WO2024061126A1 - 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备 - Google Patents

聚羟基脂肪酸酯含量的检测方法、装置、系统、设备 Download PDF

Info

Publication number
WO2024061126A1
WO2024061126A1 PCT/CN2023/119116 CN2023119116W WO2024061126A1 WO 2024061126 A1 WO2024061126 A1 WO 2024061126A1 CN 2023119116 W CN2023119116 W CN 2023119116W WO 2024061126 A1 WO2024061126 A1 WO 2024061126A1
Authority
WO
WIPO (PCT)
Prior art keywords
pha
raman spectrum
fermentation broth
sample
spectrum information
Prior art date
Application number
PCT/CN2023/119116
Other languages
English (en)
French (fr)
Inventor
周瑶瑶
吴雅琨
汪东升
李腾
张浩千
Original Assignee
北京蓝晶微生物科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京蓝晶微生物科技有限公司 filed Critical 北京蓝晶微生物科技有限公司
Publication of WO2024061126A1 publication Critical patent/WO2024061126A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
    • G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
    • G01N21/65—Raman scattering

Definitions

  • the present application relates to the field of biological detection, and in particular to a detection method, device, system and equipment for polyhydroxyalkanoate content.
  • PHA Polyhydroxyalkanoates
  • This application provides a detection method, device, system, and equipment for polyhydroxyalkanoate content to solve the technical defects of hysteresis, high cost, and low throughput in the existing PHA detection technology.
  • This application can detect fermentation The PHA Raman spectrum in the fermentation broth is modeled, and the detection results are effectively detected through the established analysis model to achieve non-destructive, efficient and accurate detection of PHA in the fermentation broth.
  • this application provides a method for detecting polyhydroxyalkanoate content, including:
  • the PHA content detection model includes a quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value
  • the quantitative relationship is based on the Raman spectral information of sample fermentation broth under different fermentation conditions and different Quantitative relationship obtained by training PHA content values of samples under the same fermentation conditions;
  • the PHA content value of the sample is determined based on gas chromatography detection of the sample fermentation broth.
  • the Raman spectrum information includes the frequency shift interval where the Raman peak corresponding to the PHA in the fermentation broth to be detected is located and the frequency shift interval described in the frequency shift interval.
  • the quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value is a quantitative relationship obtained through training based on the wavenumber intensity in the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the PHA content value.
  • the detection model based on the PHA content processes the Raman spectrum information, including:
  • the PHA content value in the fermentation broth to be detected is output.
  • the detection method for polyhydroxyalkanoate content after collecting the Raman spectrum information of the fermentation broth to be detected, it also includes:
  • the preprocessing method includes at least one of convolution smoothing processing, baseline correction processing, multivariate scattering correction processing, orthogonal signal correction processing, standard normal transformation processing, normalization processing, Gaussian filtering processing, and median filtering processing. kind.
  • the different fermentation conditions include:
  • a detection device for polyhydroxyalkanoate content including:
  • Collection unit used to collect Raman spectrum information of the fermentation liquid to be detected
  • Input unit used to input the Raman spectrum information to the detection model of polyhydroxyalkanoate PHA content
  • Processing unit used to process the Raman spectrum information based on the detection model of the PHA content, and output the PHA content value in the fermentation broth to be detected;
  • the PHA content detection model includes a quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value
  • the quantitative relationship is a quantitative relationship obtained by training based on the Raman spectrum information of sample fermentation broth under different fermentation conditions and the PHA content values of samples under different fermentation conditions;
  • the PHA content value of the sample is determined based on gas chromatography detection of the sample fermentation broth.
  • the processing unit further includes:
  • Mapping subunit used to map the frequency shift interval where the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is located to the corresponding frequency shift interval of the fingerprint sample in the detection model;
  • Determination subunit used to determine the wave number intensity of the Raman wave peak corresponding to the PHA in the fermentation broth to be detected in the frequency shift interval based on the corresponding fingerprint sample frequency shift interval;
  • Output subunit used to output the PHA content value in the fermentation broth to be detected through the quantitative relationship between the wave number intensity and the PHA content.
  • the detection device further includes:
  • Preprocessing unit used to preprocess the Raman spectrum information to obtain denoised Raman spectrum information
  • the preprocessing method includes at least one of convolution smoothing processing, baseline correction processing, multivariate scattering correction processing, orthogonal signal correction processing, standard normal transformation processing, normalization processing, Gaussian filtering processing, and median filtering processing. kind.
  • a detection system for polyhydroxyalkanoate content including:
  • Testing container used to provide a testing environment for the fermentation broth
  • the probe is used to immerse into the detection cell to collect Raman spectrum information
  • Optical fiber used for signal transmission between the probe and the excitation light source, as well as signal transmission between the probe and the signal detector;
  • Excitation light source used to provide detection light source for fermentation broth
  • a signal detector for converting the optical signal into a data signal
  • the detection device is used to analyze and process the collected Raman spectrum information and output the polyhydroxyalkanoate content in the fermentation broth to be measured.
  • an electronic device including a memory, a processor, and a computer program stored on the memory and executable on the processor.
  • the processor executes the program, the Method for detecting polyhydroxyalkanoate content.
  • This application provides a method, device, system, and equipment for detecting polyhydroxyalkanoate content.
  • a detection model for polyhydroxyalkanoate PHA content By inputting Raman spectrum information collected from the fermentation broth to be detected into a detection model for polyhydroxyalkanoate PHA content, Obtain the PHA content value in the fermentation broth to be detected; since the PHA content detection model is trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the sample PHA content values under different fermentation conditions, This makes the detection of the PHA content value finally obtained accurate.
  • This application can overcome the technical problems of being unable to detect in real time due to the complex components and long fermentation cycle of PHA during fermentation. It can effectively detect and detect through the constructed polyhydroxyalkanoate PHA detection model. As a result, the detection error value can be controlled within 8%, especially within 5%, achieving non-destructive, efficient and accurate detection of PHA in the fermentation broth.
  • FIG1 is a schematic diagram of a method for detecting polyhydroxyalkanoate content provided by the present application.
  • Figure 2 is a schematic flow chart of processing the Raman spectrum information provided by this application.
  • Figure 3 is the second schematic flow chart of the detection method for polyhydroxyalkanoate content provided by this application.
  • Figure 4 is a schematic structural diagram of the polyhydroxyalkanoate content detection system provided by the present application.
  • Figure 5 is a schematic structural diagram of a detection device for polyhydroxyalkanoate content provided by the present application.
  • Figure 6 is a schematic structural diagram of an electronic device provided by this application.
  • Polyhydroxyalkanoate is a high-molecular polyester compound produced by a variety of microorganisms in nature. As an energy reserve substance, it is widely present in microbial cells. Due to its excellent biodegradability and plasticity, it is widely used in agriculture and food. , medical and pharmaceutical industries. This application aims to provide a method for real-time online detection of PHA in fermentation broth to overcome the hysteresis, high cost, and low throughput of existing PHA detection technology. technical problem.
  • Figure 1 is one of the schematic flow diagrams of a method for detecting polyhydroxyalkanoate content provided by this application.
  • This application provides a method for detecting polyhydroxyalkanoate content, which includes:
  • the detection model based on the PHA content processes the Raman spectrum information and outputs the PHA content value in the fermentation broth to be detected;
  • the PHA content detection model includes a quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value
  • the quantitative relationship is a quantitative relationship obtained by training based on the Raman spectrum information of sample fermentation broth under different fermentation conditions and the PHA content values of samples under different fermentation conditions;
  • the PHA content value of the sample is determined based on gas chromatography detection of the sample fermentation broth.
  • the Raman spectrum information includes the frequency shift interval where the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is located and the wave number intensity of the Raman wave peak in the frequency shift interval.
  • This application collects The Raman spectrum information of the fermentation broth to be detected, the frequency shift interval where the Raman wave peak is located and the wave number intensity of the Raman wave peak in the frequency shift interval are extracted from the Raman spectrum information, the fermentation broth to be detected It can be the fermentation broth extracted from the detection pool, or it can be directly penetrated into the fermentation broth to be detected through a probe to collect the Raman spectrum information of the fermentation broth to be detected.
  • the Raman spectrum information is a kind of scattering spectrum, which is made by using molecules.
  • the difference in wavelength of the scattered light emitted after laser irradiation is used to destroy the material. It is a technology that performs characterization analysis. Due to the advantages of Raman spectroscopy detection, such as fast, non-destructive and high sensitivity, it is widely used in various fields, especially in the field of fermentation. Raman spectroscopy can obtain the status information of the current reaction in a timely manner without the need for Through the steps of sampling, processing, and re-testing, the throughput of detection can be greatly enhanced, while interference to the reaction caused by volume changes is also avoided.
  • the collected Raman spectra can optionally be preprocessed and mapped preprocessed.
  • the Raman spectrum information is then transferred to the preset band interval, the Raman wave peak is determined, the frequency shift interval where the Raman wave peak is located is determined as the target frequency shift interval, and the Raman wave peak is determined to be in the frequency shift interval.
  • the target wavenumber intensity in .
  • the Raman spectrum information is input into the detection model of polyhydroxyalkanoate PHA content.
  • the frequency shift interval and the wave number intensity are vectorized, and according to the The frequency shift interval and the vectorized representation of the wave number intensity construct a feature input matrix, and the feature input matrix is input to the polyhydroxyalkanoate PHA detection model to obtain the PHA content detection value output by the PHA detection model.
  • the Raman spectrum information is processed based on the detection model of the PHA content, and the PHA content value in the fermentation broth to be detected is output.
  • the detection model of the PHA content includes the Raman spectrum information of the fermentation broth.
  • the quantitative relationship with the PHA content value; the quantitative relationship is a quantitative relationship obtained by training based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the PHA content value of the sample under different fermentation conditions.
  • the Raman spectrum information into the detection model of polyhydroxyalkanoate PHA content also includes: based on sample data under different fermentation conditions and different fermentation times, determining a sample training set and a sample test set according to a preset ratio, A detection model for PHA content is constructed based on the sample training set and the sample test set.
  • the PHA content detection model is trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the PHA content values of the samples under different fermentation conditions. Specifically: the Raman spectrum information of the fermentation broth and the PHA content.
  • the quantitative relationship of the values is a quantitative relationship obtained through training based on the wavenumber intensity and PHA content value in the Raman spectrum information of the sample fermentation broth under different fermentation conditions, as follows:
  • Y represents the content value of PHA
  • the quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value is the quantitative relationship between the wave number intensity and the PHA content value, that is, the wave number intensity corresponds to the PHA content in the fermentation broth.
  • the PHA detection model is based on each sample frequency shift interval of the sample fermentation liquid under different fermentation conditions and different times, and the sample wave number of the Raman wave peak in each sample frequency shift interval. intensity and the sample PHA content value of the sample fermentation broth at different times. The sample PHA content value is determined based on the gas chromatography detection of the sample fermentation broth.
  • the sample frequency shift The PHA detection model is trained on a sample set constructed from the interval, the sample wave number intensity and the sample PHA content value of the sample fermentation broth, so that when the target frequency shift interval and the target wave number intensity are input, the PHA detection model can output the target PHA content value, that is, PHA content detection value.
