More specifically, in step (1), the final concentration of the fifth generation polyamide-amine dendrimer solution is 10 -8 -10 -6 M.

[0016] More specifically, in step (1), the particle size of the gold nanoparticle solution A is 10-20 nm; the particle size of the solution B is 5-10 nm.

[0017] Preferably, the particle size of the gold nanoparticle solution A is 15 nm; the particle size of solution B is 5 nm.

[0018] More specifically, in step (2), the microorganism includes but is not limited to: at least one of Escherichia coli (E. coli), Pseudomonas aeruginosa (P. aeruginosa), Staphylococcus aureus (S. aureus) and Salmonella Paratyphi B (S. Paratyphi B).

[0019] More specifically, in step (2), the culture conditions of the microorganism are: temperature 30-35° C., time 18-48 hours.

[0020] Specifically, in step (3), the SERS spectrum is collected under the following conditions: excitation wavelength 532 nm, scanning spectrum range 50-3000 cm -1 , laser intensity 1-300mW, acquisition time 0.5-20s / time, accumulation times 1-300 times, 50-1000 spectra were collected for each sample.

[0021] More specifically, in step (4), the preprocessing method includes but is not limited to: at least one of baseline removal, Savitzky-Golay smoothing, first-order derivative, and second-order derivative.

[0022] More specifically, in step (4), the machine learning algorithm includes but is not limited to: decision tree, random forest, K nearest neighbor, naive Bayes, support vector machine, neural network, GBDT (Gradient Boosting Decison Tree), XGBoost (eXtreme Gradient Boosting), LightGBM (Light Gradient Boosting Machine).

[0023] Preferably, the machine learning algorithm is LightGBM, and the parameters are: boosting: gbdt; objective: binary; metric: auc; nthread: 1-9; learning_rate: 0.05-0.25; num_leaves: 1-99; max_depth: 1-9; max_bin: 1-999; min_data_in_leaf: 1-499; feature_fraction: 0.1-1.0; bagging_fraction: 0.1-1.0; bagging_freq: 1-99; lambda_l1: 0-1; lambda_l2: 0-1; min_split_gain: 0-0.5; min_sum_hessian_in_leaf: 0-0.01; random_state: 0-9999; verbose: -1; others are default values.

[0024] In another aspect, the present invention provides application of the above method in product quality control, wherein the product includes: food and / or medicine.

[0025] Specifically, the food is milk and dairy products and / or beverages.

[0026] Preferably, the food includes milk, juice and sparkling water.

[0027] Specifically, the medicines include: Chinese medicinal materials, Chinese medicinal pieces, Chinese medicinal extracts and / or Chinese medicinal preparations.

[0028] Preferably, the medicine comprises Dendrobium officinale and / or Ganoderma lucidum.

[0029] Specifically, the method comprises the following steps:

[0030] (a) Preparation of SERS substrate materials;

[0031] (b) Preparation of bacteria-containing samples:

[0032] Take samples for sterilization, add mixed bacterial sample solution, shake and resuspend them fully to obtain bacteria-containing samples;

[0033] (c) SERS spectrum collection;

[0034] (d) Optimization of microbial identification model:

[0035] Based on the optimal model S A The SERS spectrum data of the bacteria-containing sample obtained in step (c) is preprocessed; based on S A Fine-tune parameters and determine the optimal model S B; Use performance evaluation indicators to evaluate model performance; the model performance evaluation indicator is accuracy (ACC);

[0036] (e) Identification analysis of bacteria-containing samples.

[0037] More specifically, in step (c), the platform for collecting the SERS spectrum includes a spontaneous Raman system, a laser Raman system, a stimulated Raman system, a Fourier transform infrared Raman spectroscopy, a portable Raman system, an atomic force Raman system, a surface enhanced Raman system and a tip enhanced Raman system.

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] (1) Compared with the existing methods, the present invention provides a method for detecting and identifying microorganisms. It does not require separation, purification, cultivation or enrichment operations, but only requires simple sample preparation to achieve simultaneous identification of multiple microorganisms in a mixed microbial sample. It has a short detection time and high accuracy, saves a lot of time and manpower costs required for testing, improves analysis efficiency, and reduces detection costs.

