Mass spectrometry-based rapid detection system for respiratory sputum components in critically ill patients

Abstract

The present application relates to the technical field of component detection, in particular to a rapid detection system for respiratory tract sputum components of severe patients based on mass spectrometry. The system can realize the following steps through the cooperation between multiple modules: obtaining a target mass spectrum corresponding to a respiratory tract sputum sample of a patient to be detected, and identifying a target characteristic peak and its corresponding isotope peak cluster in the target mass spectrum; determining a pathogen compliance index, protein flooding possibility and isotope peak distribution rationality corresponding to each target characteristic peak; screening protein flooding characteristic peaks from all target characteristic peaks; segmenting target pathogen peaks from the protein flooding characteristic peaks, and detecting respiratory tract sputum components according to the target characteristic peaks and the segmented target pathogen peaks. The present application realizes the segmentation of protein flooding characteristic peaks by analyzing the protein flooding condition in the characteristic peaks, thereby facilitating the identification of the flooded pathogen, and further improving the accuracy of respiratory tract sputum component detection.

Description

Technical Field

[0001] This invention relates to the field of component detection technology, specifically to a rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry. Background Technology

[0002] Mass spectrometry-based component detection is the cornerstone of modern component analysis, with extremely wide applications. For example, it can be used for the rapid detection of components in the respiratory sputum of critically ill patients. Currently, the common method for component detection based on mass spectrometry is to match the characteristic peaks in the acquired mass spectrum with an existing database, and then perform component detection based on the matching results.

[0003] However, when detecting respiratory sputum components by matching the characteristic peaks in the mass spectrum of a respiratory sputum sample with the characteristic peaks of existing sputum components recorded in the database, the following technical problems often arise:

[0004] In reality, the intensity of the mass spectrometry peaks of high-abundance host proteins in sputum often far exceeds that of other signals. Their strong baseline may "submerge" low-abundance pathogen peaks, making the latter difficult to identify directly, which may lead to missed detection of pathogens in sputum. Therefore, when detecting respiratory sputum components directly based on the matching of characteristic peaks, the accuracy of respiratory sputum component detection is often poor. Summary of the Invention

[0005] To address the technical problem of poor accuracy in detecting respiratory sputum components, this invention proposes a rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry.

[0006] In a first aspect, the present invention provides a rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry, the system comprising:

[0007] The acquisition and identification module is used to acquire the target mass spectrum corresponding to the respiratory sputum sample of the patient to be tested, and to identify the target characteristic peaks and their corresponding isotope peak clusters in the target mass spectrum.

[0008] The indicator determination module is used to determine the pathogen conformity index corresponding to each target feature peak based on the curvature distribution of different points on the fitted curve corresponding to each target feature peak.

[0009] The probability determination module is used to determine the probability of protein submersion for each target feature peak based on the intersection between the standard protein feature peak corresponding to the pre-acquired standard host protein and each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak.

[0010] The rationality determination module is used to determine the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), front width, and back width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak.

[0011] The screening module is used to screen out protein flooding characteristic peaks from all target characteristic peaks based on the probability of protein flooding and the rationality of isotope peak distribution.

[0012] The segmentation and detection module is used to segment the target pathogen peak from the protein flooding characteristic peak, and to detect respiratory sputum components based on the target characteristic peak and the segmented target pathogen peak.

[0013] In conjunction with the first aspect above, in one possible implementation, determining the pathogen compliance index corresponding to each target feature peak based on the curvature distribution at different points on the fitted curve corresponding to each target feature peak includes:

[0014] Based on the maximum value of the curvature of all coordinate points on the fitted curve corresponding to each target feature peak, and the standard deviation of the curvature of all coordinate points on the fitted curve corresponding to each target feature peak, the pathogen compliance index corresponding to each target feature peak is determined.

[0015] In conjunction with the first aspect above, in one possible implementation, determining the protein flooding probability corresponding to each target feature peak based on the intersection between the standard protein feature peak corresponding to the pre-acquired standard host protein and each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, includes:

[0016] Based on the mass-to-charge ratio ranges corresponding to all standard protein characteristic peaks and the mass-to-charge ratio range corresponding to each target characteristic peak, the matching protein characteristic peaks corresponding to each target characteristic peak are selected from all standard protein characteristic peaks.

[0017] The probability of protein flooding for each target feature peak is determined by the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak.

[0018] In conjunction with the first aspect above, in one possible implementation, the step of selecting the matching protein characteristic peak corresponding to each target characteristic peak from all standard protein characteristic peaks based on the mass-to-charge ratio ranges corresponding to all standard protein characteristic peaks and the mass-to-charge ratio range corresponding to each target characteristic peak includes:

[0019] Any target characteristic peak is identified as the labeled characteristic peak, and the length of the intersection between the mass-to-charge ratio range corresponding to each standard protein characteristic peak and the mass-to-charge ratio range corresponding to the labeled characteristic peak is identified as the target intersection length between each standard protein characteristic peak and the labeled characteristic peak.

[0020] The standard protein characteristic peak with the largest target intersection length with the labeled characteristic peak is selected from all standard protein characteristic peaks and used as the matching protein characteristic peak corresponding to the labeled characteristic peak.

[0021] In conjunction with the first aspect above, in one possible implementation, determining the protein flooding probability corresponding to each target feature peak based on the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, includes:

[0022] The length of the intersection between the mass-to-charge ratio range of each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak is determined as the matching intersection length for each target feature peak.

[0023] The length of the mass-to-charge ratio range of the matching protein feature peaks corresponding to each target feature peak is determined as the length of the matching peak corresponding to each target feature peak.