  • the PHA detection model can also be based on different fermentation conditions, the sample wave number intensity of the Raman wave peak of the sample fermentation liquid at different times, and the sample PHA content value of the sample fermentation liquid at different times. Obtained by training, in such an embodiment, the sample PHA content value can also be determined based on gas chromatography detection of the sample fermentation broth. Correspondingly, the sample wave number intensity and the sample PHA content of the sample fermentation broth are The PHA detection model is trained on the sample set constructed by the values, so that the PHA detection model can output the target PHA content value, that is, the PHA content detection value, only when the target wavenumber intensity is input.
  • the different fermentation conditions include different fermentation containers, or different strains corresponding to different monomers of PHA, or different fermentation substrates, or strains with different activities.
  • the different fermentation containers include stainless steel fermentation tanks, glass fermentation tanks, and plastic fermentation tanks. Tanks, etc.
  • the different strains include strains corresponding to PHA of different monomers
  • the PHA of different monomers include poly- ⁇ -hydroxybutyrate PHB, 3-hydroxybutyrate and 3-hydroxyvalerate copolymer PHBV , 3-hydroxybutyric acid and 3-hydroxycaproic acid copolyester PHBHHx, poly-3-hydroxybutyrate-4-hydroxybutyrate P34HB.
  • the sample PHA content value is determined based on gas chromatography detection of the sample fermentation broth.
  • the gas chromatography detection method for PHA content can be determined in the following way: First, take 10 mL The fermentation broth is placed in a weighed 15 ml centrifuge tube, and 10 mL of ethanol is added and centrifuged. The conditions for the centrifugal treatment are to run at a speed of 10,000 rpm for 5 minutes. After centrifugation, 20 mL of ethanol is added to wash the cells. Centrifuge under the same conditions. After centrifugation, discard the supernatant and dry it in an oven at 65°C.
  • the preparation method of the esterification liquid is: weigh 0.5g of benzoic acid and add it to a methanol reagent bottle containing 485ml, slowly add 15ml of concentrated sulfuric acid into the methanol reagent bottle, mix and complete the preparation of the esterification liquid; finally, Add 1 mL of ultrapure water and vortex to perform extraction. Let it stand for 30 to 60 minutes to allow stratification. Remove the lower organic phase for gas chromatography analysis. The conditions for gas chromatography analysis must be met: the injection volume is 1 ⁇ L. The column flow rate is 35mL/min, the column temperature is 240°C, the flow rate is 23.4cm/s, the purge flow rate is 3mL/min, and the split ratio is 39.
  • This application adopts a method for real-time detection of PHA based on Raman spectroscopy.
  • the detection method can detect PHA in real time. , significantly shortening the detection time and greatly improving the detection throughput.
  • the fermentation broth samples do not need to be treated with chemical reagents during the entire process, which can significantly reduce the detection cost of PHA.
  • the detection process does not require sampling, and can eliminate changes in fermentation volume caused by sampling, thus interference with fermentation.
  • This application provides a method, device, system, and equipment for detecting polyhydroxyalkanoate content.
  • a detection model for polyhydroxyalkanoate PHA content By inputting Raman spectrum information collected from the fermentation broth to be detected into a detection model for polyhydroxyalkanoate PHA content, Obtain the PHA content value in the fermentation broth to be detected; since the PHA content detection model is trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the sample PHA content values under different fermentation conditions, This makes the detection of the PHA content value finally obtained accurate.
  • This application can overcome the technical problems of being unable to detect in real time due to the complex components and long fermentation cycle of PHA during fermentation. It can effectively detect and detect through the constructed polyhydroxyalkanoate PHA detection model. As a result, non-destructive, efficient and accurate detection of PHA in fermentation broth was achieved.
  • FIG2 is a schematic diagram of a process for processing the Raman spectrum information provided by the present application.
  • the detection model based on the PHA content processes the Raman spectrum information, including:
  • the PHA content value in the fermentation broth to be detected is output.
  • the application first divides the preset band intervals to obtain the frequency shift intervals of all fingerprint samples, and maps the frequency shift interval where the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is to the corresponding frequency shift interval in the detection model.
  • the frequency shift interval of the fingerprint sample For example, in a coordinate system with the frequency shift interval as the X-axis and the wave number intensity as the Y-axis, the preset band interval can be divided into intervals from 1800 to 1600 along the X-axis direction. , the interval from 1600 to 1400, the interval from 1400 to 1200, the interval from 1200 to 1000 and the interval from 1000 to 800.
  • the wave number intensity range can be set sequentially along the Y-axis direction from 0 to 120000.
  • step 1032 based on the corresponding fingerprint sample frequency shift interval, the wave number intensity of the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is determined in the frequency shift interval, and the PHA in the fermentation broth to be detected is The frequency shift interval where the corresponding Raman wave peak is located is mapped to the corresponding frequency shift interval of the fingerprint sample in the detection model to determine the frequency shift interval of the fingerprint sample associated with the PHA in the fermentation broth to be detected, and then determine the frequency shift interval of the fingerprint sample to be detected The wave number intensity of the Raman wave peak corresponding to the PHA in the fermentation broth in the frequency shift interval. Specifically, the wave number intensity in the frequency shift interval of the fingerprint sample is determined based on the coordinate position of the peak point of the Raman wave peak.
  • the wave number intensity will be determined by the corresponding position of the Y-axis of the wave peak point, for example, determine the target frequency shift interval of the Raman wave peak between 1600 and 1400, and then, according to the coordinates of the peak point of the Raman wave peak Position, determine the wave number intensity value to be 100000.
  • step 1033 based on the quantitative relationship between the wave number intensity and the PHA content, the PHA content value in the fermentation broth to be detected is output.
  • the quantitative relationship can be as shown in the above formula (1), where Y represents the content value of PHA and X represents the wave number intensity.
  • This application inputs the wave number intensity into the above-mentioned quantitative relationship, and then calculates and outputs the PHA content value in the fermentation broth to be detected.
  • collecting the target Raman spectrum information of the fermentation broth to be detected includes:
  • the probe of the Raman spectrum signal detector is immersed in the fermentation broth to be detected during the acquisition stage of target Raman spectrum information.
  • the probe of the Raman spectrum signal detector is a device that can obtain target Raman spectrum information in the fermentation liquid to be detected by immersing it in the fermentation liquid to be detected. It can be connected to the signal detector via an optical fiber to obtain target Raman spectrum information.
  • the present application adopts a technical solution for online real-time monitoring of target Raman spectrum information in the fermentation liquid to be detected, so as to efficiently and quickly analyze the wavenumber intensity of the Raman peak in the target Raman spectrum information.
  • the collection of target Raman spectrum information of the fermentation broth to be detected includes:
  • Each target fermentation broth has the same volume.
  • the preset number of parts can be 3 parts, 5 parts or even more.
  • This application can extract 5 mL of PHA fermentation liquid sample from the corresponding fermentation liquid carrying device at any time during the fermentation liquid fermentation.
  • the fermentation liquid carrying device It can be a fermentation tank. Shake for 2 minutes and mix evenly. Then add it to the detection tank and immerse the probe. Collect the fermentation broth Raman signal from the fermentation broth sample. During the collection process, the fermentation broth will be irradiated with an excitation light source. Optionally, set the excitation wavelength to 785 nm, the integration time to 5 s, and randomly scan the sample 5 times.
  • the extracted target fermentation broth to be detected is input into the Raman spectrum signal detector for direct detection.
  • the target Raman spectrum information if 5 target fermentation broths to be detected are obtained at the same time, then 5 target Raman spectrum information corresponding to the target fermentation broth to be detected are determined, among which, in order to ensure the accuracy of the detection results , the volume of each target fermentation broth is the same.
  • all the Raman spectrum information to be detected is averaged to obtain target Raman spectrum information.
  • all the Raman spectrum information to be detected is processed in the form of average processing, so as to use the average processing result as the target Raman spectrum information.
  • this application averages all the Raman spectrum information to be detected to obtain the initial Raman spectrum information, uses first-order derivation to process the initial Raman spectrum information, and obtains the denoised Raman spectrum information.
  • the spectral information is determined as the target Raman spectrum information within the preset band interval intercepted from the denoised Raman spectrum information.
  • Raman spectrum information to be detected is not the target Raman spectrum information, but the initial Raman spectrum information, which is the Raman spectrum information that has not been subjected to de-noising and truncation processing.
  • select PHA fermentation broths of different fermentation times and use the initial fermentation culture medium as a blank control, use Raman spectroscopy detection equipment for detection, and collect PHA fermentation broths and blank controls of different fermentation times through incident laser light sources of specific wavelengths.
  • the Raman spectrum information was used to determine the characteristic Raman peaks of PHA.
  • the collected Raman spectra were preprocessed, and the collected Raman spectra of PHA fermentation broth were analyzed in the full band. , and successively remove the fluorescence signal through standard normal transformation, derivation, and baseline correction.
  • this application can remove the fluorescence signal from the measured Raman spectrum data of the PHA fermentation broth through first-order derivation, The baseline is calibrated simultaneously and all data are normalized.
  • preprocessing methods include but are not limited to convolution smoothing, baseline correction, multivariate scattering correction, orthogonal signal correction, standard normal transformation, normalization, and Gaussian filtering. , median filtering processing.
  • the Raman spectrum information within the preset band interval intercepted from the denoised Raman spectrum information is determined as the target Raman spectrum information.
  • the present application When analyzing Mann spectral data, the fingerprint region spectrum with wave number from 800cm-1 to 1800cm-1 is selected.
  • the PHA detection model before obtaining the PHA content detection value output by the PHA detection model, it also includes starting a timer from the moment when the sample fermentation liquid is inoculated and fermented, and collecting samples of the sample fermentation liquid at preset intervals.
  • Raman spectrum information based on the sample Raman spectrum information at each moment, obtain the sample frequency shift interval of the sample fermentation broth at each moment and the sample wave number intensity of the Raman wave peak within each sample frequency shift interval; based on gas chromatography detection
  • the sample fermentation broth is tested at each moment to determine the sample PHA content value corresponding to the sample fermentation broth at each moment; according to the sample frequency shift interval of the sample fermentation broth at each moment, the Raman wave peak in each sample frequency shift interval
  • the sample wave number intensity and the sample PHA content value corresponding to the sample fermentation broth at each moment are used to construct a sample data set; the sample data set is divided according to a preset ratio and a sample training set and a sample test set are determined to determine the sample training set according to the sample training set And the
  • 5 mL of PHA fermentation broth sample was taken from the fermentation tank, shaken for 2 min to mix, and then added to the detection cell.
  • the Raman probe was immersed in the fermentation broth.
  • the fermentation broth Raman signal was collected from the sample, and the Raman spectrum acquisition parameters were: excitation wavelength 785 nm, integration time 5 s, random scanning of the sample 5 times, and taking the average spectrum of 5 times to represent the sample spectral information.
  • 3 parallel samples of 10 mL each are taken every 2 hours, and the samples taken at each time point are added to the detection pool.
  • the Raman probe is immersed in the fermentation broth sample to collect signals.
  • the sample frequency shift interval of the sample fermentation broth at each moment and the sample wavenumber intensity of the Raman wave peak within each sample frequency shift interval are obtained.
  • the above sample frequency shift interval and the sample wavenumber intensity are obtained.
  • the strong technical solution is to remove the background noise (such as fluorescence signal) from the Raman spectrum information of the collected PHA fermentation broth sample, correct the baseline, normalize all data, and analyze the Raman spectrum of the sample.
  • the fingerprint area map of 800 to 1800cm-1 is selected, the Raman spectrum information of the sample is mapped to the preset band interval, and the frequency shift interval where the sample Raman wave peak is located is determined as the sample frequency shift interval.
  • the sample wave number intensity of the Raman wave peak in the sample frequency shift interval is traversed through all sample Raman spectral information to obtain the sample frequency shift interval of the sample fermentation liquid at each moment and the Raman value in each sample frequency shift interval.
  • the sample wavenumber intensity of the wave crest is selected from the Raman spectrum information of the sample is mapped to the preset band interval, and the frequency shift interval where the sample Raman wave peak is located.
