Metagenome sequencing diagnosis method and system for Aspergillus lung infection

Abstract

The invention relates to the technical field of biological information detection, in particular to a metagenome sequencing diagnosis method and system for Aspergillus pulmonary infection, and the method comprises the following steps: acquiring lavage fluid lysis monitoring data, adjusting shear power output, analyzing fluorescence accumulation trend, adjusting temperature control frequency and substrate interval, identifying mutation points and constructing indexes, screening entropy inversion fragments, and obtaining the metagenome sequencing diagnosis result of Aspergillus pulmonary infection. According to the method, through dynamic matching of the lysate flow rate and the shearing power, the nucleic acid release balance is improved, the temperature control frequency and the substrate interval are adjusted according to the fluorescence signal accumulation trend in the amplification stage, the reaction equilibrium state is established, and the amplification deviation is corrected; and identifying a mutation boundary and reconstructing an index hierarchy according to basic group joint distribution and feature vector change, judging and screening feature sections in combination with information entropy inversion, establishing a probability convergence structure based on coverage depth and repetition frequency, and realizing stable identification of pathogen features and convergence output of classification confidence.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of biological information detection, in particular to a macro-genome sequencing diagnosis method and system for pulmonary aspergillosis. BACKGROUND

[0002] The technical field of biological information detection involves the use of principles of cross-disciplines such as biology, informatics and computer science, through high-throughput sequencing, biochips, mass spectrometry, microfluidic detection and other means, to collect, identify and analyze biological information such as nucleic acids, proteins and metabolites, to achieve the purposes of disease diagnosis, genetic analysis and pathogen detection, etc. Specifically, it includes extraction of genetic material in biological samples, nucleic acid sequence determination, data cleaning and alignment, sequence classification and annotation, and bioinformatics analysis. By combining experimental detection and information calculation, comprehensive molecular identification of complex samples is achieved, providing molecular-level detection basis for clinical medicine, environmental monitoring and public health. Among them, the traditional macro-genome sequencing diagnosis method for pulmonary aspergillosis refers to a detection method that extracts total nucleic acid from a patient's lung sample and uses a high-throughput sequencing platform to sequence all microbial genetic material in the sample. Typically, bronchoalveolar lavage fluid, sputum or lung tissue is used as the detection object. By chemical lysis and centrifugal separation, host cells are removed, and microbial DNA or RNA in the sample is extracted. After fragmentation, ligation of sequencing adapters and amplification purification through library construction reaction, the obtained sequencing data is subjected to quality control processing and host sequence removal. Using alignment algorithms, the sequences are matched with a microbial database to identify the presence or absence of Aspergillus and its species in the sample. This method relies on the cooperation of experimental sequencing and information analysis to achieve macro-genome level detection of pulmonary aspergillosis.

[0003] Traditional technology mainly relies on chemical lysis and centrifugal separation to release microbial nucleic acid in the macro-genome sequencing diagnosis of pulmonary aspergillosis. The uneven distribution of lysis energy leads to insufficient integrity of nucleic acid extraction, affecting the representativeness of subsequent sequencing fragments. In the amplification reaction stage, fixed temperature and substrate concentration are usually maintained, which cannot dynamically correct the reaction differences between different templates, causing the signal of low-abundance pathogens to be masked. Sequence alignment relies on fixed databases and single matching rules, lacks precise identification of sequence mutations and information entropy changes, resulting in fuzzy positioning of Aspergillus characteristic sequences and biased pathogen identification results, unstable detection sensitivity and confidence. SUMMARY

[0004] To solve the technical problems existing in the prior art, the present application provides a macro-genome sequencing diagnosis method for pulmonary aspergillosis, comprising the following steps:

[0005] S1: Obtain the lysis reaction monitoring data of the perfusion liquid, analyze the flow characteristics and the dispersion state of the microparticles in the lysis liquid, compare the synchronism of the liquid flow rate variation trend and the shear equipment output power, adjust the output curve gradient of the shear power, and generate the lysis energy distribution coordination parameter;

[0006] S2: Based on the lysis energy distribution coordination parameter, analyze the fluorescence signal intensity accumulation trend, compare the consistency of the amplification rate slope and the standard curve, judge the synchronous relationship of the reaction rate and the substrate trajectory, adjust the temperature control frequency and the substrate addition interval, and generate the amplification reaction balance information;

[0007] S3: According to the amplification reaction balance information, analyze the base composition difference of the amplification fragment, construct a feature vector and identify a mutation point, draw a mapping boundary line, reconstruct an index structure level, calculate a vector correlation coefficient between fragments, and generate a sequence index mapping coefficient;