[0024] Based on the difference between the matching intersection length and the matching peak length corresponding to each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, the probability of protein submersion corresponding to each target feature peak is determined.

[0025] In conjunction with the first aspect above, in one possible implementation, determining the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), leading edge width, and trailing edge width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak includes:

[0026] The product of the number of atoms corresponding to each isotope peak and the peak intensity is used to determine the initial intensity representative value corresponding to each isotope peak.

[0027] The ratio between the initial intensity representative values ​​corresponding to every two isotope peaks in the isotope peak cluster is determined as the initial intensity ratio between every two isotope peaks in the isotope peak cluster.

[0028] Based on the leading edge width and trailing edge width of each isotope peak, determine the characteristic value of the leading and trailing edge widths of each isotope peak.

[0029] Based on the difference between the half-width at half-maximum (WHM) of each pair of isotope peaks in the isotope peak cluster, the initial intensity ratio between each pair of isotope peaks, and the difference between the leading and trailing edge width eigenvalues ​​of each pair of isotope peaks, the target intensity ratio representative value between each pair of isotope peaks in the isotope peak cluster is determined.

[0030] Based on the representative values ​​of the target intensity ratios among isotope peaks in the isotope peak clusters corresponding to all standard protein characteristic peaks, determine the theoretical representative value of each target characteristic peak.

[0031] The rationality of the isotope peak distribution corresponding to each target characteristic peak is determined based on the difference between the target intensity ratio representative value and its corresponding theoretical ratio representative value in the isotope peak cluster corresponding to each target characteristic peak.

[0032] In conjunction with the first aspect above, in one possible implementation, determining the theoretical proportional representative value corresponding to each target characteristic peak based on the target intensity proportional representative value among the isotope peaks in the isotope peak clusters corresponding to all standard protein characteristic peaks includes:

[0033] Select the matching protein characteristic peaks corresponding to each target characteristic peak from all standard protein characteristic peaks;

[0034] The average of the target intensity ratios among all isotope peaks in the isotope peak cluster corresponding to the matching protein characteristic peak for each target characteristic peak is determined as the theoretical ratio representative value for each target characteristic peak.

[0035] In conjunction with the first aspect above, in one possible implementation, the step of selecting protein-submerged characteristic peaks from all target characteristic peaks based on the probability of protein submersion and the rationality of isotope peak distribution includes:

[0036] Based on the probability of protein smothering and the rationality of isotope peak distribution corresponding to each target feature peak, the target masking index corresponding to each target feature peak is determined.

[0037] If the target masking index corresponding to the target feature peak is greater than the preset masking threshold, then the target feature peak is determined as a protein smothering feature peak.

[0038] In conjunction with the first aspect above, in one possible implementation, determining the leading and trailing edge width characteristic values ​​corresponding to each isotope peak based on the leading edge width and trailing edge width corresponding to each isotope peak includes:

[0039] The ratio between the leading edge width and the trailing edge width of each isotope peak is determined as the characteristic value of the leading and trailing edge widths of each isotope peak.

[0040] In conjunction with the first aspect above, in one possible implementation, the step of segmenting the target pathogen peak from the protein flooding characteristic peak includes:

[0041] Based on the pre-acquired characteristic peaks corresponding to standard pathogens, as well as the protein flooding characteristic peaks and their corresponding target masking indices, the target pathogen peaks characterizing the pathogens are segmented from the protein flooding characteristic peaks using CVAE.

[0042] Secondly, the present invention provides a method for rapid detection of respiratory sputum components in critically ill patients based on mass spectrometry, implemented using a mass spectrometry-based rapid detection system for respiratory sputum components. The method includes:

[0043] Obtain the target mass spectrum corresponding to the respiratory sputum sample of the patient to be tested, and identify the target characteristic peaks and their corresponding isotope peak clusters in the target mass spectrum;

[0044] Based on the curvature distribution of different points on the fitted curve corresponding to each target feature peak, the pathogen compliance index corresponding to each target feature peak is determined.

[0045] Based on the intersection between the standard protein characteristic peaks corresponding to the pre-obtained standard host proteins and each target characteristic peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target characteristic peak, the probability of protein submersion corresponding to each target characteristic peak is determined.

[0046] Based on the number of atoms, peak intensity, full width at half maximum (FWHM), front width, and back width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak, determine the rationality of the isotope peak distribution corresponding to each target characteristic peak.

[0047] Based on the probability of protein immersion and the rationality of isotope peak distribution, protein immersion characteristic peaks are selected from all target characteristic peaks.

[0048] The target pathogen peak is separated from the protein flooding characteristic peak, and the respiratory sputum components are detected based on the target characteristic peak and the separated target pathogen peak.

[0049] Thirdly, a server is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, enabling the device to perform the aforementioned mass spectrometry-based rapid detection method for respiratory sputum components in critically ill patients.

[0050] Fourthly, a computer program product is provided, comprising: computer program code, which, when run on a computer, causes the computer to execute the aforementioned rapid detection method for respiratory sputum components in critically ill patients based on mass spectrometry.

[0051] Fifthly, a computer-readable storage medium is provided that stores computer program code, which, when executed on a computer, causes the computer to perform the aforementioned rapid detection method for respiratory sputum components in critically ill patients based on mass spectrometry.