  • the sample fermentation broth at each moment is detected to determine the sample PHA content value corresponding to the sample fermentation broth at each moment.
  • This application will determine the corresponding sample fermentation broth at each moment based on the gas chromatography detection PHA content method.
  • the sample PHA content value is determined.
  • the preparation method of the esterification liquid was as follows: weigh 0.5g benzoic acid and add it to a methanol reagent bottle containing 485ml. Take 15ml concentrated sulfuric acid and slowly add it to the methanol reagent bottle. After mixing, Complete the preparation of the esterification liquid; finally, add 1 mL of ultrapure water, vortex, extract, and let stand for 30 to 60 minutes to cause stratification, and then remove the lower organic phase for gas chromatography analysis. The gas chromatography analysis The conditions need to be met: injection volume 1 ⁇ L, column flow 35mL/min, column temperature 240°C, flow rate 23.4cm/s, purge flow 3mL/min and split ratio 39.
  • the sample data is constructed based on the sample frequency shift interval of the sample fermentation liquid at each time, the sample wave number intensity of the Raman wave peak within each sample frequency shift interval, and the sample PHA content value corresponding to the sample fermentation liquid at each time.
  • Set use the sample frequency shift interval, sample wave number intensity and sample PHA content value at each moment as labels to form a sample data set, and then determine all sample data sets at all times based on the sample fermentation broth at all times.
  • the sample training set and the sample test set after dividing the sample data set according to a preset ratio, so as to construct a PHA detection model based on the sample training set and the sample test set.
  • the preset ratio may be 7:3. , 8:2 or other ratios.
  • this application will process the results of Raman spectroscopy and gas chromatography measurements, for example, determine a total of 34 sample data sets, as labels to establish Partial Least Squares regression (Partial Least Squares Regression (PLSR) model, wherein the sample data set is divided into a sample training set and a sample test set, and the proportions of the sample training set and the sample test set are determined to be 70% and 30% respectively, so that according to the sample training set And the sample test set is used to establish the PHA detection model.
  • Partial Least Squares regression Partial Least Squares Regression
  • This application divides the processed Raman spectrum data set into a training set and a test set according to different proportions, and establishes a PHA detection model based on Raman spectrum.
  • the PHA results at different fermentation times detected by gas chromatography are used as the real values.
  • Partial least squares regression models the Raman spectrum data of the training set to establish a quantitative relationship between wave number intensity and PHA characteristics.
  • This application brings the Raman spectrum data of the test set into the detection model, and based on the detection results, the detection The model is modified to improve the generalization ability of the model, and the PHA characteristic is PHA content or PHA concentration.
  • preprocess the Raman spectrum information to obtain denoised Raman spectrum information.
  • the preprocessing method includes convolution smoothing processing, baseline correction processing, multivariate scattering correction processing, orthogonal signal correction processing, At least one of standard normal transformation processing, normalization processing, Gaussian filtering processing, and median filtering processing.
  • Figure 3 is the second flow chart of the method for detecting the content of polyhydroxyalkanoates provided by the present application.
  • the present application collects standard samples according to Raman spectroscopy, then intercepts the spectrum by wavelength range, and preprocesses the spectrum.
  • the preprocessing includes removing the baseline, taking derivatives, and normalizing.
  • the detection model is trained to determine the accuracy of the model. When the accuracy of the model is insufficient, the step of returning to the selected wavelength range to intercept the spectrum is performed. When the accuracy of the model reaches the preset accuracy, the detection model is determined. After the detection model is determined, the spectrum collects data in real time, and the Raman data is preprocessed accordingly, and imported into the detection model to obtain the corresponding detection results.
  • constructing a model for detecting PHA based on Raman spectroscopy is implemented in the following manner:
  • Raman spectrum collection of PHA fermentation broth At different fermentation times, take 5mL of PHA fermentation broth sample from the fermentation tank, shake for 2 minutes and mix well, add it to the detection tank, and immerse the Raman probe into the fermentation broth sample to collect Fermentation broth Raman signal, Raman spectrum collection parameters are: excitation wavelength 785nm, integration time 5s, randomly scan the sample 5 times, and take the average spectrum of 5 times to represent the sample spectral information.
  • Preprocessing of Raman spectrum data remove the fluorescence signal of the measured Raman spectrum data of PHA fermentation broth through first-order derivation, calibrate the baseline at the same time, and normalize all data.
  • the fingerprint area spectrum from 800 to 1800cm-1 was selected.
  • Detection PHA model construction based on Raman spectrum Use the processed results of Raman spectrum and gas chromatography measurement as labels to establish a PLSR model.
  • the Raman spectrum data set is divided into a training set and a test set, and the ratio of the training set to the test set 70% and 30%, establish model parameters as shown in Table 1:
  • R2 is the coefficient of determination (R-square). The closer it is to 1, the stronger the ability to explain the regression equation, and the better the model fits the data.
  • the established detection model is then used to verify and analyze the Raman spectrum data of the fermentation time point that is not used as a label.
  • the Raman spectrum data processing method of the fermentation time point that is not used as a label is the same as the Raman spectrum data processing method of the label.
  • the model detection results are shown in Table 2.
  • the gas chromatography detection PHA results have a good correlation with the Raman spectroscopy modeling detection results, and the error value is within 4%.
  • a fermentation tank is optionally used as the fermentation carrier of the fermentation liquid in sample sampling.
  • the fermentation tanks include but are not limited to stainless steel fermentation tanks, glass fermentation tanks, and plastic fermentation tanks.
  • the Raman spectrum acquisition conditions are: using a wavelength of 785 nm Laser, collection range is 300-3200cm-1, resolution is 5cm-1, laser power is 500mW, collection frequency is 10-30s/time, cumulative collection is 1-10 times, optionally, collection is 5 times.
  • this application uses offline sampling from glass fermentation tanks to detect PHA content, including the following processes:
  • the Raman spectrum collection parameters are: excitation wavelength 785nm, integration time 5s, randomly scan the sample 5 times, and take the average spectrum of 5 times to represent the sample spectral information.
  • the processing of Raman spectrum data remove the background noise in the collected Raman spectrum data of PHA fermentation broth, such as the removal of fluorescence signal, correct the baseline, and normalize all data.
  • the fingerprint area spectrum of 800 to 1800cm-1 was selected.
  • Raman spectrum modeling and detection analysis use the processed results of Raman spectrum and gas chromatography measurements as labels to establish a PLSR model.
  • the sample data set is divided into a training set and a test set, and the ratio of the training set and the test set is 70%. and 30%.
  • the established PLSR model was then used to verify and analyze the Raman spectrum data at other fermentation time points that were not used as labels.
  • the processing method of the Raman spectrum data at other fermentation time points that were not used as labels was consistent with the method used as label Raman spectrum data.
  • the model detection results are shown in Table 3. It can be seen that the gas chromatography detection PHA results have a good correlation with the Raman spectrum modeling detection results, and the error value is within 7%.
  • this application can take 1 spectral data to represent the sample spectral information, or 5 spectral data to represent the sample spectral information.
  • this application uses stainless steel fermentation
  • the correlation analysis between the PHA content results detected by gas chromatography and the Raman spectrum modeling detection results includes:
  • the Raman spectrum probe is immersed in a 75L stainless steel fermentation tank. After sterilization and inoculation, the Raman signal of the fermentation broth is collected in real time from the beginning of fermentation. Among them, the PHA fermentation broth It is poly-3-hydroxybutyrate-co-3-hydroxyhexanoate (PHBHHx).
  • the Raman spectrum collection parameters are: excitation wavelength 785nm, integration time 10s, randomly scan the sample once, and take 1 spectral data to represent the sample spectrum. information.
  • process the Raman spectrum conduct full-band analysis on the collected Raman spectrum of PHA fermentation broth, and remove the fluorescence signal through standard normal transformation and baseline correction.
  • Raman spectrum modeling and detection analysis use the processed results of Raman spectrum and gas chromatography measurements as labels to establish a PLSR model.
  • the sample data set is divided into a training set and a test set, and the ratio of the training set and the test set is 70%. and 30%, and then use the established PLSR model to verify and analyze the Raman spectrum data of other fermentation time points that are not used as labels.
  • the processing method of Raman spectrum data of other fermentation time points that are not used as labels is the same as that of Raman spectra as labels.
  • the data are in the same manner, and the detection results of the model are shown in Table 4.
  • the PHA content detection results by gas chromatography have a good correlation with the Raman spectroscopy modeling detection results, and the error value is within 8%.
  • this application can also use 5 spectral data to represent sample spectral information.
  • this application uses stainless steel fermentation tanks for online detection to determine that 5 spectral data represent
  • the correlation analysis between the PHA content detection results by gas chromatography and the Raman spectrum modeling detection results includes:
  • the Raman spectrum probe is immersed in a 75L stainless steel fermentation tank. After sterilization and inoculation, the Raman signal of the fermentation broth is collected in real time from the beginning of fermentation. Among them, the PHA fermentation broth The product is PHBHHx.
  • the Raman spectrum collection parameters are: excitation wavelength 785nm, integration time 10s, randomly scan the sample 5 times, and take the average spectrum of 5 times to represent the sample spectral information.
  • the Raman spectrum was processed: the collected Raman spectrum of PHA fermentation broth was analyzed in the whole band, and the fluorescence signal was removed by standard normal transformation and baseline correction.
  • Raman spectrum modeling and detection analysis use the processed results of Raman spectrum and gas chromatography measurements as labels to establish a PLSR model.
  • the sample data set is divided into a training set and a test set. The ratio of the training set and the test set is 70%. and 30%.
  • the established PLSR model was then used to verify and analyze the Raman spectrum data at other fermentation time points that were not used as labels.
  • the processing method of the Raman spectrum data at other fermentation time points that were not used as labels was consistent with the method used as label Raman spectrum data.
  • the model detection results are shown in Table 5.
  • the PHA content detection results by gas chromatography have a good correlation with the Raman spectroscopy modeling detection results, and the error value is within 5%.
  • the initial Raman spectrum information without first-order derivation processing is provided.
  • the technical solution is to construct a detection model based on the Raman spectrum that has not undergone first-order derivation processing.
  • the first-order derivation is used to process the initial Raman spectrum information, the denoised Raman spectrum can be obtained.
  • Man spectral information in order to construct a detection model based on the Raman spectrum processed by first-order derivation, specifically includes:
  • the Raman spectrum of PHA fermentation broth was collected in real time: the Raman spectrum probe was immersed in a 75L stainless steel fermenter. After sterilization and inoculation, the Raman signal of the fermentation broth was collected in real time from the beginning of fermentation.
  • the PHA fermentation broth product was PHBHHx.
  • the Raman spectrum collection parameters were: excitation wavelength 785nm, integration time 10s, random scanning of the sample 5 times, and the average spectrum of 5 times represented the sample spectral information.
  • process the Raman spectrum conduct full-band analysis on the collected Raman spectrum of PHA fermentation broth, and remove the fluorescence signal through standard normal transformation and baseline correction.
  • Raman spectrum modeling and detection analysis use the processed results of Raman spectrum and gas chromatography measurements as labels to establish a PLSR model.
  • the sample data set is divided into a training set and a test set. The ratio of the training set and the test set is 70%. and 30%, and then use the established PLSR model to verify and analyze the Raman spectrum data of other fermentation time points that are not used as labels.
  • the processing method of Raman spectrum data of other fermentation time points that are not used as labels is the same as that of Raman spectra as labels.
  • the data are in the same manner, and the model detection results are shown in Table 6.
  • the PHA content detection results by gas chromatography have a good correlation with the Raman spectroscopy modeling detection results, and the error value is within 3%.
  • this application will also verify the correlation between the PHA content results detected by gas chromatography and the Raman spectroscopy modeling detection results using different PHA concentrations, specifically including:
  • the Raman spectrum probe is immersed in a 75L stainless steel fermentation tank. After sterilization and inoculation, the Raman signal of the fermentation broth is collected in real time from the beginning of fermentation. Among them, the PHA fermentation broth The product is poly3-hydroxybutyrate PHB.