[0008] S4: Based on the sequence index mapping coefficient, construct a fixed length sliding window, count the joint frequency distribution of continuous adjacent base combinations in the window and calculate the information entropy value, according to the entropy value difference of adjacent windows, filter the reverse fragments, adjust the window span, calculate the segment information increment index, and generate a set of Aspergillus characteristic identification fragments;

[0009] S5: According to the set of Aspergillus characteristic identification fragments, analyze the coverage depth sequence and the repetition frequency, judge the matching of uniformity and stability, adjust the weight calculation structure, correct the deviation direction of the rule classification probability, analyze and adjust the convergence process of the probability distribution, calculate the classification probability variation trend of the pathogenic area, and generate a classification confidence convergence index.

[0010] As a further scheme of the present application, the lysis energy distribution coordination parameter includes shear energy gradient coverage range, liquid phase particle motion density matching value, and power output response continuity, the amplification reaction balance information specifically refers to fluorescence signal accumulation rate consistency, enzyme reaction and substrate consumption synchronous trajectory, and amplification curve smoothness index, the sequence index mapping coefficient includes base joint distribution weight vector, mutation point aggregation tightness, and index level correlation coefficient, and the set of Aspergillus characteristic identification fragments specifically refers to entropy value change reverse fragment, window information increment comparison result, and continuous distribution structure mapping fragment, and the classification confidence convergence index includes sequencing coverage balance coefficient, probability deviation correction parameter, and pathogen distribution convergence trend.

[0011] As a further scheme of the present application, the acquisition step of the lysis energy distribution coordination parameter specifically refers to:

[0012] S101: Obtain the perfusion liquid lysis reaction monitoring data, analyze the flow characteristics and the particle dispersion state of the lysis liquid, compare the liquid flow rate change trend with the synchronism of the shear equipment output power, judge the stability degree of the shear energy distribution in the sample, and generate a shear energy distribution parameter;

[0013] S102: According to the shear energy distribution parameter, the local energy offset region position and the change direction are identified, the gradient of the shear power output curve is adjusted, the power coverage ratio of each region of the lysis cavity is optimized, and the power ratio coordination configuration is established.

[0014] S103: Based on the power ratio coordination configuration, the matching relationship between the liquid phase particle motion density and the shear power change trend is calculated, the corresponding relationship between the particle motion density change and the shear power fluctuation of each region is analyzed, and the lysis energy distribution coordination parameter is obtained.

[0015] As a further scheme of the present application, the acquisition step of the amplification reaction balance information is specifically:

[0016] S201: Based on the lysis energy distribution coordination parameter, the intensity change data of the fluorescence signal in the equal time interval in the amplification stage is collected, the cumulative trend of the signal intensity with time is analyzed, the average amplitude consistency of the amplification curve and the standard amplification curve is compared, the difference ratio between the two is calculated and the cumulative signal sequence is recorded, and the fluorescence signal cumulative trend parameter is generated.

[0017] S202: According to the fluorescence signal cumulative trend parameter, the change direction of the product generation rate curve and the enzymatic reaction rate curve is analyzed, the continuity of the substrate consumption track and the synchronization section of the product generation rate are compared, the phase offset degree between the enzymatic reaction and the substrate reaction is judged, the dynamic difference ratio between the two is calculated, and the enzyme substrate reaction synchronization coefficient is generated.

[0018] S203: The enzyme substrate reaction synchronization coefficient is called, the heating unit temperature control frequency and the substrate addition interval of the liquid supply unit are adjusted, and the amplification reaction balance information is generated.

[0019] As a further scheme of the present application, the acquisition step of the sequence index mapping coefficient is specifically:

[0020] S301: According to the amplification reaction balance information, the base arrangement data of the sequence fragments after amplification is obtained, the composition ratio of the base types in each fragment is analyzed, the joint appearance frequency of adjacent base pairs is calculated, the base joint distribution matrix is established, the numerical set of base joint probability in the fragment is extracted, and the base joint distribution parameter is generated.

[0021] S302: Construct a feature vector of each sequence segment based on the base joint distribution parameter, compare the numerical difference of the feature vectors between adjacent segments, determine the continuity breakpoint of the weight change direction, delimit the mutation point as the demarcation line of the mapping section, calculate the distribution trend of the segment weight difference in the demarcation section, and generate a structural mutation demarcation coefficient;

[0022] S303: Call the structural mutation demarcation coefficient, analyze the segment aggregation tightness of each mapping section, reconstruct the hierarchical distribution of the index structure according to the aggregation degree, compare the correlation degree of the segment feature vectors between multiple levels, calculate the correlation coefficient of the vectors between adjacent segments, establish the corresponding relationship between the index level and the correlation coefficient, and generate a sequence index mapping coefficient.