[0052] The present invention has the following beneficial effects:

[0053] This invention provides a rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry. By analyzing the protein submersion status of characteristic peaks, it achieves the segmentation of protein-submerged characteristic peaks, thereby facilitating the identification of submerged pathogens. This solves the technical problem of poor accuracy in respiratory sputum component detection and improves its accuracy. Specifically, this invention analyzes the target mass spectrum corresponding to the respiratory sputum sample, quantifying multiple indicators related to the protein submersion status of characteristic peaks, such as pathogen conformity indicators, protein submersion probability, and the rationality of isotope peak distribution. This allows for the screening of protein-submerged characteristic peaks, and the segmentation of target pathogen peaks representing submerged pathogens from these peaks. This facilitates subsequent identification of submerged pathogens, thereby improving the accuracy of respiratory sputum component detection. Attached Figure Description

[0054] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a schematic diagram of the structure of the mass spectrometry-based rapid detection system for respiratory sputum components in critically ill patients according to the present invention.

[0056] Figure 2 This is a flowchart of the rapid detection method for respiratory sputum components in critically ill patients based on mass spectrometry according to the present invention;

[0057] Figure 3 This is a schematic diagram of the structure of a computer device according to the present invention. Detailed Implementation

[0058] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the technical solution proposed according to the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0060] In mass spectrometry-based sputum detection, the acquired mass spectrometry data often contains rich biomolecular information, which together form an interpretation of the patient's respiratory status. In this embodiment of the invention, characteristic peaks of sputum are extracted by mass spectrometry and matched with pathogen information to achieve rapid and accurate identification of pathogens infecting critically ill patients.

[0061] For complex samples such as sputum containing a large amount of host proteins, drug residues and other interference, the true pathogen characteristics can be separated from the mass spectrometry signal, which can avoid missed diagnosis to a certain extent.

[0062] refer to Figure 1 A schematic diagram of the structure of a mass spectrometry-based rapid detection system for respiratory sputum components in critically ill patients according to the present invention is shown. This mass spectrometry-based rapid detection system for respiratory sputum components in critically ill patients includes:

[0063] The acquisition and identification module 101 is used to acquire the target mass spectrum corresponding to the respiratory sputum sample of the patient to be tested, and to identify the target characteristic peaks and their corresponding isotope peak clusters in the target mass spectrum.

[0064] The patients to be tested can be critically ill patients (ICU patients, Intensive Care Unit patients) undergoing respiratory sputum composition analysis. Critically ill patients are often at extremely high risk of respiratory infection due to underlying diseases, immunosuppression, mechanical ventilation, etc. Underlying diseases can include, but are not limited to: severe trauma, major surgery, sepsis, and acute exacerbations of chronic diseases. The respiratory sputum sample from the patient to be tested can be a collected sputum sample from the patient to be tested. The target mass spectrum can be the mass spectrum of the respiratory sputum sample from the patient to be tested. The target characteristic peaks can be the characteristic peaks in the target mass spectrum. Characteristic peaks mainly include molecular ion peaks and fragment ion peaks. Each peak in the mass spectrum often represents an ion signal.

[0065] As an example, the acquisition and recognition module 101 can specifically implement the following steps:

[0066] The first step is to use a mass spectrometer to collect the mass spectrum of the respiratory sputum sample from the patient to be tested, and record it as the target mass spectrum.

[0067] The second step involves using mass spectrometry data processing software to identify the characteristic peaks in the target mass spectrum and their corresponding isotope peak clusters, and then recording the characteristic peaks in the target mass spectrum as target characteristic peaks.

[0068] It should be noted that sputum examination is performed on patients to clarify the nature of the sputum, its cellular composition, and any possible pathogens, which helps doctors diagnose asthma, assess the condition, and formulate treatment plans. Mass spectrometry uses lasers to excite the proteins, nucleic acids, and lipids of pathogens, causing them to ionize and fly. The resulting spectrum, or mass spectrum, is formed based on the mass-to-charge ratio (m / z). The core principle is to identify different substances by utilizing the differences in their mass-to-charge ratio (m / z).

[0069] Sputum samples from patients can be obtained and pre-processed to remove interfering components as much as possible, highlighting pathogen signals. These interfering components may include mucus, epithelial cells, inflammatory cells, and impurities. Mass spectrometry is then used to obtain characteristic spectra. Each peak on the mass spectrum represents one or a class of ions with a specific mass-to-charge ratio (m / z). The peak position (the horizontal axis of the mass spectrum) corresponds to the ion's mass-to-charge ratio, which can be used to identify the type of substance. Different substances have specific mass-to-charge ratios. Relative abundance (the vertical axis of the mass spectrum) reflects the relative content of ions and can be used for semi-quantitative analysis; the stronger the peak, the higher the content of the corresponding substance.

[0070] The obtained mass spectra can also be analyzed using peak detection algorithms (detection algorithms based on signal-to-noise ratio and peak shape characteristics). A characteristic peak is a peak in a mass spectrum that represents a specific molecule or fragment; its mass-to-charge ratio (m / z) reflects key components in sputum (such as proteins, metabolites, or pathogen-related molecules). During mass spectrometry detection, proteins are typically cleaved into small peptides using enzymes such as trypsin. In mass spectrometry, any characteristic peak may correspond to small peptides of different host proteins or pathogens. Mass spectrometry data is essentially a relationship between mass-to-charge ratio and intensity, but converting it into biomedical information about sputum components often requires multiple levels of transformation: the first level is the directly detected molecular signal, and the second level is the biological entity corresponding to these signals (such as specific pathogen proteins). However, in the characteristic peak detection results of mass spectrometry, direct component identification can lead to the masking of some characteristic peaks due to the overlap of mass-to-charge ratios of different components. Therefore, this embodiment of the invention performs masking analysis to facilitate the subsequent segmentation of submerged pathogen features.