  • the Raman spectrum collection parameters are: excitation wavelength 785nm, integration time 10s, randomly scan the sample 5 times, and take the average spectrum of 5 times to represent the sample spectral information.
  • process the Raman spectrum perform full-band analysis on the collected Raman spectrum of PHA fermentation broth. After analysis, the fluorescence signal was removed through standard normal transformation, derivation, and baseline correction.
  • Raman spectrum modeling and detection analysis use the processed results of Raman spectrum and gas chromatography measurements as labels to establish a PLSR model.
  • the sample data set is divided into a training set and a test set. The ratio of the training set and the test set is 70%. and 30%, and then use the established PLSR model to detect and analyze other samples that are not labeled.
  • the model detection results are shown in Table 7. It can be seen that the PHA content detected by gas chromatography at different PHA concentrations has a good correlation with the Raman spectroscopic modeling detection results, and the error value is within 5%.
  • Figure 4 is a schematic structural diagram of the polyhydroxyalkanoate content detection system provided by this application, including;
  • Testing container 1 is used to provide a testing environment for the fermentation liquid
  • Probe 2 is used to immerse into the detection cell to collect Raman spectrum information
  • Optical fiber 3 used for signal transmission between the probe and the excitation light source, as well as signal transmission between the probe and the signal detector;
  • Excitation light source 4 used to provide detection light source for fermentation broth
  • Signal detector 5 used to convert optical signals into data signals
  • the polyhydroxyalkanoate content detection device can be used as a whole to analyze and process the collected Raman spectrum information, and then output the polyhydroxyalkanoate content in the fermentation broth to be tested, and optionally implement
  • the detection device for polyhydroxyalkanoate content can be further subdivided into:
  • Data collection unit 6 used to collect Raman spectrum information
  • the analysis and processing unit 7 is used to analyze and process the collected Raman spectrum information, and then output the polyhydroxyalkanoate content in the fermentation broth to be tested.
  • a device for detecting PHA in the fermentation broth based on Raman spectroscopy includes a detection container 1, the detection container is a detection pool, and the excitation
  • the light source 4 is connected to the probe 2 through the optical fiber 3
  • the probe 2 is connected to the signal detector 5 through the optical fiber 3.
  • the signal detector 5 transmits the collected signals to the data collection unit 6 and the analysis and processing unit 7 through the data line or wireless network.
  • the polyhydroxyalkanoate content detection system can be used in two scenarios: online detection and offline detection.
  • the probe 2 is immersed in the fermentation tank and collects real-time data during the fermentation process.
  • the component signal inside the fermentation broth, then the signal detector 5 transmits the detected Raman spectrum signal to the data collection unit 6, and the analysis and processing unit 7 is used to further process and analyze the Raman signal collected by the data collection unit 6 to obtain the PHA ingredient information;
  • the offline scenario first take the fermentation broth sample from the fermentation tank, and then place the sample in the detection container 1.
  • the inside of the detection container 1 is completely in a light-proof environment.
  • the probe 2 is immersed in the fermentation broth sample to collect the fermentation broth.
  • the signal detector 5 then transmits the detected Raman spectrum signal to the data collection unit 6, and the analysis and processing unit 7 is used to further process and analyze the Raman signal collected by the data collection unit 6 to obtain the composition information of PHA.
  • the present application also includes a memory and a program or instruction stored on the memory and executable on the analysis processing unit 7 .
  • the program or instruction is executed by the analysis processing unit 7 .
  • a method for detecting polyhydroxyalkanoate content which method includes: collecting Raman spectrum information of fermentation broth to be detected; inputting the Raman spectrum information into a detection model of polyhydroxyalkanoate PHA content; based on the PHA content
  • the detection model processes the Raman spectrum information and outputs the PHA content value in the fermentation broth to be detected;
  • the PHA content detection model includes a quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value; the quantitative relationship The relationship is a quantitative relationship trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the PHA content value of the sample under different fermentation conditions; the sample PHA content value is based on the gas chromatography detection of the sample fermentation broth. definite.
  • This application provides a method, device, system, and equipment for detecting polyhydroxyalkanoate content.
  • a detection model for polyhydroxyalkanoate PHA content By inputting Raman spectrum information collected from the fermentation broth to be detected into a detection model for polyhydroxyalkanoate PHA content, Obtain the PHA content value in the fermentation broth to be detected; since the PHA content detection model is trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the sample PHA content values under different fermentation conditions, This makes the detection of the PHA content value finally obtained accurate.
  • This application can overcome the technical problems of being unable to detect in real time due to the complex components and long fermentation cycle of PHA during fermentation. It can effectively detect and detect through the constructed polyhydroxyalkanoate PHA detection model. As a result, non-destructive, efficient and accurate detection of PHA in fermentation broth was achieved.
  • FIG. 5 is a schematic structural diagram of a detection device for polyhydroxyalkanoate content provided by this application.
  • This application also provides a detection device for polyhydroxyalkanoate content, including a collection unit 51: used to collect fermentation broth to be detected For Raman spectrum information, the working principle of the collection unit 51 can be referred to the aforementioned step 101, which will not be described again here.
  • the polyhydroxyalkanoate content detection device also includes an input unit 52: used to input the Raman spectrum information into the polyhydroxyalkanoate PHA content detection model.
  • the working principle of the input unit 52 can be referred to the previous steps. 102, which will not be discussed in detail here.
  • the polyhydroxyalkanoate content detection device also includes a processing unit 53: used to process the Raman spectrum information based on the detection model of the PHA content, and output the PHA content value in the fermentation broth to be detected, so
  • a processing unit 53 used to process the Raman spectrum information based on the detection model of the PHA content, and output the PHA content value in the fermentation broth to be detected, so
  • the processing unit 53 For the working principle of the processing unit 53, reference can be made to the aforementioned step 103, which will not be described again here.
  • the detection model of PHA content includes Raman spectrum information of fermentation broth and determination of PHA content value. Quantity relationship;
  • the quantitative relationship is a quantitative relationship obtained by training based on the Raman spectrum information of sample fermentation broth under different fermentation conditions and the PHA content values of samples under different fermentation conditions;
  • the PHA content value of the sample is determined based on gas chromatography detection of the fermentation broth of the sample.
  • the processing unit also includes a mapping subunit 531: used to map the frequency shift interval where the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is located to the corresponding frequency shift interval of the fingerprint sample in the detection model.
  • a mapping subunit 531 used to map the frequency shift interval where the Raman wave peak corresponding to the PHA in the fermentation broth to be detected is located to the corresponding frequency shift interval of the fingerprint sample in the detection model.
  • the working principle of the mapping subunit 531 can be referred to the aforementioned step 1031, which will not be described again here.
  • the processing unit further includes a determining subunit 532: for determining the wavenumber intensity of the Raman peak corresponding to the PHA in the fermentation liquid to be detected in the frequency shift interval based on the corresponding fingerprint sample frequency shift interval.
  • the working principle of the determining subunit 532 can refer to the aforementioned step 1032 and will not be described in detail here.
  • the processing unit also includes an output subunit 533: used to output the PHA content value in the fermentation broth to be detected through the quantitative relationship between the wave number intensity and the PHA content.
  • the working principle of the output subunit 533 can be Refer to the aforementioned step 1033, which will not be described again here.
  • the detection device also includes:
  • Preprocessing unit 54 used to preprocess the Raman spectrum information to obtain denoised Raman spectrum information.
  • the preprocessing methods include convolution smoothing processing, baseline correction processing, multivariate scattering correction processing, and orthogonal signal processing. At least one of correction processing, standard normal transformation processing, normalization processing, Gaussian filtering processing, and median filtering processing.
  • This application provides a method, device, system, and equipment for detecting polyhydroxyalkanoate content.
  • a detection model for polyhydroxyalkanoate PHA content By inputting Raman spectrum information collected from the fermentation broth to be detected into a detection model for polyhydroxyalkanoate PHA content, Obtain the PHA content value in the fermentation broth to be detected; since the PHA content detection model is trained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the sample PHA content values under different fermentation conditions, This makes the detection of the finally obtained PHA content value accurate.
  • This application can overcome the technical problems of inability to detect in real time due to the complex components and long fermentation cycle of PHA during fermentation, and effectively detect and detect the polyhydroxyalkanoate PHA detection model constructed As a result, non-destructive, efficient and accurate detection of PHA in fermentation broth is achieved.
  • Figure 6 is a schematic structural diagram of an electronic device provided by this application.
  • the electronic device may include: a processor (processor) 610, a communication interface (Communications Interface) 620, a memory (memory) 630, and a communication bus 640.
  • the processor 610, the communication interface 620, and the memory 630 pass through The communication bus 640 completes mutual communication.
  • the processor 610 can call the logic instructions in the memory 630 to execute a method for detecting polyhydroxyalkanoate content.
  • the method includes: collecting Raman spectrum information of the fermentation broth to be detected; inputting the Raman spectrum information to the polyhydroxyalkanoate.
  • a detection model for acid ester PHA content processes the Raman spectrum information and outputs the PHA content value in the fermentation broth to be detected; the detection model for the PHA content includes the Raman spectrum of the fermentation broth.
  • the detection is determined by testing the sample fermentation broth.
  • the above-mentioned logical instructions in the memory 630 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.
  • the technical solution of the present application essentially contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium and includes a number of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program code. .
  • the present application also provides a computer program product.
  • the computer program product includes a computer program.
  • the computer program can be stored on a non-transitory computer-readable storage medium.
  • the computer program can Execute a method for detecting polyhydroxyalkanoate content provided by each of the above methods.
  • the method includes: collecting Raman spectrum information of the fermentation broth to be detected; inputting the Raman spectrum information into the PHA content of polyhydroxyalkanoate.
  • the detection model based on the PHA content processes the Raman spectrum information and outputs the PHA content value in the fermentation broth to be detected;
  • the detection model of the PHA content includes the Raman spectrum information of the fermentation broth and PHA
  • the quantitative relationship of the content value is a quantitative relationship obtained based on the Raman spectrum information of the sample fermentation broth under different fermentation conditions and the sample PHA content value training;
  • the sample PHA content value is based on the gas chromatography detection of the sample Determined by fermentation broth testing.
  • the present application also provides a non-transitory computer-readable storage medium on which a computer program is stored.
  • the computer program is implemented when executed by a processor to perform the above methods to provide a detection method for polyhydroxyalkanoate content.
  • the method includes: collecting Raman spectrum information of the fermentation broth to be detected; inputting the Raman spectrum information into a detection model of polyhydroxyalkanoate PHA content; processing the Raman spectrum information based on the detection model of the PHA content , and output the PHA content value in the fermentation broth to be detected;
  • the detection model of the PHA content includes the quantitative relationship between the Raman spectrum information of the fermentation broth and the PHA content value; the quantitative relationship is based on the sample fermentation under different fermentation conditions.
  • the quantitative relationship obtained by training the Raman spectrum information of the liquid and the PHA content value of the sample; the PHA content value of the sample is determined based on the gas chromatography detection of the sample fermentation liquid.
  • the device embodiments described above are only illustrative.
  • the units described as separate components may or may not be physically separated.
  • the components shown as units may or may not be physical units, that is, they may be located in One location, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.
  • each embodiment can be implemented by software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., including a number of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments.