[0040] (2) The present invention provides a method for detecting and identifying microorganisms. Based on the excellent feature extraction capability of machine learning, it successfully achieves accurate identification of multiple microorganisms simultaneously under the interference of complex background Raman signals of food and traditional Chinese medicine related products, which is beneficial to the actual production and use of factories, and provides important technical support for food and drug quality risk assessment and management, and is of great significance.

[0041] (3) The present invention provides a method for detecting and identifying microorganisms, which is convenient and fast, requires a small amount of sample, has a wide range of applications, and can be used in conjunction with a portable Raman spectrometer to conduct on-site and field experiments. It can be expanded to more types of microorganisms and detection scenarios, and has the potential for actual on-site detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the present invention.

[0043] Figure 2 These are the SERS spectra of different microorganisms in the present invention after baseline removal pretreatment. DETAILED DESCRIPTION

[0044] The present invention will be further described in detail below in conjunction with specific examples. The following examples are not intended to limit the present invention, but are only intended to illustrate the present invention. The experimental methods used in the following examples are generally conventional, unless otherwise specified, and the materials, reagents, etc. used in the following examples are commercially available, unless otherwise specified.

[0045] The flow chart of the present invention is shown in Figure 1 The reagents and equipment used are shown in Table 1.

[0046] Table 1 Manufacturer, batch number / model of reagents / equipment

[0047]

[0048]

[0049] Example 1 Construction of a method for rapid detection and identification of microorganisms

[0050] (1) Preparation of SERS substrate materials:

[0051] Take 500 μL 10 -6 The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0052] (2) Preparation of single bacterial sample solution:

[0053] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ), and the concentration of the bacterial solution was calculated by comparing with the standard curve. The four microbial stock solutions were taken separately, diluted with sterile water, and the gradient concentrations were 10, 10 2 , 10 3 , 10 4 , 10 5, 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of a single bacterial species sample.

[0054] (3) Collection of SERS spectra:

[0055] The gradient concentrations were 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 Four microbial single strain samples with CFU / mL and sterile water (negative control) were dripped onto the surface of the SERS substrate material and the SERS spectra were collected after drying. The collection conditions were: excitation wavelength of 532nm, scanning spectrum range of 50-3000cm 1 , the laser intensity was 30 mW, the acquisition time was 5 s / time, the accumulation times were 200 times, and 100 spectra were collected for different types of samples.

[0056] (4) Construction and optimization of microbial identification model:

[0057] The SERS data of the single bacterial sample obtained in step (3) were preprocessed using algorithms such as baseline removal, Savitzky-Golay smoothing, first-order derivative, second-order derivative, baseline removal followed by Savitzky-Golay smoothing, baseline removal followed by first-order derivative, and baseline removal followed by second-order derivative. The preprocessed data are recorded as data set S A-1 .

[0058] Use LightGBM algorithm to establish and train microbial classification model, set classification model training data set and detection data set; randomly select data set S A-1 The 75% subset is S A-1 Classification model training data set, 25% subset as S A-1 Detection data set. Using accuracy (ACC) as the performance evaluation indicator, the optimal data preprocessing method is determined to be removing the baseline first and then the second-order derivative algorithm.

[0059] The model was trained using a ten-fold crossover operation to optimize the model parameters; the accuracy (ACC) was used as the performance evaluation indicator to determine S A-1 Optimal model parameters. The accuracy results of the optimal model corresponding to the four microbial classifications are shown in Table 2. The accuracy is the proportion of the correct number of model predictions to the total number calculated based on the actual results and the predicted results.

[0060] Table 2 Classification accuracy of microbial identification model for four microorganisms of a single species

[0061] Accuracy (%) Escherichia coli Pseudomonas aeruginosa Staphylococcus aureus Salmonella Paratyphi B Single strain 94.9 97.6 97.4 95.5

[0062] Example 2 Construction of a method for rapid simultaneous detection and identification of multiple microorganisms

[0063] (1) Preparation of SERS substrate materials:

[0064] Take 500 μL 10 -6 The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0065] (2) Preparation of mixed bacterial sample solution:

[0066] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ), and the concentration of the bacterial solution was calculated by comparing with the standard curve. The original solutions of the four microorganisms were diluted with sterile water to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9CFU / mL single strain sample solution, and then four microbial single strain sample solutions were prepared according to the proportions in Table 3 to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of mixed bacterial sample solution, and sterile water as a negative control.