[0023] As a further scheme of the present application, the acquisition step of the Aspergillus feature recognition segment set is specifically:

[0024] S401: Based on the sequence index mapping coefficient, a fixed length sliding window is constructed, the joint frequency distribution of consecutive adjacent base combinations in each window is counted, the distribution set of base combinations in each window is obtained, and a base joint frequency parameter is generated;

[0025] S402: According to the base joint frequency parameter, the information entropy value of each window is calculated, the information entropy values between adjacent windows are compared, the entropy value change direction is judged, the position where the information entropy change direction is reversed is identified, the entropy change reversal segment is screened, and the information entropy reversal segment interval is obtained.

[0026] S403: According to the information entropy reversal segment interval, the stability degree of the information entropy difference value on both sides of the reversal point is compared, the window span is adjusted by considering the continuity of the data distribution in the segment, the segment information increment index is calculated, and an Aspergillus feature recognition segment set is generated.

[0027] As a further scheme of the present application, the process of identifying the position where the information entropy change direction is reversed and screening the entropy change reversal segment is specifically:

[0028] The information entropy sequence constructed based on the base joint frequency parameter is subjected to continuous difference, the information entropy increase and decrease direction between adjacent sliding windows is calculated, and the positive and negative change points in the continuous difference result are trend judged, and the position where the positive difference changes to negative difference or the negative difference changes to positive difference is selected as the information entropy reversal candidate point;

[0029] The information entropy values of each two windows before and after each information entropy inversion candidate point are acquired, the absolute values of the mean difference of the information entropy of the windows on both sides of the target inversion candidate point are calculated respectively, the difference is compared with the set information entropy inversion threshold value, the inversion candidate point meeting the condition that the mean difference of the information entropy is greater than the information entropy inversion threshold value is marked as an effective inversion point, and a window is expanded to both sides with each effective inversion point as the center to obtain an information entropy inversion segment interval;

[0030] The information entropy inversion threshold value is obtained by calculating the average of the information entropy difference values between all windows in the information entropy sequence and superimposing the standard deviation of the information entropy difference value set.

[0031] As a further scheme of the present application, the step of obtaining the classification confidence convergence index is specifically:

[0032] S501: According to the Aspergillus characteristic recognition fragment set, the sequencing coverage depth distribution sequence of the pathogenic region is analyzed, the sequencing coverage degree of each fragment is counted, the coverage depth change trend and the sequencing repetition frequency are extracted, and the sequencing depth coverage parameter is generated;

[0033] S502: Based on the sequencing depth coverage parameter, the coverage gradient change amplitude of adjacent fragments is compared, the matching relationship between the sequencing balance and the fragment stability is judged, the coverage fluctuation section is screened, the weight calculation mode is adjusted, and the pathogenic fragment weight coefficient is obtained;

[0034] S503: According to the pathogenic fragment weight coefficient, the deviation direction of the pathogenic region classification probability is corrected, the convergence process of the probability distribution is analyzed, the change trend of the classification probability with time is calculated, the convergence criterion is established, and the classification confidence convergence index is generated.

[0035] The metagenomic sequencing diagnosis system for pulmonary aspergillosis infection comprises:

[0036] The lysis energy coordination module acquires lavage fluid lysis reaction monitoring data, analyzes the flow characteristics and particle dispersion state of the lysis liquid, compares the synchronism of the liquid flow rate change trend and the shear equipment output power, adjusts the output curve gradient of the shear power, and generates a lysis energy distribution coordination parameter;

[0037] The amplification reaction regulation module analyzes the fluorescence signal intensity accumulation trend based on the lysis energy distribution coordination parameter, compares the consistency of the amplification rate slope and the standard curve, judges the synchronization relationship between the reaction rate and the substrate trajectory, adjusts the temperature control frequency and the substrate addition interval, and generates amplification reaction balance information.

[0038] The sequence mapping reconstruction module analyzes the base composition difference of the amplification fragments according to the amplification reaction balance information, constructs a feature vector and identifies a mutation point, draws a mapping boundary line, reconstructs an index structure hierarchy, calculates a fragment inter-vector correlation coefficient, and generates a sequence index mapping coefficient.