Landscapes

  • Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)

Abstract

一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备。聚羟基脂肪酸酯含量的检测方法包括：采集待检测发酵液的拉曼光谱信息(101)；输入拉曼光谱信息至PHA含量的检测模型(102)；基于PHA含量的检测模型处理拉曼光谱信息，并输出PHA含量值(103)；PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息及样本PHA含量值训练得到的定量关系。

Description

聚羟基脂肪酸酯含量的检测方法、装置、系统、设备
相关申请的交叉引用
本申请要求于2022年09月23日提交的申请号为202211167876.6，发明名称为“聚羟基脂肪酸酯含量的检测方法、装置、系统、设备”的中国专利申请的优先权，其通过引用方式全部并入本申请。
技术领域
本申请涉及生物检测领域，尤其涉及一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备。
背景技术
聚羟基脂肪酸酯(Polyhydroxyalkanoates，PHA)主要通过微生物发酵的方式来制备，为了有效提高生产效率和降低生产成本，监测发酵过程中的PHA含量、浓度、纯度等指标对发酵过程控制至关重要。
传统的PHA的检测方法大多需要经过取样洗涤、离心干燥、甲醇-氯仿消解、有机萃取等处理后，再经过气相色谱法来进行检测，这种检测方法虽然能获得细胞内PHA的含量，但耗时长，样品前处理也非常麻烦，且不同发酵条件导致的PHA发酵过程具有组分复杂、发酵周期长、无法实时检测的特殊性，传统的PHA的检测方法无法实时反映发酵液中产物的变化。
发明内容
本申请提供一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备，用以解决现有PHA检测技术存在的滞后性、成本高、通量低的技术缺陷，本申请能够通过对发酵液中的PHA拉曼光谱进行建模，通过建立的分析模型有效检测检测结果，实现对发酵液中PHA的无损、高效、准确检测。
第一方面，本申请提供了一种聚羟基脂肪酸酯含量的检测方法，包括：
采集待检测发酵液的拉曼光谱信息；
输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；
基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；
所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；
所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不 同发酵条件下的样本PHA含量值训练得到的定量关系；
所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
根据本申请提供的聚羟基脂肪酸酯含量的检测方法，所述拉曼光谱信息包括所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间以及在所述频移区间中所述拉曼波峰的波数强度；
所述发酵液的拉曼光谱信息与PHA含量值的定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息中的所述波数强度以及PHA含量值通过训练得到的定量关系。
根据本申请提供的聚羟基脂肪酸酯含量的检测方法，所述基于所述PHA含量的检测模型处理所述拉曼光谱信息，包括：
将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间；
基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度；
通过所述波数强度与发酵液中PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
根据本申请提供的聚羟基脂肪酸酯含量的检测方法，在采集待检测发酵液的拉曼光谱信息之后，还包括：
预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息；
所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
根据本申请提供的聚羟基脂肪酸酯含量的检测方法，所述不同发酵条件包括：
不同的发酵容器；
或，不同单体的PHA对应的不同菌株；
或，不同发酵基质；
或，不同活性的菌株。
第二方面，还提供了一种聚羟基脂肪酸酯含量的检测装置，包括：
采集单元：用于采集待检测发酵液的拉曼光谱信息；
输入单元：用于输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；
处理单元：用于基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；
所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；
所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；
所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
根据本申请提供的聚羟基脂肪酸酯含量的检测装置，所述处理单元还包括：
映射子单元：用于将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间；
确定子单元：用于基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度；
输出子单元：用于通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
根据本申请提供的聚羟基脂肪酸酯含量的检测装置，所述检测装置还包括：
预处理单元：用于预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息；
所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
第三方面，还提供了一种聚羟基脂肪酸酯含量的检测系统，包括：
检测容器，用于为发酵液提供检测环境；
探头，用于浸入至检测池中采集拉曼光谱信息；
光纤，用于探头与激发光源的信号传输，以及探头与信号检测器的信号传输；
激发光源，用于为发酵液提供检测光源；
信号检测器，用于将光信号转换为数据信号；
还包括所述聚羟基脂肪酸酯含量的检测装置，所述检测装置用于对所收集的拉曼光谱信息分析处理后，输出待测发酵液中的聚羟基脂肪酸酯含量。
第四方面，还提供了一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现所述的聚羟基脂肪酸酯含量的检测方法。
本申请提供了一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备，通过将从待检测发酵液中采集到的拉曼光谱信息输入至聚羟基脂肪酸酯PHA含量的检测模型，获取所述待检测发酵液中的PHA含量值；由于所述PHA含量的检测模型是根据不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的，以使得最终获取的PHA含量值检测准确，本申请能够克服因PHA发酵时的组分复杂、发酵周期长而导致无法实时检测的技术问题，通过所构建的聚羟基脂肪酸酯PHA检测模型有效检测检测结果，可将检测误差值控制在8％以内，尤其是可以控制在5％以内，实现对发酵液中PHA的无损、高效、准确检测。
附图说明
为了更清楚地说明本申请或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出 创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本申请提供的聚羟基脂肪酸酯含量的检测方法的流程示意图之一；
图2是本申请提供的处理所述拉曼光谱信息的流程示意图；
图3是本申请提供的聚羟基脂肪酸酯含量的检测方法的流程示意图之二；
图4是本申请提供的聚羟基脂肪酸酯含量的检测系统的结构示意图；
图5是本申请提供的聚羟基脂肪酸酯含量的检测装置的结构示意图；
图6是本申请提供的电子设备的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚，下面将结合本申请中的附图，对本申请中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
聚羟基脂肪酸酯(PHA)是自然界中由多种微生物产生高分子聚酯化合物，作为一种能源储备物质广泛存在于微生物细胞中，由于其优异的生物可降解性和可塑性，在农业、食品、医疗和制药工业中具有广泛的应用前景，本申请旨在提供一种能够实时在线检测发酵液中PHA的方法，以克服现有PHA检测技术所存在的滞后性、成本高、通量低等技术问题。
图1是本申请提供的聚羟基脂肪酸酯含量的检测方法的流程示意图之一，本申请提供了一种聚羟基脂肪酸酯含量的检测方法，包括：
采集待检测发酵液的拉曼光谱信息；
输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；
基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；
所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；
所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；
所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
在步骤101中，所述拉曼光谱信息包括所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间以及在所述频移区间中所述拉曼波峰的波数强度，本申请采集待检测发酵液的拉曼光谱信息，从所述拉曼光谱信息中提取出拉曼波峰所在频移区间以及在所述频移区间中所述拉曼波峰的波数强度，所述待检测发酵液可以是从检测池中提取出的发酵液，也可以通过探头直接渗入待检测发酵液中采集所述待检测发酵液的拉曼光谱信息，拉曼光谱信息是一种散射光谱，是利用分子被激光照射后所出的散射光的波长差别来对物质进 行表征分析的技术，由于拉曼光谱检测具有快速、无损、灵敏度高等优点，被广泛应用于各个领域，尤其是在发酵领域，拉曼光谱可以及时地获取到当前反应的状态信息，而不需要经过取样、处理、再检测等环节，可极大增强检测的通量，同时还避免了由于体积变化对反应造成的干扰。
本领域技术人员理解，在采集待检测发酵液的拉曼光谱信息之前，为减少背景噪音的干扰，尤其是荧光信号的干扰，可选地对采集到的拉曼光谱进行预处理，映射预处理后的所述拉曼光谱信息至预设波段区间，确定出拉曼波峰，将所述拉曼波峰所在的频移区间确定为目标频移区间，确定所述拉曼波峰在所述频移区间中的目标波数强度。
在步骤102中，输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型，在一个可选地实施例中，向量化所述频移区间以及所述波数强度，并根据所述频移区间以及所述波数强度的向量化表示构建特征输入矩阵，将所述特征输入矩阵输入至聚羟基脂肪酸酯PHA检测模型，以获取所述PHA检测模型输出的PHA含量检测值。
在步骤103中，基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值，所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系。在输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型之前，还包括：基于不同发酵条件下，不同发酵时间的样本数据，按照预设比例确定样本训练集以及样本测试集，以根据所述样本训练集以及样本测试集构建PHA含量的检测模型。
所述PHA含量的检测模型是根据不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的，具体地：所述发酵液的拉曼光谱信息与PHA含量值的定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息中的所述波数强度以及PHA含量值通过训练得到的定量关系，如下：
式(1)中，Y代表PHA的含量值，X代表波数强度，采用偏最小二乘法回归(PLSR)对训练集的拉曼光谱数据进行建模，建立波数强度与PHA特性的定量关系。
在一个可选地实施例中，所述发酵液的拉曼光谱信息与PHA含量值的定量关系是所述波数强度以及PHA含量值的定量关系，即波数强度对应于发酵液中PHA含量。
作为本申请的一个可选地实施例，所述PHA检测模型是根据不同发酵条件、不同时刻下样本发酵液的每一样本频移区间、在每一样本频移区间内拉曼波峰的样本波数强度以及不同时刻下样本发酵液的样本PHA含量值训练得到的，所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的，在这样的实施例中，将样本频移区间、样本波数强度以及样本发酵液的样本PHA含量值所构建的样本集进行PHA检测模型的训练，以使得在输入目标频移区间以及目标波数强度的情况下，所述PHA检测模型能够输出目标 PHA含量值，即PHA含量检测值。
而作为本申请的另一个可选地实施例，所述PHA检测模型还可以根据不同发酵条件、不同时刻下样本发酵液拉曼波峰的样本波数强度以及不同时刻下样本发酵液的样本PHA含量值训练得到的，在这样的实施例中，所述样本PHA含量值同样可以是基于气相色谱检测对所述样本发酵液检测而确定的，相应地，将样本波数强度以及样本发酵液的样本PHA含量值所构建的样本集进行PHA检测模型的训练，以使得仅在输入目标波数强度的情况下，所述PHA检测模型能够输出目标PHA含量值，即PHA含量检测值。
所述不同发酵条件包括不同的发酵容器，或不同单体的PHA对应的不同菌株，或不同发酵基质，或不同活性的菌株，所述不同的发酵容器包括不锈钢发酵罐、玻璃发酵罐、塑料发酵罐等，所述不同菌株包括不同单体的PHA对应的菌株，不同的单体的PHA包括聚-β-羟丁酸PHB，3-羟基丁酸酯和3-羟基戊酸酯的共聚物PHBV、3-羟基丁酸与3-羟基己酸的共聚酯PHBHHx、聚-3-羟基丁酸酯-4-羟基丁酸酯P34HB。
所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的，在一个可选地实施例中，所述的气相色谱检测PHA含量方法可以通过如下方式确定：首先，取10mL发酵液置于称重后的15ml离心管中，加入10mL乙醇后采取离心处理，所述离心处理的条件为以每分钟10000转的速度运行5分钟，离心结束后加入20mL乙醇，洗涤菌体，以同样条件进行离心处理，离心后倒掉上清，在65℃的烘箱内烘干；然后，称取50mg烘干样品于试管中，加入2mL氯仿和2mL酯化液在100℃反应4小时，其中，酯化液的配置方法为：称取0.5g苯甲酸加入装有485ml的甲醇试剂瓶中，取15ml浓硫酸缓慢加入至甲醇试剂瓶中，混匀后完成酯化液的配制；最后，加入1mL超纯水涡旋振荡，进行萃取，静置30分钟至60分钟，使其产生分层，取下层有机相进行气相色谱分析，所述气相色谱分析的条件需满足：进样量1μL，色谱柱流量35mL/min，柱温240℃，流速23.4cm/s，吹扫流量3mL/min以及分流比39。
本申请采用基于拉曼光谱实时检测PHA的方法，利用PHA发酵不同时刻特征峰的变化，可以有效检测到PHA的变化，相比传统气相色谱检测技术，本申请所提供的检测方法可以实时检测PHA，显著缩短了检测时间，极大提高检测通量，整个过程发酵液样品不需要化学试剂处理，可显著降低PHA的检测成本，同时检测过程无须取样，并可排除由取样导致发酵体积变化，从而对发酵带来的干扰。