[0067] Table 3 Mixed bacterial sample solution preparation ratio

[0068]

[0069]

[0070] (3) Collection of SERS spectra:

[0071] The gradient concentrations were 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of mixed bacterial sample solution and sterile water (negative control) were dripped onto the surface of the SERS substrate material, and the SERS spectrum was collected after drying. The collection conditions were: excitation wavelength of 532nm, scanning spectrum range of 50-3000cm 1 , the laser intensity was 30 mW, the acquisition time was 5 s / time, the accumulation times were 200 times, and 100 spectra were collected for different types of samples.

[0072] (4) Construction and optimization of microbial identification model:

[0073] Based on the optimal data preprocessing method determined in Example 1, the SERS spectrum data of the mixed bacterial sample solution obtained in step (3) is preprocessed and recorded as data set S A-2 .

[0074] The LightGBM algorithm was used to establish and train the microbial classification model, and the processed data set was input into the trained and optimized S A-1 Classification model: Randomly select data set S A-2 The 75% subset is S A-2 Classification model training data set, 25% subset as S A-2 Detection data set, and use data set S A-1 A random 50% subset expansion SA-2 Classification model training data set; based on S A-1 The optimal model is fine-tuned, and the model is trained using a ten-fold crossover operation. The accuracy (ACC) is calculated based on the actual results and the predicted results to determine S A-2 Optimal model parameters. The minimum detection limits of the four microorganisms in the mixed bacterial solution and the corresponding classification accuracy of the optimal model are shown in Table 4, where the accuracy is calculated as the proportion of the correct number predicted by the model to the total number.

[0075] Table 4 Minimum detection limits and classification accuracy of the microbial identification model for four microorganisms in mixed strains

[0076] Escherichia coli Pseudomonas aeruginosa Staphylococcus aureus Salmonella Paratyphi B Minimum detection limit (CFU / mL) 7.55×10 6.98×10 6.82×10 7.63×10 Mixed strains 94.3 96.7 96.9 94.6

[0077] Example 3 Application of a method for rapid simultaneous detection and identification of multiple microorganisms in milk

[0078] (1) Preparation of SERS substrate materials:

[0079] Take 500 μL 10 -6 The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0080] (2) Preparation of milk-containing bacteria sample solution:

[0081] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ) and calculate the bacterial concentration by comparing with the standard curve.

[0082] Take the four microbial stock solutions and dilute them with sterile water to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL single strain sample solution. The four microbial single strain sample solutions were prepared according to Table 3 to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of mixed bacterial sample solution, and sterile water as a negative control.

[0083] Take pasteurized milk and add the prepared mixed bacterial sample solution under aseptic operation, shake it thoroughly and resuspend it to obtain milk bacteria samples.

[0084] (3) Collection of SERS spectra:

[0085] The bacteria-containing milk sample and sterile water (negative control) were added to the surface of the SERS substrate material, and the SERS spectra were collected after drying. The collection conditions were: the excitation wavelength was 532nm, and the scanning spectrum range was 50-3000cm 1, the laser intensity was 30 mW, the acquisition time was 5 s / time, the accumulation times were 200 times, and 100 spectra were collected for different types of samples.

[0086] (4) Identification and analysis of bacteria-containing milk samples:

[0087] Based on the optimal data preprocessing method determined in Example 1, the SERS spectrum data of the milk-containing bacteria sample obtained in step (3) is preprocessed and recorded as data set S B-1 .

[0088] The LightGBM algorithm was used to establish and train the microbial classification model, and the processed data set was input into the trained and optimized S A-2 Classification model: Randomly select data set S B-1 The 75% subset is S B-1 Classification model training data set, 25% subset as S B-1 Detection data set, and use data set S A-2 A random 50% subset expansion S B-1 Classification model training data set; based on S A-2 The optimal model is fine-tuned, and the model is trained using a ten-fold crossover operation. The accuracy (ACC) is calculated based on the actual results and the predicted results to determine S B-1 Optimal model parameters.

[0089] The data in the reserved data set were input into the optimal model, and the minimum detection limit and classification accuracy of various microorganisms in milk were obtained as output (Table 5).