[0071] The index determination module 102 is used to determine the pathogen conformity index corresponding to each target feature peak based on the curvature distribution of different points on the fitting curve corresponding to each target feature peak.

[0072] It should be noted that, in addition to pathogens, sputum often contains a large number of endogenous substances, the characteristic peaks of which may overlap with the characteristic peaks of pathogens in the same m / z region, thus masking or misinterpreting the target peak. This is because the concentration of host proteins (such as mucins and immunoglobulins) in patient sputum can reach 10 times that of pathogen proteins. 4 This can lead to host protein masking, which manifests in mass spectrometry as follows: low-abundance pathogen characteristic peaks are masked by high-intensity host protein peaks; furthermore, some pathogens share sequence similarities with human homologous proteins, resulting in similar peptide mass-to-charge ratios (m / z), making characteristic peaks difficult to distinguish. Strong host protein peaks often cause local baseline elevation (due to ion suppression effects), while the signal-to-noise ratio (S / N) of low-abundance peaks decreases. In addition, the half-width at half-maximum (HWHM) of strong peaks may increase (due to signal saturation), while weak pathogen protein peaks are usually sharp but masked by the baseline, resulting in broad plateau peaks formed by host proteins that mask the narrow peaks of neighboring pathogens.

[0073] As an example, the pathogen compliance index corresponding to each target feature peak can be determined based on the maximum value of the curvature corresponding to all coordinate points on the fitted curve corresponding to each target feature peak, and the standard deviation of the curvature corresponding to all coordinate points on the fitted curve corresponding to each target feature peak.

[0074] For example, the formula for determining the pathogen that meets the target characteristic peak can be:

[0075] ;

[0076] in, It is the pathogen conformity index corresponding to the x-th target characteristic peak. x is the sequence number of the target characteristic peak. is the maximum curvature of all coordinate points on the fitted curve corresponding to the x-th target feature peak. k is the maximum curvature of all coordinate points on the fitted curve corresponding to all target feature peaks. It is the standard deviation of the curvature of all coordinate points on the fitted curve corresponding to the x-th target feature peak.

[0077] It should be noted that in reality, pathogen peaks are usually very sharp, with a large internal curvature, and due to their narrow width, the curvature variation across the entire peak range is often significant. In contrast, host protein peaks typically have smaller curvature and less variation in curvature. Therefore, when A larger value often indicates that the curvature within the x-th target characteristic peak is likely to be larger and the curvature change is likely to be larger. This often indicates that the x-th target characteristic peak is more likely to characterize the pathogen and that protein flooding is less likely to have occurred.

[0078] The probability determination module 103 is used to determine the probability of protein submersion corresponding to each target feature peak based on the intersection between the standard protein feature peak corresponding to the pre-acquired standard host protein and each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak.

[0079] The standard host protein can be a human protein from a library of historically known protein types. The standard protein characteristic peak can be the characteristic peak in the mass spectrum corresponding to the standard host protein. The full width at half maximum (FWHM) of the target characteristic peak can be the mass-to-charge ratio width corresponding to half of the peak intensity.

[0080] As an example, determining the probability of protein flooding corresponding to each target feature peak may include the following steps:

[0081] The first step is to select the matching protein characteristic peaks corresponding to each target characteristic peak from all the standard protein characteristic peaks, based on the mass-to-charge ratio ranges corresponding to all standard protein characteristic peaks and the mass-to-charge ratio range corresponding to each target characteristic peak.

[0082] For example, the method for screening matching protein characteristic peaks corresponding to a target characteristic peak may include the following sub-steps:

[0083] The first sub-step involves identifying any target characteristic peak as the labeled characteristic peak, and determining the length of the intersection between the mass-to-charge ratio range corresponding to each standard protein characteristic peak and the mass-to-charge ratio range corresponding to the labeled characteristic peak as the target intersection length between each standard protein characteristic peak and the labeled characteristic peak.

[0084] The second sub-step involves selecting the standard protein characteristic peak with the largest target intersection length with the aforementioned labeled characteristic peak from all standard protein characteristic peaks, and using it as the matching protein characteristic peak corresponding to the aforementioned labeled characteristic peak.

[0085] The second step, determining the protein flooding probability corresponding to each target feature peak based on the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, may include the following sub-steps:

[0086] The first sub-step is to determine the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, and thus determine the matching intersection length for each target feature peak.

[0087] The second sub-step is to determine the length of the matching peak corresponding to each target feature peak by determining the length of the mass-to-charge ratio range of the matching protein feature peaks corresponding to each target feature peak.

[0088] The third sub-step determines the protein flooding probability corresponding to each target feature peak based on the difference between the matching intersection length and the matching peak length corresponding to each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak.

[0089] For example, the formula for determining the probability of protein flooding corresponding to a target feature peak can be:

[0090] ;

[0091] in, This represents the probability of protein saturation corresponding to the x-th target characteristic peak. x is the index of the target characteristic peak. It is the full width at half maximum (FWHM) of the x-th target characteristic peak. It is an exponential function with the natural constant as its base. It is an absolute value function. It is the matching intersection length corresponding to the x-th target feature peak. It is the matching peak length corresponding to the x-th target feature peak. It is the pathogen conformity index corresponding to the xth target characteristic peak.

[0092] It should be noted that when A larger value often indicates a larger half-width at half-maximum (WHM) of the x-th target characteristic peak, which often suggests a greater likelihood of protein immersion. It can characterize the deviation between the x-th target characteristic peak and the host protein standard peak; the larger the value, the more likely protein flooding has not occurred. A larger value generally indicates a greater likelihood of curvature and curvature variation within the x-th target characteristic peak, suggesting that the x-th target characteristic peak is more likely to characterize the pathogen and that protein flooding is less likely to have occurred. Therefore, when The larger the value, the more likely protein saturation is to occur within the x-th target characteristic peak.