本申请提供了一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备，通过将从待检测发酵液中采集到的拉曼光谱信息输入至聚羟基脂肪酸酯PHA含量的检测模型，获取所述待检测发酵液中的PHA含量值；由于所述PHA含量的检测模型是根据不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的，以使得最终获取的PHA含量值检测准确，本申请能够克服因PHA发酵时的组分复杂、发酵周期长而导致无法实时检测的技术问题，通过所构建的聚羟基脂肪酸酯PHA检测模型有效检测检测结果，实现对发酵液中PHA的无损、高效、准确检测。
图2是本申请提供的处理所述拉曼光谱信息的流程示意图，所述基于所述PHA含量的检测模型处理所述拉曼光谱信息，包括：
将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间；
基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度；
通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
在步骤1031中，本申请首先划分预设波段区间，以获取所有指纹样本频移区间，将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间，例如，在一个以频移区间为X轴，以波数强度为Y轴的坐标系中，划分预设波段区间，可沿着X轴方向，划分为1800至1600的区间，1600至1400的区间，1400至1200的区间，1200至1000的区间以及1000至800的区间，而相对应Y轴而言，则可沿着Y轴方向依次设定波数强度的区间为0至120000。
在步骤1032中，基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度，将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间，以确定所述待检测发酵液中的PHA相关联的指纹样本频移区间，进而确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度，具体地，根据所述拉曼波峰的波峰点所在的坐标位置确定在所述指纹样本频移区间的波数强度，所述波数强度将通过所述波峰点的Y轴相应位置确定，例如，确定所述拉曼波峰在1600至1400的目标频移区间，紧接着，根据所述拉曼波峰的波峰点所在的坐标位置，确定波数强度值为100000。
在步骤1033中，通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
在一个可选地实施例中，定量关系可以由如上式(1)中所示，其中，Y代表PHA的含量值，X代表波数强度。本申请输入波数强度至上述定量关系中，进而通过计算输出所述待检测发酵液中的PHA含量值。
在一个可选地实施例中，所述采集所述待检测发酵液的目标拉曼光谱信息，包括：
根据拉曼光谱信号检测器的探头获取目标拉曼光谱信息；
所述拉曼光谱信号检测器的探头在目标拉曼光谱信息的获取阶段，被浸入至所述待检测发酵液中。
可选地，拉曼光谱信号检测器的探头是一种能够通过浸入至所述待检测发酵液中，获取待检测发酵液中目标拉曼光谱信息的设备，其可以通过光纤连接信号检测器，以获取目标拉曼光谱信息，在这样的实施例中，本申请采用了一种在线实时监测待检测发酵液中目标拉曼光谱信息的技术方案，以高效快速的分析出目标拉曼光谱信息中拉曼波峰的波数强度。
可选地，所述采集所述待检测发酵液的目标拉曼光谱信息，包括：
从所述待检测发酵液中提取预设份数的目标待检测发酵液；
输入每一目标待检测发酵液至拉曼光谱信号检测器，以获取所述拉曼光谱信号检测器输出的待检测拉曼光谱信息；
均值化处理所有待检测拉曼光谱信息，获取目标拉曼光谱信息；
每份目标发酵液的体积相同。
所述预设份数可以为3份、5份甚至更多，本申请可以在发酵液发酵的任意时刻从相应地发酵液承载装置中提取出5mL的PHA发酵液样品，所述发酵液承载装置可以为发酵罐，振荡2分钟混匀后加入至检测池中，将探头浸入 至发酵液样品中采集发酵液拉曼信号，在采集的过程中，将通过激发光源照射所述发酵液，可选地，设置激发波长为785nm，积分时间为5s，随机扫描样品5次。
在一个可选地实施例中，并不需要将探头浸入至发酵液样品中采集发酵液拉曼信号，而是将提取的目标待检测发酵液输入至拉曼光谱信号检测器中直接进行检测，以获取目标拉曼光谱信息，若在同一时刻获取了5份目标待检测发酵液，则确定5份与目标待检测发酵液相对应的目标拉曼光谱信息，其中，为保证检测结果的准确性，每份目标发酵液的体积相同。
可选地，均值化处理所有待检测拉曼光谱信息，获取目标拉曼光谱信息，本申请将通过均值化处理的形式处理所有的待检测拉曼光谱信息，以将均值化处理结果作为目标拉曼光谱信息。
在另一个可选地实施例中，本申请均值化处理所有待检测拉曼光谱信息，获取初始拉曼光谱信息，采用一阶求导处理所述初始拉曼光谱信息，获取去噪后拉曼光谱信息，将从所述去噪后拉曼光谱信息中截取出的预设波段区间内的拉曼光谱信息确定为目标拉曼光谱信息。
可选地，均值化处理所有待检测拉曼光谱信息后所获取的并不是目标拉曼光谱信息，而是初始拉曼光谱信息，所述初始拉曼光谱信息是并未经过去噪、截取处理的拉曼光谱信息。
可选地，选取不同发酵时间的PHA发酵液，同时以初始发酵培养基作为空白对照，采用拉曼光谱检测设备进行检测，通过特定波长的入射激光光源采集不同发酵时间的PHA发酵液和空白对照的拉曼光谱信息，确定PHA的特征拉曼峰，为减少背景噪音，例如荧光信号的干扰，对采集到的拉曼光谱进行预处理，将采集到的PHA发酵液拉曼光谱进行全波段分析，先后经过标准正态变换、求导、基线校正去除荧光信号，在这样的实施例中，本申请能够通过一阶求导对所测得的PHA发酵液的拉曼光谱数据进行荧光信号去除，同时校准基线，并对所有数据进行归一化处理。
本领域技术人员理解，所述的预处理方法包括但不限于卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理。
可选地，将从所述去噪后拉曼光谱信息中截取出的预设波段区间内的拉曼光谱信息确定为目标拉曼光谱信息，在一个可选地实施例中，本申请对拉曼光谱数据进行分析时选取波数为800cm-1至1800cm-1的指纹区图谱。
在一个可选地实施例中，在获取所述PHA检测模型输出的PHA含量检测值之前，还包括从样本发酵液接种发酵时刻开启计时，每间隔预设时长，采集所述样本发酵液的样本拉曼光谱信息；根据每一时刻的样本拉曼光谱信息，获取每一时刻样本发酵液的样本频移区间以及在每一样本频移区间内拉曼波峰的样本波数强度；基于气相色谱检测对每一时刻的样本发酵液检测，确定每一时刻的样本发酵液相对应的样本PHA含量值；根据每一时刻样本发酵液的样本频移区间、在每一样本频移区间内拉曼波峰的样本波数强度以及每一时刻的样本发酵液相对应的样本PHA含量值构建样本数据集；根据预设比例划分所述样本数据集后确定样本训练集以及样本测试集，以根据所述样本训练集以及样本测试集构建PHA检测模型。
在一个可选地实施例中，在不同的发酵时间，从发酵罐中取5mL的PHA发酵液样品，振荡2min混匀后，加入至检测池中，将拉曼探头浸入至发酵液 样品中采集发酵液拉曼信号，拉曼光谱采集参数为：激发波长785nm，积分时间5s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
而在另一个可选地实施例中，还可以从2L玻璃发酵罐接种发酵开始后，每隔2小时取3个平行样，每份样品10mL，将每个时间点取得的样品加入到检测池中，拉曼探头浸入至发酵液样品中采集信号。
根据每一时刻的样本拉曼光谱信息，获取每一时刻样本发酵液的样本频移区间以及在每一样本频移区间内拉曼波峰的样本波数强度，上述样本频移区间以及样本波数强度的确定，可以参考步骤101中采集待检测发酵液的拉曼光谱信息，从所述拉曼光谱信息中提取出拉曼波峰所在频移区间以及在所述频移区间中所述拉曼波峰的波数强度的技术方案，即将采集到的PHA发酵液样本拉曼光谱信息中的背景噪音(如荧光信号)去除，并对基线进行校正处理，并对所有数据做归一化处理，对样本拉曼光谱数据进行分析时均选取800～1800cm-1的指纹区图谱，映射所述样本拉曼光谱信息至所述预设波段区间，将样本拉曼波峰所在的频移区间确定为样本频移区间，确定所述拉曼波峰在所述样本频移区间中的样本波数强度，遍历所有样本拉曼光谱信息，以获取每一时刻样本发酵液的样本频移区间以及在每一样本频移区间内拉曼波峰的样本波数强度。
基于气相色谱检测对每一时刻的样本发酵液检测，确定每一时刻的样本发酵液相对应的样本PHA含量值，本申请将根据气相色谱检测PHA含量方法确定每一时刻的样本发酵液相对应的样本PHA含量值。
可选地，首先，取10mL样本发酵液置于称重后的15ml离心管中，加入10mL乙醇后采取离心处理，所述离心处理的条件为以每分钟10000转的速度运行5分钟，离心结束后加入20mL乙醇，洗涤菌体，以同样条件进行离心处理，离心后倒掉上清，在65℃的烘箱内烘干；然后，称取50mg烘干样品于试管中，加入2mL氯仿和2mL酯化液在100℃反应4小时，其中，酯化液的配置方法为：称取0.5g苯甲酸加入装有485ml的甲醇试剂瓶中，取15ml浓硫酸缓慢加入至甲醇试剂瓶中，混匀后完成酯化液的配制；最后，加入1mL超纯水涡旋振荡，进行萃取，静置30分钟至60分钟，使其产生分层，取下层有机相进行气相色谱分析，所述气相色谱分析的条件需满足：进样量1μL，色谱柱流量35mL/min，柱温240℃，流速23.4cm/s，吹扫流量3mL/min以及分流比39。
可选地，根据每一时刻样本发酵液的样本频移区间、在每一样本频移区间内拉曼波峰的样本波数强度以及每一时刻的样本发酵液相对应的样本PHA含量值构建样本数据集，将每一时刻的样本频移区间、样本波数强度以及样本PHA含量值作为标签组成一个样本数据集，进而根据所有时刻的样本发酵液，确定所有时刻的所有样本数据集。
可选地，根据预设比例划分所述样本数据集后确定样本训练集以及样本测试集，以根据所述样本训练集以及样本测试集构建PHA检测模型，所述预设比例可以为7:3、8:2或者其他比例，可选地，本申请将处理后到拉曼光谱和气相色谱测定的结果，例如共确定34个样本数据集，作为标签来建立偏最小二乘回归(Partial Least Squares Regression，PLSR)模型，其中，所述样本数据集分为样本训练集以及样本测试集，确定所述样本训练集和所述样本测试集的比例分别为70％和30％，以根据样本训练集以及样本测试集建立PHA检测模型。
本申请将处理后的拉曼光谱数据集按不同比例划分为训练集、测试集，建立基于拉曼光谱的PHA检测模型，以气相色谱法检测得到的不同发酵时间的PHA结果作为真实值，采用偏最小二乘法回归对训练集的拉曼光谱数据进行建模，建立波数强度与PHA特性的定量关系，本申请将测试集的拉曼光谱数据带入至检测模型中，根据检测结果，对检测模型进行修正以提高模型的泛化能力，所述PHA特性即为PHA含量或PHA浓度。
可选地，预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息，所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
图3是本申请提供的聚羟基脂肪酸酯含量的检测方法的流程示意图之二，本申请根据拉曼光谱采集标准样品，然后通过波长范围截取光谱，对光谱进行预处理，所述预处理包括去基线、求导、归一化，然后训练检测模型，判断模型的准确度，在模型准确度不足的情况下，重新返回选定波长范围截图光谱的步骤，在模型准确度达到预设准确度的情况下，确定检测模型，在确定检测模型后，光谱实时采集数据，并对拉曼数据进行相应地预处理，将其导入到所述检测模型中，以获取相应的检测结果。
在一个可选地实施例中，构建基于拉曼光谱用于检测PHA的模型通过如下方式实现：
PHA发酵液的拉曼光谱采集：在不同的发酵时间，从发酵罐中取5mL的PHA发酵液样品，振荡2min混匀后，加入至检测池中，将拉曼探头浸入至发酵液样品中采集发酵液拉曼信号，拉曼光谱采集参数为：激发波长785nm，积分时间5s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
拉曼光谱数据的预处理：通过一阶求导对所测得的PHA发酵液的拉曼光谱数据进行荧光信号去除，同时校准基线，并对所有数据进行归一化处理。在对拉曼光谱数据进行分析时均选取800～1800cm-1的指纹区图谱。
基于拉曼光谱的检测PHA模型构建：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，拉曼光谱数据集中分为训练集和测试集，训练集和测试集的比例70％和30％，如表1所示建立模型参数：
表1基于拉曼光谱PHA检测模型参数
其中，R2为确定系数(R-square)，越接近1表明对回归方程的解释能力越强，其模型对数据拟合的越好。
之后利用建立的检测模型来对未作为标签的发酵时间点的拉曼光谱数据进行验证分析，其中，未作为标签的发酵时间点的拉曼光谱数据处理方式与作为标签的拉曼光谱数据处理方式一致，模型检测结果如表2所示，气相色谱检测PHA结果与拉曼光谱建模检测结果具有较好的相关关系，误差值在4％以内。
表2模型检测结果
本申请中可选地采用发酵罐作为样品采样中发酵液的发酵载体，所述的发酵罐包括但不限于不锈钢发酵罐、玻璃发酵罐、塑料发酵罐，拉曼光谱采集条件为：使用785nm波长激光、采集范围300-3200cm-1、分辨率为5cm-1、激光功率为500mW、采集频次为10-30s/次、累计采集1-10次，可选地，采集5次。
在一个可选地实施例中，本申请利用玻璃发酵罐离线取样检测PHA含量，包括如下过程：
首先，PHA发酵液的拉曼光谱采集：从2L玻璃发酵罐接种发酵开始后，每隔2h取3个平行样，每份样品10mL，将每个时间点取得的样品加入到检测池中，拉曼探头浸入至发酵液样品中采集信号，其中PHA发酵液产品为PHBHHx，拉曼光谱采集参数为：激发波长785nm，积分时间5s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
然后，拉曼光谱数据的处理：将采集到的PHA发酵液拉曼光谱数据中的背景噪音去除，例如荧光信号的去除，并对基线进行校正处理，并对所有数据做归一化处理，对拉曼光谱数据进行分析时均选取800～1800cm-1的指纹区图谱。
最后，拉曼光谱建模检测分析：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，样本数据集中分为训练集和测试集，训练集和测试集的比例70％和30％。之后利用建立的PLSR模型来对未作为标签的其他发酵时间点的拉曼光谱数据进行验证分析，未作为标签的其他发酵时间点的拉曼光谱数据处理方式与作为标签拉曼光谱数据的方式一致，模型检测结果如表3所示，可以看出，气相色谱检测PHA结果与拉曼光谱建模检测结果具有较好的相关关系，误差值在7％以内。
表3玻璃罐发酵检测结果
本领域技术人员理解，本申请可以取1次光谱数据代表样品光谱信息，也可以5次光谱数据代表样品光谱信息，在取1次光谱数据代表样品光谱信息的实施例中，本申请利用不锈钢发酵罐在线检测，确定取1次光谱数据代表样品光谱信息的情况下，气相色谱检测PHA含量结果与拉曼光谱建模检测结果的相关性分析，具体包括：
首先，PHA发酵液的拉曼光谱实时采集：将拉曼光谱探头浸入式安装在75L不锈钢发酵罐中，经过灭菌和接种后，从发酵开始，实时采集发酵液拉曼信号，其中PHA发酵液为聚3-羟基丁酸酯-共-3-羟基己酸酯(PHBHHx)，拉曼光谱采集参数为：激发波长785nm，积分时间10s，随机扫描样品1次，取1次光谱数据代表样品光谱信息。
然后，处理拉曼光谱：将采集到的PHA发酵液拉曼光谱进行全波段分析，先后经过标准正态变换、基线校正去除荧光信号。
最后，拉曼光谱建模检测分析：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，样本数据集中分为训练集和测试集，训练集和测试集的比例70％和30％，之后利用建立的PLSR模型来对未作为标签的其他发酵时间点的拉曼光谱数据进行验证分析，未作为标签的其他发酵时间点的拉曼光谱数据处理方式与作为标签拉曼光谱数据的方式一致，模型的检测结果如表4所示，气相色谱检测PHA含量结果与拉曼光谱建模检测结果具有很好的相关关系，误差值在8％以内。
表4不锈钢罐发酵1次连续光谱检测结果