[0090] Table 5 Minimum detection limit and classification accuracy of microorganisms in milk by microbial identification model

[0091] Escherichia coli Pseudomonas aeruginosa Staphylococcus aureus Salmonella Paratyphi B Minimum detection limit (CFU / mL) 7.78×10 7.23×10 7.39×10 7.94×10 Accuracy (%) 93.4 95.1 96.0 92.8

[0092] Example 4 Application of a method for rapid simultaneous detection and identification of multiple microorganisms in a Dendrobium officinale leaf beverage

[0093] (1) Preparation of SERS substrate materials:

[0094] Take 500 μL 10 -6 The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0095] (2) Preparation of Dendrobium officinale leaf beverage containing bacteria sample liquid:

[0096] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ) and calculate the bacterial concentration by comparing with the standard curve.

[0097] Take the four microbial stock solutions and dilute them with sterile water to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL single strain sample solution. The four microbial single strain sample solutions were prepared according to Table 3 to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of mixed bacterial sample solution, and sterile water as a negative control.

[0098] The Dendrobium officinale leaf beverage was heated at 88°C for 1 min, and after cooling, the prepared mixed bacterial sample solutions were added under sterile operation, and the mixture was fully shaken and resuspended to obtain the Dendrobium officinale leaf beverage containing bacteria sample.

[0099] (3) Collection of SERS spectra:

[0100] The bacteria-containing sample of Dendrobium officinale leaf beverage and sterile water (negative control) were added to the surface of the SERS substrate material, and the SERS spectrum was collected after drying. The collection conditions were: the excitation wavelength was 532nm, and the scanning spectrum range was 50-3000cm 1 , the laser intensity was 30 mW, the acquisition time was 5 s / time, the accumulation times were 200 times, and 100 spectra were collected for different types of samples.

[0101] (4) Identification and analysis of bacteria-containing samples of Dendrobium officinale leaf beverage:

[0102] Based on the optimal data preprocessing method determined in Example 1, the SERS spectrum data of the Dendrobium officinale leaf beverage containing bacteria sample obtained in step (3) was preprocessed and recorded as data set S B-2 .

[0103] The LightGBM algorithm was used to establish and train the microbial classification model, and the processed data set was input into the trained and optimized S A-2 Classification model: Randomly select data set S B-2 The 75% subset is S B-2 Classification model training data set, 25% subset as S B-2 Detection data set, and use data set S A-2 A random 50% subset expansion S B-2 Classification model training data set; based on S A-2 The optimal model is fine-tuned, and the model is trained using a ten-fold crossover operation. The accuracy (ACC) is calculated based on the actual results and the predicted results to determine S B-2 Optimal model parameters.

[0104] The data in the reserved data set were input into the optimal model, and the minimum detection limit and classification accuracy of various microorganisms in the Dendrobium officinale leaf beverage were obtained as output (Table 6).

[0105] Table 6 Minimum detection limit and classification accuracy of microorganisms in Dendrobium officinale leaf beverage by microbial identification model

[0106] Escherichia coli Pseudomonas aeruginosa Staphylococcus aureus Salmonella Paratyphi B Minimum detection limit (CFU / mL) 7.21×10 6.84×10 6.73×10 7.46×10 Accuracy (%) 93.6 95.6 96.3 93.2

[0107] Example 5 Application of a method for rapid simultaneous detection and identification of multiple microorganisms in Ganoderma lucidum spore powder

[0108] (1) Preparation of SERS substrate materials:

[0109] Take 500 μL 10 -6The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0110] (2) Preparation of Ganoderma lucidum spore powder containing bacteria sample:

[0111] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ) and calculate the bacterial concentration by comparing with the standard curve.

[0112] Take the four microbial stock solutions and dilute them with sterile water to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL single strain sample solution. The four microbial single strain sample solutions were prepared according to Table 3 to obtain gradient concentrations of 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of mixed bacterial sample solution, and sterile water as a negative control.

[0113] The Ganoderma lucidum spore powder was heated at 121°C for 15 min, and after cooling, the prepared mixed bacterial sample solution was added under sterile operation, and the mixture was fully shaken and resuspended to obtain the Ganoderma lucidum spore powder containing bacteria sample.

[0114] (3) Collection of SERS spectra:

[0115] Ganoderma lucidum spore powder and sterile water (negative control) were added to the surface of the SERS substrate material, and the SERS spectrum was collected after drying. The collection conditions were: excitation wavelength of 532nm, scanning spectrum range of 50-3000cm 1 , the laser intensity was 30 mW, the acquisition time was 5 s / time, the accumulation times were 200 times, and 100 spectra were collected for different types of samples.