[0039] The feature entropy recognition module constructs a fixed-length sliding window based on the sequence index mapping coefficient, counts the joint frequency distribution of continuous adjacent base combinations in the window and calculates the information entropy value, screens the reverse fragments according to the adjacent window entropy value difference, adjusts the window span, calculates the section information increment index, and generates a set of Aspergillus feature recognition fragments;

[0040] The classification confidence calculation module analyzes the coverage depth sequence and the repetition frequency according to the set of Aspergillus feature recognition fragments, judges the matching of uniformity and stability, adjusts the weight calculation structure, corrects the deviation direction of the rule classification probability, analyzes and adjusts the convergence process of the probability distribution, calculates the classification probability trend of the pathogenic region, and generates a classification confidence convergence index.

[0041] Compared with the prior art, the advantages and positive effects of the present application are that:

[0042] In the present application, by dynamically matching the lysis liquid flow rate and the shear power, the nucleic acid release uniformity is improved, the temperature control frequency and the substrate interval are adjusted according to the fluorescence signal accumulation trend in the amplification stage, the reaction equilibrium state is established, the amplification deviation is corrected, the mutation boundary is recognized and the index hierarchy is reconstructed according to the base joint distribution and the feature vector change, the characteristic section is screened according to the information entropy reverse judgment, the probability convergence structure is established based on the coverage depth and the repetition frequency, and the stable identification of the pathogen feature and the convergence output of the classification confidence are realized. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 It is a step flowchart of the present application.

[0045] Figure 2 It is a S1 refinement diagram of the present application.

[0046] Figure 3 It is a S2 refinement diagram of the present application.

[0047] Figure 4 It is a S3 refinement diagram of the present application.

[0048] Figure 5 It is a S4 refinement diagram of the present application.

[0049] Figure 6 It is a S5 refinement diagram of the present application.

[0050] Figure 7 A system module diagram of the present application. DETAILED DESCRIPTION

[0051] The technical solutions in the present application will be described below with reference to the drawings.

[0052] In the embodiments of the present application, the words such as "example", "for example" are used to represent an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0053] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized. "Of", "corresponding" and "corresponding" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized.

[0054] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1, and the meanings expressed are consistent when the distinction is not emphasized.

[0055] To make the technical problems, technical solutions and advantages to be solved by the present application clearer, specific embodiments will be described in detail below with reference to the drawings.

[0056] Please refer to Figure 1 The present application provides a metagenomic sequencing diagnosis method for pulmonary aspergillus infection, comprising the following steps:

[0057] S1: Obtain the monitoring data of the lavage fluid lysis reaction, analyze the flow characteristics and the dispersion state of the lysis fluid, compare the synchronism of the liquid flow rate change trend and the output power of the shearing equipment, adjust the output curve gradient of the shearing power, and generate lysis energy distribution coordination parameters;

[0058] S2: Based on the lysis energy distribution coordination parameters, analyze the cumulative trend of the fluorescence signal intensity, compare the consistency of the amplification rate slope and the standard curve, judge the synchronous relationship between the reaction rate and the substrate trajectory, adjust the temperature control frequency and the substrate addition interval, and generate amplification reaction balance information;

[0059] S3: According to the amplification reaction balance information, the base composition difference of the amplification fragment is analyzed, the feature vector is constructed, the mutation point is identified, the mapping boundary line is drawn, the index structure level is reconstructed, the vector correlation coefficient between fragments is calculated, and the sequence index mapping coefficient is generated;

[0060] S4: Based on the sequence index mapping coefficient, a fixed length sliding window is constructed, the joint frequency distribution of continuous adjacent base combination in the window is counted and the information entropy value is calculated, according to the difference of adjacent window entropy value, the reverse fragment is screened, the window span is adjusted, the section information increment index is calculated, and the aspergillus characteristic identification fragment set is generated.

[0061] S5: According to the aspergillus characteristic identification fragment set, the coverage depth sequence and the repetition frequency are analyzed, the balance and stability matching are judged, the weight calculation structure is adjusted, the deviation direction of the rule classification probability is corrected, the convergence process of the probability distribution is analyzed and adjusted, the classification probability trend of the pathogenic area is calculated, and the classification confidence convergence index is generated.

[0062] The cracking energy distribution coordination parameters include shear energy gradient coverage range, liquid phase particle motion density matching value, power output response continuity, the amplification reaction balance information is specifically fluorescence signal accumulation rate consistency, enzyme reaction and substrate consumption synchronous trajectory, amplification curve smoothness index, sequence index mapping coefficient includes base joint distribution weight vector, mutation point aggregation tightness, index level correlation coefficient, aspergillus characteristic identification fragment set is specifically entropy value change reverse fragment, window information increment comparison result, continuous distribution structure mapping fragment, classification confidence convergence index includes sequencing coverage balance coefficient, probability deviation correction parameter, pathogen distribution convergence trend.