本领域技术人员理解，本申请也可以5次光谱数据代表样品光谱信息，在取5次光谱数据代表样品光谱信息的实施例中，本申请利用不锈钢发酵罐在线检测，确定取5次光谱数据代表样品光谱信息的情况下，气相色谱检测PHA含量结果与拉曼光谱建模检测结果的相关性分析，具体包括：
首先，PHA发酵液的拉曼光谱实时采集：将拉曼光谱探头浸入式安装在75L不锈钢发酵罐中，经过灭菌和接种后，从发酵开始，实时采集发酵液拉曼信号，其中PHA发酵液产品为PHBHHx，拉曼光谱采集参数为：激发波长785nm，积分时间10s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
然后，处理拉曼光谱：将采集到的PHA发酵液拉曼光谱进行全波段分析，先后经过标准正态变换、基线校正去除荧光信号。
最后，拉曼光谱建模检测分析：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，样本数据集中分为训练集和测试集，训练集和测试集的比例70％和30％。之后利用建立的PLSR模型来对未作为标签的其他发酵时间点的拉曼光谱数据进行验证分析，未作为标签的其他发酵时间点的拉曼光谱数据处理方式与作为标签拉曼光谱数据的方式一致，模型检测结果如表5所示，气相色谱检测PHA含量结果与拉曼光谱建模检测结果具有很好的相关关系，误差值在5％以内。
表5不锈钢罐发酵检测5次连续光谱结果
在表5中，提供了一种未经过一阶求导处理所述初始拉曼光谱信息， 以根据未经过一阶求导处理的拉曼光谱构建检测模型的技术方案，而在表5实施例的基础上，若采用一阶求导处理所述初始拉曼光谱信息，获取去噪后拉曼光谱信息，以根据一阶求导处理的拉曼光谱构建检测模型，则具体包括：
首先，PHA发酵液的拉曼光谱实时采集：将拉曼光谱探头浸入式安装在75L不锈钢发酵罐中，经过灭菌和接种后，从发酵开始，实时采集发酵液拉曼信号，其中PHA发酵液产品为PHBHHx，拉曼光谱采集参数为：激发波长785nm，积分时间10s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
然后，处理拉曼光谱：将采集到的PHA发酵液拉曼光谱进行全波段分析，先后经过标准正态变换、基线校正去除荧光信号。
最后，拉曼光谱建模检测分析：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，样本数据集中分为训练集和测试集，训练集和测试集的比例70％和30％，之后利用建立的PLSR模型来对未作为标签的其他发酵时间点的拉曼光谱数据进行验证分析，未作为标签的其他发酵时间点的拉曼光谱数据处理方式与作为标签拉曼光谱数据的方式一致，模型检测结果如表6所示，气相色谱检测PHA含量结果与拉曼光谱建模检测结果具有很好的相关关系，误差值在3％以内。
表6不锈钢罐发酵的一阶求导检测结果
在另一个可选地实施例中，本申请还将结合不同PHA浓度对气相色谱检测PHA含量结果与拉曼光谱建模检测结果的相关关系进行验证，具体地，包括：
首先，PHA发酵液的拉曼光谱实时采集：将拉曼光谱探头浸入式安装在75L不锈钢发酵罐中，经过灭菌和接种后，从发酵开始，实时采集发酵液拉曼信号，其中PHA发酵液产品为聚3-羟基丁酸酯PHB，拉曼光谱采集参数为：激发波长785nm，积分时间10s，随机扫描样品5次，取5次平均光谱代表样品光谱信息。
然后，处理拉曼光谱：将采集到的PHA发酵液拉曼光谱进行全波段分 析，先后经过标准正态变换、求导、基线校正去除荧光信号。
最后，拉曼光谱建模检测分析：将处理后到拉曼光谱和气相色谱测定的结果作为标签来建立PLSR模型，样本数据集中分为训练集和测试集，训练集和测试集的比例70％和30％，之后利用建立的PLSR模型来对其他未作为标签的样本进行检测分析，模型检测结果如表7所示。可以看出，不同PHA浓度的气相色谱检测PHA含量结果与拉曼光谱建模检测结果具有很好的相关关系，误差值在5％以内。
表7不同PHA浓度发酵检测结果
图4是本申请提供的聚羟基脂肪酸酯含量的检测系统的结构示意图，包括；
检测容器1，用于为发酵液提供检测环境；
探头2，用于浸入至检测池中采集拉曼光谱信息；
光纤3，用于探头与激发光源的信号传输，以及探头与信号检测器的信号传输；
激发光源4，用于为发酵液提供检测光源；
信号检测器5，用于将光信号转换为数据信号；
还包括所述的聚羟基脂肪酸酯含量的检测装置，所述检测装置用于对所收集的拉曼光谱信息分析处理后，输出待测发酵液中的聚羟基脂肪酸酯含量。
所述聚羟基脂肪酸酯含量的检测装置可以作为一个整体，对所收集的拉曼光谱信息分析处理后，输出待测发酵液中的聚羟基脂肪酸酯含量，而在另一个可选地实施例中，所述聚羟基脂肪酸酯含量的检测装置可以进一步地细分为：
数据收集单元6，用于收集拉曼光谱信息；
分析处理单元7，用于对所收集的拉曼光谱信息分析处理后，输出待测发酵液中的聚羟基脂肪酸酯含量。
如图4所示，为了实时检测发酵液中PHA的含量，基于拉曼光谱检测发酵液中PHA的装置包括检测容器1，所述检测容器为检测池，所述激发 光源4通过光纤3与探头2相连，探头2通过光纤3与信号检测器5相连，信号检测器5将收集的信号经过数据线或无线网络传输至数据收集单元6和分析处理单元7。
本领域技术人员理解，所述聚羟基脂肪酸酯含量的检测系统可用于在线检测和离线检测两种场景，在在线检测场景下，所述探头2浸入至发酵罐中，在发酵过程中实时采集发酵液内部的成分信号，之后信号检测器5将检测到的拉曼光谱信号传输至数据收集单元6，采用分析处理单元7对数据收集单元6收集的拉曼信号进一步进行处理分析后获得PHA的成分信息；
在离线场景下，首先从发酵罐中取发酵液样品，之后将样品放置在检测容器1中，所述检测容器1内部完全处于避光环境，然后将探头2浸入发酵液样品中，采集发酵液样品的成分信号，之后信号检测器5将检测到的拉曼光谱信号传输至数据收集单元6，采用分析处理单元7对数据收集单元6收集的拉曼信号进一步进行处理分析获得PHA的成分信息。
更为具体地，本申请还包括存储器及存储在所述存储器上并可在所述分析处理单元7上运行的程序或指令，所述程序或指令被所述分析处理单元7执行时执行所述聚羟基脂肪酸酯含量的检测方法，该方法包括：采集待检测发酵液的拉曼光谱信息；输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
本申请提供了一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备，通过将从待检测发酵液中采集到的拉曼光谱信息输入至聚羟基脂肪酸酯PHA含量的检测模型，获取所述待检测发酵液中的PHA含量值；由于所述PHA含量的检测模型是根据不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的，以使得最终获取的PHA含量值检测准确，本申请能够克服因PHA发酵时的组分复杂、发酵周期长而导致无法实时检测的技术问题，通过所构建的聚羟基脂肪酸酯PHA检测模型有效检测检测结果，实现对发酵液中PHA的无损、高效、准确检测。
图5是本申请提供的聚羟基脂肪酸酯含量的检测装置的结构示意图，本申请还提供了一种聚羟基脂肪酸酯含量的检测装置，包括采集单元51：用于采集待检测发酵液的拉曼光谱信息，所述采集单元51的工作原理可以参考前述步骤101，在此不予赘述。
所述聚羟基脂肪酸酯含量的检测装置还包括输入单元52：用于输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型，所述输入单元52的工作原理可以参考前述步骤102，在此不予赘述。
所述聚羟基脂肪酸酯含量的检测装置还包括处理单元53：用于基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值，所述处理单元53的工作原理可以参考前述步骤103，在此不予赘述。
所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定 量关系；
所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；
所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
可选地，所述处理单元还包括映射子单元531：用于将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间，所述映射子单元531的工作原理可以参考前述步骤1031，在此不予赘述。
所述处理单元还包括确定子单元532：用于基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度，所述确定子单元532的工作原理可以参考前述步骤1032，在此不予赘述。
所述处理单元还包括输出子单元533：用于通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值，所述输出子单元533的工作原理可以参考前述步骤1033，在此不予赘述。
可选地，所述检测装置还包括：
预处理单元54：用于预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息，所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
本申请提供了一种聚羟基脂肪酸酯含量的检测方法、装置、系统、设备，通过将从待检测发酵液中采集到的拉曼光谱信息输入至聚羟基脂肪酸酯PHA含量的检测模型，获取所述待检测发酵液中的PHA含量值；由于所述PHA含量的检测模型是根据不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的，以使得最终获取的PHA含量值检测准确，本申请能够克服因PHA发酵时的组分复杂、发酵周期长而导致无法实时检测的技术问题，通过所构建的聚羟基脂肪酸酯PHA检测模型有效检测检测结果，实现对发酵液中PHA的无损、高效、准确检测。
图6是本申请提供的电子设备的结构示意图。如图6所示，该电子设备可以包括：处理器(processor)610、通信接口(Communications Interface)620、存储器(memory)630和通信总线640，其中，处理器610，通信接口620，存储器630通过通信总线640完成相互间的通信。处理器610可以调用存储器630中的逻辑指令，以执行聚羟基脂肪酸酯含量的检测方法，该方法包括：采集待检测发酵液的拉曼光谱信息；输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及样本PHA含量值训练得到的定量关系；所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
此外，上述的存储器630中的逻辑指令可以通过软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该 计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
另一方面，本申请还提供一种计算机程序产品，所述计算机程序产品包括计算机程序，计算机程序可存储在非暂态计算机可读存储介质上，所述计算机程序被处理器执行时，计算机能够执行上述各方法所提供的一种聚羟基脂肪酸酯含量的检测方法，该方法包括：采集待检测发酵液的拉曼光谱信息；输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及样本PHA含量值训练得到的定量关系；所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
又一方面，本申请还提供一种非暂态计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现以执行上述各方法提供聚羟基脂肪酸酯含量的检测方法，该方法包括：采集待检测发酵液的拉曼光谱信息；输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及样本PHA含量值训练得到的定量关系；所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是：以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (10)