[0116] (4) Identification and analysis of Ganoderma lucidum spore powder containing bacteria:

[0117] Based on the optimal data preprocessing method determined in Example 1, the SERS spectrum data of the Ganoderma lucidum spore powder containing bacteria sample liquid obtained in step (3) was preprocessed and recorded as data set S B-3 .

[0118] The LightGBM algorithm was used to establish and train the microbial classification model, and the processed data set was input into the trained and optimized S A-2 Classification model: Randomly select data set S B-3 The 75% subset is S B-3 Classification model training data set, 25% subset as S B-3 Detection data set, and use data set S A-2 A random 50% subset expansion S B-3 Classification model training data set; based on S A-2 The optimal model is fine-tuned, and the model is trained using a ten-fold crossover operation. The accuracy (ACC) is calculated based on the actual results and the predicted results to determine S B-3 Optimal model parameters.

[0119] The data in the reserved data set were input into the optimal model, and the minimum detection limit and classification accuracy of various microorganisms in Ganoderma lucidum spore powder were obtained as output (Table 7).

[0120] Table 7 Minimum detection limit and classification accuracy of microorganisms in Ganoderma lucidum spore powder by microbial identification model

[0121]

[0122]

[0123] Comparative Example 1 Construction of a method for rapid detection of microorganisms

[0124] (1) Preparation of SERS substrate materials:

[0125] Take 500 μL 10 -6 The fifth generation polyamide-amine dendrimer aqueous solution was added to 5 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 15 nm. After stirring for 2 hours in a dark environment, 15 mL of a 0.05 mg / mL gold nanoparticle solution with a particle size of 5 nm was added dropwise to the above mixture. Stirring continued for 6 hours, and then added to the mixture after 10 -8 The surface solution was evaporated at room temperature on a silicon wafer pretreated with a fifth-generation polyamide-amine dendrimer aqueous solution to obtain a SERS substrate material.

[0126] (2) Preparation of single bacterial sample solution:

[0127] Escherichia coli (E.coli, CMCC(B)44102), Pseudomonas aeruginosa (P.aeruginosa, CMCC(B)10104), Staphylococcus aureus (S.aureus, CMCC(B)26003) and Salmonella Paratyphi B (S.Paratyphi B, CMCC(B)50094) were inoculated into tryptic soy bean liquid medium and cultured at 35°C for 24 hours. 5 mL of bacterial solution was taken into 15 mL centrifuge tubes, centrifuged at 4000×g for 10 minutes, and the supernatant was discarded. 5 mL of sterile water was added to the precipitate, and the mixture was mixed by pipetting and centrifugation. After repeating 2-3 times, 1 mL of sterile water was added to dilute the mixture, which was the stock solution of microorganisms. Take 100 μL and measure the absorbance value (OD) at 600 nm in an ELISA instrument. 600 ), and the concentration of the bacterial solution was calculated by comparing with the standard curve. The four microbial stock solutions were taken separately, diluted with sterile water, and the gradient concentrations were 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 CFU / mL of a single bacterial species sample.

[0128] (3) Collection of SERS spectra:

[0129] The gradient concentrations were 10, 10 2 , 103 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 and 10 9 Four microbial single strain samples with CFU / mL and sterile water (negative control) were dripped onto the surface of the SERS substrate material and the SERS spectra were collected after drying. The collection conditions were: excitation wavelength of 532nm, scanning spectrum range of 50-3000cm 1 , laser intensity is 30mW, acquisition time is 5s / time, accumulation times are 200 times, and 100 spectra are collected for different types of samples. Baseline removal is performed on the collected SERS spectra. Figure 2 Table 8 shows the SERS spectra of the four microorganisms after baseline removal. Table 8 shows the minimum detection limit results of the four microorganisms.