[0093] The rationality determination module 104 is used to determine the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), front width, and back width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak.

[0094] Here, "atoms number" represents the total number of atoms. "Peak intensity" represents the relative abundance. "Leading edge width" represents the width of the rising edge of the peak. "Falling edge width" represents the width of the falling edge of the peak.

[0095] It should be noted that due to their high concentration, the host protein's mass spectrum peak intensity is often much higher than that of the pathogen's characteristic peak, causing the pathogen peak to be "submerged" (i.e., the pathogen peak's m / z overlaps with the host protein peak, or is masked by the strong signal of the host peak). However, the isotope peak clusters and multi-charge peak clusters of the host protein and pathogen exhibit pattern differences. Isotope peaks refer to a group of characteristic peak clusters formed by the presence of different isotopes (such as carbon-12 and carbon-13) in the same compound. In reality, atoms in compound molecules often naturally possess stable isotopes, and the mass differences between different isotopes will manifest as a small shift in the mass-to-charge ratio (m / z) in the mass spectrum, thus forming a series of secondary peaks surrounding the "main peak".

[0096] As an example, determining the reasonableness of the isotope peak distribution corresponding to each target feature peak may include the following steps:

[0097] The first step is to determine the initial intensity representative value for each isotope peak by multiplying the number of atoms corresponding to each isotope peak by the peak intensity.

[0098] The second step is to determine the ratio between the initial intensity representative values ​​of each pair of isotope peaks in the isotope peak cluster as the initial intensity ratio between each pair of isotope peaks in the isotope peak cluster.

[0099] For example, the formula for determining the initial intensity ratio between different isotope peaks in an isotope peak cluster can be:

[0100] ;

[0101] in, It represents the initial intensity ratio between the i-th and j-th isotope peaks in the isotope peak cluster. i and j are the indices of the different isotope peaks in the isotope peak cluster. It is the initial intensity representative value corresponding to the i-th isotope peak in the isotope peak cluster. It is the initial intensity representative value corresponding to the j-th isotope peak in the isotope peak cluster. It is the number of atoms corresponding to the i-th isotope peak in the isotope peak cluster. It is the number of atoms corresponding to the j-th isotope peak in the isotope peak cluster. It is the peak intensity corresponding to the i-th isotope peak in the isotope peak cluster. It is the peak intensity corresponding to the j-th isotope peak in the isotope peak cluster.

[0102] It should be noted that the peak intensity ratio only reflects the proportion of intensity and may be affected by external factors such as fluctuations in ionization efficiency and matrix suppression. However, "physical calibration" of intensity using the "number of isotopic atoms" correlates the intensity with the actual number of isotopes in the molecule, more closely approximating the inherent laws of isotopic distribution. Therefore, It can characterize the ratio of the initial intensity between the i-th isotope peak and the j-th isotope peak.

[0103] The third step is to determine the characteristic values ​​of the front and rear edge widths of each isotope peak based on the front and rear edge widths of each isotope peak.

[0104] For example, the ratio between the leading edge width and the trailing edge width of each isotope peak can be determined as the characteristic value of the leading and trailing edge widths of each isotope peak.

[0105] The fourth step is to determine the target intensity ratio representative value between two isotope peaks in the isotope peak cluster based on the difference between the full width at half maximum (FWHM) of each pair of isotope peaks, the initial intensity ratio between each pair of isotope peaks, and the difference between the front and rear edge width eigenvalues ​​of each pair of isotope peaks.

[0106] For example, the formula for determining the representative value of the target intensity ratio between different isotope peaks in an isotope peak cluster can be:

[0107] ;

[0108] in, This represents the target intensity ratio between the i-th and j-th isotope peaks in the isotope peak cluster. i and j are the indices of the different isotope peaks in the isotope peak cluster. It is the initial intensity ratio between the i-th isotope peak and the j-th isotope peak in the isotope peak cluster. It is an exponential function with the natural constant as its base. It is an absolute value function. It is the full width at half maximum (FWHM) of the i-th isotope peak in the isotope peak cluster. It is the full width at half maximum (FWHM) of the j-th isotope peak in the isotope peak cluster. It is the characteristic value of the leading and trailing edge widths corresponding to the i-th isotope peak in the isotope peak cluster. It is the characteristic value of the leading and trailing edge widths corresponding to the j-th isotope peak in the isotope peak cluster. It is the front width corresponding to the i-th isotope peak in the isotope peak cluster. It is the front width corresponding to the j-th isotope peak in the isotope peak cluster. It is the trailing edge width corresponding to the i-th isotope peak in the isotope peak cluster. It is the trailing edge width corresponding to the j-th isotope peak in the isotope peak cluster.

[0109] It should be noted that in reality, all isotopes in the same molecule often undergo the same ionization, transport, and detection processes in mass spectrometry because they are in the same protein / peptide biomolecule. Therefore, the peak shapes (half-width at half maximum, symmetry) are often highly similar. The peak shapes between isotopes should be consistent. Furthermore, combining the peak shapes can supplement the intensity ratio of isotope peak clusters. Starting from the unity of molecular origin, the reliability of the isotope peak attribution can be verified. It can characterize the ratio of the initial intensity between the i-th isotope peak and the j-th isotope peak. It can characterize the consistency of the full width at half maximum (FWHM) between the i-th isotope peak and the j-th isotope peak. The smaller the value, the more consistent the FWHM between the i-th isotope peak and the j-th isotope peak tends to be. This value can characterize the symmetry and consistency between the i-th and j-th isotope peaks. A smaller value generally indicates a more consistent peak shape between the i-th and j-th isotope peaks, meaning they are more likely to originate from the same molecule, which is more helpful in determining their intensity ratio. Therefore, It can characterize the final intensity ratio between the i-th isotope peak and the j-th isotope peak.