  1. 一种聚羟基脂肪酸酯含量的检测方法，包括：
    采集待检测发酵液的拉曼光谱信息；
    输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；
    基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；
    所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；
    所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；
    所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
  2. 根据权利要求1所述的聚羟基脂肪酸酯含量的检测方法，其中，所述拉曼光谱信息包括所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间以及在所述频移区间中所述拉曼波峰的波数强度；
    所述发酵液的拉曼光谱信息与PHA含量值的定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息中的所述波数强度以及PHA含量值通过训练得到的定量关系。
  3. 根据权利要求2所述的聚羟基脂肪酸酯含量的检测方法，其中，所述基于所述PHA含量的检测模型处理所述拉曼光谱信息，包括以下步骤：
    将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间；
    基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度；
    通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
  4. 根据权利要求1所述的聚羟基脂肪酸酯含量的检测方法，其中，在采集待检测发酵液的拉曼光谱信息之后，还包括：
    预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息；
    所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
  5. 根据权利要求1-4中任一项所述的聚羟基脂肪酸酯含量的检测方法，其中，所述不同发酵条件包括：
    不同的发酵容器；
    或，不同单体的PHA对应的不同菌株；
    或，不同发酵基质；
    或，不同活性的菌株。
  6. 一种聚羟基脂肪酸酯含量的检测装置，其中，包括：
    采集单元：用于采集待检测发酵液的拉曼光谱信息；
    输入单元：用于输入所述拉曼光谱信息至聚羟基脂肪酸酯PHA含量的检测模型；
    处理单元：用于基于所述PHA含量的检测模型处理所述拉曼光谱信息，并输出所述待检测发酵液中的PHA含量值；
    所述PHA含量的检测模型包括发酵液的拉曼光谱信息与PHA含量值的定量关系；
    所述定量关系是基于不同发酵条件下样本发酵液的拉曼光谱信息以及不同发酵条件下的样本PHA含量值训练得到的定量关系；
    所述样本PHA含量值是基于气相色谱检测对所述样本发酵液检测而确定的。
  7. 根据权利要求6所述的聚羟基脂肪酸酯含量的检测装置，其中，所述处理单元还包括：
    映射子单元：用于将所述待检测发酵液中的PHA对应的拉曼波峰所在频移区间映射至所述检测模型中对应的指纹样本频移区间；
    确定子单元：用于基于对应的指纹样本频移区间，确定所述待检测发酵液中的PHA对应的拉曼波峰在所述频移区间中的波数强度；
    输出子单元：用于通过所述波数强度与PHA含量之间的定量关系，输出所述待检测发酵液中的PHA含量值。
  8. 根据权利要求6或7所述的聚羟基脂肪酸酯含量的检测装置，其中，所述检测装置还包括：
    预处理单元：用于预处理所述拉曼光谱信息，得到去噪后的拉曼光谱信息；
    所述预处理的方式包括卷积平滑处理、基线校正处理、多元散射校正处理、正交信号校正处理、标准正态变换处理、归一化处理、高斯滤波处理、中值滤波处理中的至少一种。
  9. 一种聚羟基脂肪酸酯含量的检测系统，其中，包括：
    检测容器，用于为发酵液提供检测环境；
    探头，用于浸入至检测池中采集拉曼光谱信息；
    光纤，用于探头与激发光源的信号传输，以及探头与信号检测器的信号传输；
    激发光源，用于为发酵液提供检测光源；
    信号检测器，用于将光信号转换为数据信号；
    还包括权利要求6-8中任一项所述的聚羟基脂肪酸酯含量的检测装置，所述检测装置用于对所收集的拉曼光谱信息分析处理后，输出待测发酵液中的聚羟基脂肪酸酯含量。
  10. 一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，其中，所述处理器执行所述程序时实现如权利要求1至6任一项所述的聚羟基脂肪酸酯含量的检测方法。
PCT/CN2023/119116 2022-09-23 2023-09-15 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备 WO2024061126A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211167876.6 2022-09-23
CN202211167876.6A CN115950871B (zh) 2022-09-23 2022-09-23 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备

Publications (1)

Publication Number Publication Date
WO2024061126A1 true WO2024061126A1 (zh) 2024-03-28

Family

ID=87284767

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/119116 WO2024061126A1 (zh) 2022-09-23 2023-09-15 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备

Country Status (2)

Country Link
CN (1) CN115950871B (zh)
WO (1) WO2024061126A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115950871B (zh) * 2022-09-23 2023-12-15 北京蓝晶微生物科技有限公司 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100136609A1 (en) * 2008-10-31 2010-06-03 Biomerieux, Inc. Method for separation, characterization and/or identification of microorganisms using raman spectroscopy
KR20170053360A (ko) * 2015-11-06 2017-05-16 최종수 근적외선 분광분석기를 이용한 아미노산 발효공정의 실시간관리방법
CN110987895A (zh) * 2019-11-07 2020-04-10 江苏大学 一种快速在线检测木薯固态发酵的生长过程的方法
CN113252634A (zh) * 2020-02-13 2021-08-13 凯塞光学系统股份有限公司 使用拉曼光谱法实时监测葡萄酒发酵特性
US20210372932A1 (en) * 2020-02-13 2021-12-02 Kaiser Optical Systems Inc. Real-time raman spectroscopic monitoring of wine properties and constituents during wine production
CN114216893A (zh) * 2021-12-15 2022-03-22 河南科技大学 一种食用油反式脂肪酸的检测方法
WO2022127203A1 (zh) * 2020-12-17 2022-06-23 山东省科学院生物研究所 一种基于拉曼光谱检测发酵液中多糖的装置及其方法和应用
CN115950871A (zh) * 2022-09-23 2023-04-11 北京蓝晶微生物科技有限公司 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114264643B (zh) * 2021-12-20 2024-03-29 江南大学 基于单细胞拉曼光谱快速预测食醋发酵进程的方法

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100136609A1 (en) * 2008-10-31 2010-06-03 Biomerieux, Inc. Method for separation, characterization and/or identification of microorganisms using raman spectroscopy
KR20170053360A (ko) * 2015-11-06 2017-05-16 최종수 근적외선 분광분석기를 이용한 아미노산 발효공정의 실시간관리방법
CN110987895A (zh) * 2019-11-07 2020-04-10 江苏大学 一种快速在线检测木薯固态发酵的生长过程的方法
CN113252634A (zh) * 2020-02-13 2021-08-13 凯塞光学系统股份有限公司 使用拉曼光谱法实时监测葡萄酒发酵特性
US20210255040A1 (en) * 2020-02-13 2021-08-19 Kaiser Optical Systems Inc. Real-time monitoring of wine fermentation properties using raman spectroscopy
US20210372932A1 (en) * 2020-02-13 2021-12-02 Kaiser Optical Systems Inc. Real-time raman spectroscopic monitoring of wine properties and constituents during wine production
WO2022127203A1 (zh) * 2020-12-17 2022-06-23 山东省科学院生物研究所 一种基于拉曼光谱检测发酵液中多糖的装置及其方法和应用
CN114216893A (zh) * 2021-12-15 2022-03-22 河南科技大学 一种食用油反式脂肪酸的检测方法
CN115950871A (zh) * 2022-09-23 2023-04-11 北京蓝晶微生物科技有限公司 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备

Also Published As

Publication number Publication date
CN115950871B (zh) 2023-12-15
CN115950871A (zh) 2023-04-11

Similar Documents

Publication Publication Date Title
WO2024061126A1 (zh) 聚羟基脂肪酸酯含量的检测方法、装置、系统、设备
CN103134767B (zh) 一种红外光谱校正鉴定白酒品质的方法
CN102692388B (zh) 数字成像系统及其快速检测果蔬中有机磷农药残留的方法
Véstia et al. Predicting calcium in grape must and base wine by FT-NIR spectroscopy
CN109668858A (zh) 基于近红外光谱检测发酵过程生物量和组分浓度的方法
CN102636450A (zh) 基于近红外光谱技术无损检测枸杞中枸杞多糖含量的方法
CN101221125A (zh) 用光谱技术测定富营养化水体特征参量的方法
CN105548138A (zh) 基于拉曼光谱的酿酒发酵液酵母活力在线检测装置及方法
CN110987895B (zh) 一种快速在线检测木薯固态发酵的生长过程的方法
CN105628644A (zh) 基于原位实时光谱在线监测蛋白质酶解过程的装置和方法
CN102435575A (zh) 一种检测近红外二维相关光谱的方法
CN103439269A (zh) 近红外快速检测酒醅中理化指标的方法
CN102928379A (zh) 一种近红外光谱技术快速检测增健口服液多糖含量的方法
CN103558180A (zh) 近红外快速检测糖化发酵剂中理化指标的方法
CN111289489A (zh) 一种基于拉曼光谱的微生物单细胞生长检测方法
CN101566564A (zh) 反射干涉光谱检测乙醇浓度的检测系统及方法
CN101871898A (zh) 一种基于气味蒸发特征谱的嗅觉检测方法及其系统
CN101769867A (zh) 一种堆肥产品品质的无损检测方法
CN104483287A (zh) 基于近红外光谱的在线发酵过程生物学参数的检测装置及方法
CN102768191B (zh) 一种水中痕量铊的简便测定方法
CN102288569A (zh) 一种纤维生物质的快速分析方法
CN111595807B (zh) 一种生物基食品包装膜中己内酰胺的定量检测方法
Zhou et al. ATR-FTIR spectroscopy combined with chemometrics to assess the spectral markers of irradiated baijius and their potential application in irradiation dose control
CN103728269A (zh) 一种近红外快速检测酿酒原料中理化指标的方法
CN115718089A (zh) 一种基于菌群拉曼特征快速识别样品类别的方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23867415

Country of ref document: EP

Kind code of ref document: A1