[0130] Table 8 Minimum detection limits of four microorganisms

[0131] Escherichia coli Pseudomonas aeruginosa Staphylococcus aureus Salmonella Paratyphi B Minimum detection limit (CFU / mL) <![CDATA[6.73×10 2 ]]> <![CDATA[5.87×10 2 ]]> <![CDATA[6.21×10 2 ]]> <![CDATA[6.54×10 2 ]]>

[0132] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A method for rapid detection and identification of microorganisms, characterized in that: The method comprises the following steps: (1) Preparation of SERS substrate materials: Take the fifth generation polyamide-amine dendrimer aqueous solution and add it to the gold nanoparticle solution A. After stirring in the dark, add the gold nanoparticle solution B dropwise to the mixture. Continue stirring until the mixture has been stirred for 10 minutes. -9 -10 -6 M fifth generation polyamide-amine dendrimer aqueous solution pre-treated silicon wafer to obtain SERS substrate material; (2) Preparation of microbial sample solution: The microorganisms are inoculated into a tryptic soytone liquid medium for overnight culture to obtain a microorganism stock solution, which is then treated and diluted to obtain a single bacterial sample solution; the single bacterial sample solutions are respectively mixed according to a proportion to obtain a mixed bacterial sample solution; (3) Collection of SERS spectra: The acquisition conditions were: excitation wavelength 532 nm, scanning spectrum range 50-3000 cm -1 , laser intensity 1-300mW, acquisition time 0.5-20s / time, accumulation times 1-300 times, 50-1000 spectra were collected for different types of samples; (4) Construction and optimization of microbial identification model: The SERS spectral data of the microbial samples obtained in step (3) are preprocessed, and a microbial classification model is established and trained using a machine learning algorithm; the model parameters are optimized, and the model performance evaluation index is calculated based on the actual results and the predicted results to determine the optimal model S A ; The model performance evaluation indicator is accuracy.

2. The method according to claim 1, characterized in that In step (1), the particle size of the gold nanoparticle solution A is 10-20 nm; the particle size of the solution B is 5-10 nm.

3. The method according to claim 1, characterized in that In step (2), the microorganisms include at least one of Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus and / or Salmonella paratyphi B.

4. The method according to claim 1, characterized in that: In step (4), the preprocessing method includes at least one of baseline removal, Savitzky-Golay smoothing, first-order derivative, and second-order derivative.

5. The method according to claim 1, characterized in that In step (4), the machine learning algorithm includes decision tree, random forest, K nearest neighbor, naive Bayes, support vector machine, neural network, GBDT, XGBoost, and LightGBM.

6. The method according to claim 4, characterized in that In step (4), the machine learning algorithm is the LightGBM algorithm, and the parameters are: boosting: gbdt; objective: binary; metric: auc; nthread: 1-9; learning_rate: 0.05-0.25; num_leaves: 1-99; max_depth: 1-9; max_bin: 1-999; min_data_in_leaf: 1-499; feature_fraction: 0.1-1.0; bagging_fraction: 0.1-1.0; bagging_freq: 1-99; lambda_l1: 0-1; lambda_l2: 0-1; min_split_gain: 0-0.5; min_sum_hessian_in_leaf: 0-0.01; random_state: 0-9999; verbose: -1; others are default values.

7. Application of the method according to any one of claims 1 to 6 in product quality control, wherein the product comprises: Food and / or medicine.

8. The use according to claim 7, characterized in that: The food includes: milk, juice, and sparkling water; the medicine includes: Chinese medicinal materials, Chinese medicinal pieces, Chinese medicinal extracts and / or Chinese medicinal preparations; preferably, the medicine includes Dendrobium officinale and / or Ganoderma lucidum.

9. The use according to claim 8, characterized in that: The following steps are involved: (a) Preparation of SERS substrate materials; (b) Preparation of bacteria-containing samples: Take samples for sterilization, add mixed bacterial sample solution, shake and resuspend them fully to obtain bacteria-containing samples; (c) SERS spectrum collection; (d) Optimization of microbial identification model: Based on the optimal model S A The SERS spectrum data of the food / traditional Chinese medicine containing bacteria sample obtained in step (c) is preprocessed; based on S A Fine-tune parameters and determine the optimal model S B ; The performance of the model is evaluated using a performance evaluation indicator; the model performance evaluation indicator is accuracy; (e) Identification analysis of bacteria-containing samples.

10. The method according to claim 9, characterized in that In step (c), the platforms for collecting SERS spectral signals include spontaneous Raman system, laser Raman system, stimulated Raman system, Fourier transform infrared Raman spectroscopy, portable Raman system, atomic force Raman system, surface enhanced Raman system and tip enhanced Raman system.

Images