[0110] The fifth step, determining the theoretical proportional representative value corresponding to each target characteristic peak based on the target intensity ratio representative values ​​among the isotope peaks in the isotope peak clusters corresponding to all standard protein characteristic peaks, may include the following sub-steps:

[0111] The first sub-step involves selecting the matching protein characteristic peak corresponding to each target characteristic peak from all standard protein characteristic peaks.

[0112] The second sub-step involves determining the average of the target intensity ratio representative values ​​among all isotope peaks in the isotope peak cluster corresponding to the matching protein characteristic peak for each target characteristic peak, and using this average value as the theoretical ratio representative value for each target characteristic peak.

[0113] The sixth step is to determine the rationality of the distribution of isotope peaks corresponding to each target characteristic peak based on the difference between the representative value of the target intensity ratio among isotope peaks in the isotope peak cluster corresponding to each target characteristic peak and its corresponding theoretical representative value.

[0114] For example, the formula for determining the rationality of the isotope peak distribution corresponding to the target characteristic peak can be:

[0115] ;

[0116] ;

[0117] in, This refers to the rationality of the isotope peak distribution corresponding to the x-th target characteristic peak. x is the index of the target characteristic peak. It is an exponential function with the natural constant as its base. is the number of isotope peaks in the isotope peak cluster corresponding to the x-th target characteristic peak. a and b are the sequence numbers of the different isotope peaks in the isotope peak cluster corresponding to the x-th target characteristic peak. It is an absolute value function. It is the representative value of the target intensity ratio between the a-th isotope peak and the b-th isotope peak in the isotope peak cluster corresponding to the x-th target characteristic peak. It is the theoretical proportional representative value corresponding to the xth target characteristic peak. It is the average of the target intensity ratios among all isotope peaks in the isotope peak cluster corresponding to the x-th target characteristic peak.

[0118] It should be noted that the amino acid composition of host proteins is relatively fixed, and the intensity ratio of isotope peak clusters is often predictable and can be verified using theoretical isotope patterns in host protein databases. Pathogens, however, may contain specific amino acids, and the intensity ratio of their isotope peak clusters may differ slightly from that of host proteins. For example, at the same molecular weight, the nitrogen-15 isotope peak intensity of bacterial proteins may be slightly higher. Therefore, when... The smaller the value, the more likely the intensity ratio of the xth target characteristic peak does not conform to the isotopic pattern of the host protein. This may be the result of the superposition of the host protein peak and the pathogen peak. That is, the strong signal of the host protein masks the pathogen peak, but the isotopic characteristics of the pathogen peak may disrupt the regularity of the host peak cluster, resulting in an overall proportional deviation.

[0119] The screening module 105 is used to screen out protein flooding characteristic peaks from all target characteristic peaks based on the probability of protein flooding and the rationality of isotope peak distribution.

[0120] As an example, filtering out protein flooding characteristic peaks from all target characteristic peaks may include the following steps:

[0121] The first step is to determine the target masking index corresponding to each target characteristic peak based on the probability of protein smothering and the rationality of the isotope peak distribution.

[0122] For example, the formula for determining the target masking index corresponding to the target feature peak can be:

[0123] ;

[0124] in, It is the target masking index corresponding to the x-th target feature peak. x is the index of the target feature peak. It is a normalization function. It represents the probability of protein flooding corresponding to the xth target characteristic peak. It is an exponential function with the natural constant as its base. The rationality of the isotope peak distribution corresponding to the xth target characteristic peak.

[0125] It should be noted that when A larger value often indicates a higher likelihood of protein saturation within the x-th target characteristic peak. When The smaller the value, the less likely the intensity ratio of the x-th target characteristic peak is to deviate from the isotopic pattern of the host protein. This may be the result of the superposition of the host protein peak and the pathogen peak; that is, the strong signal of the host protein is more likely to mask the pathogen peak. However, the isotopic characteristics of the pathogen peak may disrupt the regularity of the host peak cluster, leading to an overall proportional deviation. Therefore, when A larger value often indicates a higher likelihood of protein saturation within the x-th target characteristic peak. Conversely, a smaller value indicates a lower value. The smaller the value, the less likely it is that there is no protein flooding phenomenon within the xth target characteristic peak.

[0126] The second step is to determine the target feature peak as a protein smothering feature peak if the target masking index corresponding to the target feature peak is greater than the preset masking threshold.

[0127] The preset masking threshold can be a pre-set threshold, which can be 0.5. The protein flooding characteristic peak can be a characteristic peak that represents the presence of protein flooding.

[0128] The segmentation and detection module 106 is used to segment the target pathogen peak from the protein flooding characteristic peak, and to detect respiratory sputum components based on the target characteristic peak and the segmented target pathogen peak.

[0129] As an example, the segmentation and detection module can specifically implement the following steps:

[0130] The first step is to isolate the target pathogen peak from the protein flooding characteristic peaks.

[0131] For example, based on the pre-acquired characteristic peaks corresponding to standard pathogens, as well as the protein flooding characteristic peaks and their corresponding target masking indices, the target pathogen peaks characterizing the pathogen can be segmented from the protein flooding characteristic peaks using CVAE (Conditional Variational Autoencoder).

[0132] The standard pathogen can be a pathogen from a pathogen library of historically known pathogen types.

[0133] It should be noted that for protein-masked characteristic peaks, a conditional variational autoencoder (CVAE) can be introduced for assisted reconstruction. The original mass spectrometry signal of the region and its corresponding target masking index can be used as input. Utilizing the CVAE's feature learning capability for complex signals, combined with prior knowledge of the known pathogen's peak shape, the pathogen peak signal that may be masked by the host protein can be reconstructed, thus obtaining the target pathogen peak. Finally, several combinations of characteristic peaks are obtained.

[0134] Optionally, for feature peaks with large target masking indicators, two independent peak clusters, such as host peak cluster and pathogen peak cluster, can be separated by peak shape fitting (e.g., Gaussian decomposition).

[0135] The second step is to detect respiratory sputum components based on the target characteristic peaks and the segmented target pathogen peaks.

[0136] For example, respiratory sputum components can be detected by comparing them with a standard spectral library. Specifically, the obtained target characteristic peaks and segmented target pathogen peaks can be compared with standard spectra in the database. A computer can automatically match the characteristic peaks and provide the most probable compound identification result. The standard spectra can be composed of characteristic peaks of known pathogen types.

[0137] Optionally, machine learning models (such as classifiers based on random forests or deep learning) are also used for training and identification. Specifically, standard spectra of known pathogens in the database are used as training samples. By learning the core features such as the m / z distribution and intensity patterns of different pathogen characteristic peaks, accurate identification of pathogen types in the samples can be achieved. At the same time, a calibration curve is established based on the total peak area (or total intensity) of the characteristic peak combination and the peak area of ​​known concentration standards in the database. This curve can estimate the pathogen load (the number of pathogens per unit volume) in the sample, thereby assisting in the clinical assessment of the severity of infection (a higher load indicates a more severe infection).

[0138] refer to Figure 2 Based on the same inventive concept as the above-described method embodiments, this invention provides a rapid detection method for respiratory sputum components in critically ill patients based on mass spectrometry, comprising the following steps:

[0139] Step S1: Obtain the target mass spectrum corresponding to the respiratory sputum sample of the patient to be tested, and identify the target characteristic peaks and their corresponding isotope peak clusters in the target mass spectrum.

[0140] Step S2: Based on the curvature distribution of different points on the fitted curve corresponding to each target feature peak, determine the pathogen compliance index corresponding to each target feature peak.

[0141] Step S3: Based on the intersection between the standard protein characteristic peaks corresponding to the pre-obtained standard host proteins and each target characteristic peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target characteristic peak, determine the protein submersion probability corresponding to each target characteristic peak.

[0142] Step S4: Determine the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), leading edge width, and trailing edge width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak.

[0143] Step S5: Based on the probability of protein flooding and the rationality of isotope peak distribution, screen out the protein flooding characteristic peaks from all target characteristic peaks.

[0144] Step S6: Identify the target pathogen peak from the protein flooding characteristic peak, and perform respiratory sputum component detection based on the target characteristic peak and the identified target pathogen peak.

[0145] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. For example, as shown... Figure 3 As shown, the computer device 300 includes: a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can perform the aforementioned method for rapid detection of respiratory sputum components in critically ill patients based on mass spectrometry.

[0146] Based on the same inventive concept as the above-described method embodiments, the present invention provides a server, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, causing the device to execute the above-described method for rapid detection of respiratory sputum components in critically ill patients based on mass spectrometry.

[0147] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer program product comprising: computer program code, which, when executed on a computer, causes the computer to perform the above-described method for rapid detection of respiratory sputum components in critically ill patients based on mass spectrometry.

[0148] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer-readable storage medium storing computer program code, which, when executed on a computer, causes the computer to perform the above-described method for rapid detection of respiratory sputum components in critically ill patients based on mass spectrometry.

[0149] In summary, this invention quantifies multiple indicators related to the protein submersion status of characteristic peaks by analyzing the target mass spectra corresponding to respiratory sputum samples, such as pathogen conformity indicators, protein submersion probability, and the rationality of isotope peak distribution. This allows for the screening of protein submersion characteristic peaks, and the segmentation of target pathogen peaks representing submerged pathogens from these peaks. This facilitates subsequent identification of submerged pathogens and improves the accuracy of respiratory sputum component detection.

[0150] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry, characterized in that, The system includes: The acquisition and identification module is used to acquire the target mass spectrum corresponding to the respiratory sputum sample of the patient to be tested, and to identify the target characteristic peaks and their corresponding isotope peak clusters in the target mass spectrum. The indicator determination module is used to determine the pathogen conformity index corresponding to each target feature peak based on the curvature distribution of different points on the fitted curve corresponding to each target feature peak. The probability determination module is used to determine the probability of protein submersion for each target feature peak based on the intersection between the standard protein feature peak corresponding to the pre-acquired standard host protein and each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak. The rationality determination module is used to determine the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), front width, and back width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak. The screening module is used to screen out protein flooding characteristic peaks from all target characteristic peaks based on the probability of protein flooding and the rationality of isotope peak distribution. The segmentation and detection module is used to segment the target pathogen peak from the protein flooding characteristic peak, and to detect respiratory sputum components based on the target characteristic peak and the segmented target pathogen peak.

2. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 1, characterized in that, The step of determining the pathogen compliance index corresponding to each target feature peak based on the curvature distribution of different points on the fitted curve corresponding to each target feature peak includes: Based on the maximum value of the curvature of all coordinate points on the fitted curve corresponding to each target feature peak, and the standard deviation of the curvature of all coordinate points on the fitted curve corresponding to each target feature peak, the pathogen compliance index corresponding to each target feature peak is determined.

3. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 1, characterized in that, The step of determining the protein flooding probability corresponding to each target feature peak based on the intersection between the standard protein feature peaks corresponding to the pre-obtained standard host proteins and each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, includes: Based on the mass-to-charge ratio ranges corresponding to all standard protein characteristic peaks and the mass-to-charge ratio range corresponding to each target characteristic peak, the matching protein characteristic peaks corresponding to each target characteristic peak are selected from all standard protein characteristic peaks. The probability of protein flooding for each target feature peak is determined by the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak.

4. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 3, characterized in that, The step of selecting matching protein characteristic peaks for each target characteristic peak from all standard protein characteristic peaks based on the mass-to-charge ratio ranges corresponding to all standard protein characteristic peaks and the mass-to-charge ratio range corresponding to each target characteristic peak includes: Any target characteristic peak is identified as the labeled characteristic peak, and the length of the intersection between the mass-to-charge ratio range corresponding to each standard protein characteristic peak and the mass-to-charge ratio range corresponding to the labeled characteristic peak is identified as the target intersection length between each standard protein characteristic peak and the labeled characteristic peak. The standard protein characteristic peak with the largest target intersection length with the labeled characteristic peak is selected from all standard protein characteristic peaks and used as the matching protein characteristic peak corresponding to the labeled characteristic peak.

5. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 3, characterized in that, The determination of the protein flooding probability corresponding to each target feature peak based on the length of the intersection between the mass-to-charge ratio range corresponding to each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, includes: The length of the intersection between the mass-to-charge ratio range of each target feature peak and the mass-to-charge ratio range of its corresponding matching protein feature peak is determined as the matching intersection length for each target feature peak. The length of the mass-to-charge ratio range of the matching protein feature peaks corresponding to each target feature peak is determined as the length of the matching peak corresponding to each target feature peak. Based on the difference between the matching intersection length and the matching peak length corresponding to each target feature peak, as well as the full width at half maximum (FWHM) and pathogen compliance index corresponding to each target feature peak, the probability of protein submersion corresponding to each target feature peak is determined.

6. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 1, characterized in that, The determination of the rationality of the isotope peak distribution corresponding to each target characteristic peak based on the number of atoms, peak intensity, full width at half maximum (FWHM), leading edge width, and trailing edge width of the isotope peaks in the isotope peak cluster corresponding to each target characteristic peak includes: The product of the number of atoms corresponding to each isotope peak and the peak intensity is used to determine the initial intensity representative value corresponding to each isotope peak. The ratio between the initial intensity representative values ​​corresponding to every two isotope peaks in the isotope peak cluster is determined as the initial intensity ratio between every two isotope peaks in the isotope peak cluster. Based on the leading edge width and trailing edge width of each isotope peak, determine the characteristic value of the leading and trailing edge widths of each isotope peak. Based on the difference between the half-width at half-maximum (WHM) of each pair of isotope peaks in the isotope peak cluster, the initial intensity ratio between each pair of isotope peaks, and the difference between the leading and trailing edge width eigenvalues ​​of each pair of isotope peaks, the target intensity ratio representative value between each pair of isotope peaks in the isotope peak cluster is determined. Based on the representative values ​​of the target intensity ratios among isotope peaks in the isotope peak clusters corresponding to all standard protein characteristic peaks, determine the theoretical representative value of each target characteristic peak. The rationality of the distribution of isotope peaks corresponding to each target characteristic peak is determined based on the difference between the representative value of the target intensity ratio among isotope peaks in the isotope peak cluster corresponding to each target characteristic peak and its corresponding theoretical representative value.

7. A rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry, as described in claim 6, is characterized in that... The step of determining the theoretical proportional representative value corresponding to each target characteristic peak based on the target intensity proportional representative value among the isotope peaks in the isotope peak clusters corresponding to all standard protein characteristic peaks includes: Select the matching protein characteristic peaks corresponding to each target characteristic peak from all standard protein characteristic peaks; The average of the target intensity ratios among all isotope peaks in the isotope peak cluster corresponding to the matching protein characteristic peak for each target characteristic peak is determined as the theoretical ratio representative value for each target characteristic peak.

8. The rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry according to claim 1, characterized in that, Based on the probability of protein immersion and the rationality of isotope peak distribution, protein immersion characteristic peaks are screened from all target characteristic peaks, including: Based on the probability of protein smothering and the rationality of isotope peak distribution corresponding to each target feature peak, the target masking index corresponding to each target feature peak is determined. If the target masking index corresponding to the target feature peak is greater than the preset masking threshold, then the target feature peak is determined as a protein smothering feature peak.

9. A rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry, as described in claim 6, is characterized in that... The step of determining the characteristic value of the leading and trailing edge widths of each isotope peak based on the leading and trailing edge widths of each isotope peak includes: The ratio between the leading edge width and the trailing edge width of each isotope peak is determined as the characteristic value of the leading and trailing edge widths of each isotope peak.

10. A rapid detection system for respiratory sputum components in critically ill patients based on mass spectrometry, as described in claim 8, is characterized in that... The process of segmenting the target pathogen peak from the protein flooding characteristic peaks includes: Based on the pre-acquired characteristic peaks corresponding to standard pathogens, as well as the protein flooding characteristic peaks and their corresponding target masking indices, the target pathogen peaks characterizing the pathogens are segmented from the protein flooding characteristic peaks using CVAE.

Images