[0063] Please refer to Figure 2 The acquisition steps of the cracking energy distribution coordination parameters are as follows:

[0064] S101: Obtain the lavage liquid cracking reaction monitoring data, analyze the flow characteristics and particle dispersion state of the cracking liquid, compare the liquid flow rate trend with the synchronicity of the shear device output power, judge the stability degree of the shear energy distribution in the sample, and generate the shear energy distribution parameter;

[0065] After obtaining the bronchial lavage sample, it is injected into the microfluidic lysis device, which has 5 monitoring points integrated inside for real-time acquisition of lysis reaction data. First, the liquid flow rate at each monitoring point is obtained by a micro flow sensor, and the flow rate data per second is recorded within the first 10 seconds. For example, when the device output power is constant at 20 watts (W), the average flow rates of the 5 monitoring points (numbered 1 to 5) are 1.50 milliliters per second (mL / s), 1.52 mL / s, 1.45 mL / s, 1.51 mL / s and 1.48 mL / s, respectively. At the same time, the dispersion state of the microparticles in the sample is monitored by a laser scattering turbidimeter, and the turbidity values are recorded as 350 NTU, 355 NTU, 320 NTU, 352 NTU and 340 NTU, respectively. The standard deviation of the turbidity value is 14.8 NTU, indicating that there is uneven dispersion of microparticles near monitoring point 3. Next, the time series data of the shear device output power and the time series data of the liquid flow rate at each monitoring point are collected at intervals of 0.1 seconds for cross-correlation analysis. Taking monitoring point 1 as an example, the cross-correlation coefficient between its flow rate change sequence and the total power output sequence is 0.95; the cross-correlation coefficient of monitoring point 3 is 0.82. Integrating the cross-correlation coefficients of the 5 monitoring points (0.95, 0.94, 0.82, 0.93, 0.90), it is judged that the transmission of shear energy in the sample is uneven, especially in the monitoring point 3 area, the energy distribution stability is low. The cross-correlation coefficients of the 5 monitoring points are arithmetically averaged, and the shear energy distribution parameter is calculated as: (0.95+0.94+0.82+0.93+0.90) / 5=0.908.

[0066] S102: According to the shear energy distribution parameter, identify the local energy deviation area position and change direction, adjust the gradient of the shear power output curve, optimize the power coverage ratio of each area of the lysis cavity, and establish the power ratio coordination configuration;

[0067] According to the shear energy distribution parameter 0.908 and the data of each monitoring point, it is identified that the monitoring point 3 is a local energy deviation area, the flow rate (1.45 mL / s) of which is about 2.8% lower than the average value (1.492 mL / s), and the cross-correlation coefficient (0.82) is significantly lower than that of other points, and the change direction shows that the energy transfer efficiency is reduced. In order to correct this deviation, it is necessary to adjust the shear power of the corresponding area in the lysis chamber. The lysis chamber is controlled by three independent piezoelectric ceramic shear units, which correspond to monitoring points 1-2, monitoring point 3 and monitoring points 4-5 respectively. In the initial state, the total power of 20W is distributed in the three units in the proportion of 40%, 20% and 40%. Now, aiming at the problem of insufficient energy of monitoring point 3, the gradient of the shear power output curve is adjusted, specifically, the power output curve slope of the shear unit controlling the monitoring point 3 area is increased by 15% based on the original, so that it can reach the peak power faster in the response period. At the same time, the power coverage ratio of the three units is optimized, and the total power of 20W is distributed according to the new ratio. The new distribution scheme aims to concentrate more energy in the monitoring point 3 area, and the specific configuration is shown in Table 1.

[0068] Table 1 Shear unit power ratio coordination configuration table

[0069] Shearing unit Corresponding monitoring point Initial power ratio (%) Optimized power ratio (%) 1 1,2 40 35 2 3 20 30 3 4,5 40 35

[0070] As shown in Table 1, by increasing the power ratio of shear unit 2 from 20% to 30%, and at the same time, reducing the power ratio of unit 1 and unit 3 from 40% to 35%, a new power ratio coordination configuration is established.

[0071] S103: Based on the power ratio coordination configuration, the matching relationship between the liquid phase particle motion density and the shear power change trend is calculated, the corresponding relationship between the particle motion density change of each area and the shear power fluctuation is analyzed, and the lysis energy distribution coordination parameter is obtained;

[0072] Based on the power ratio coordination configuration (unit 1: 35%, unit 2: 30%, unit 3: 35%), the lysis operation is re-performed and the data is monitored. At this time, the particle image velocimetry (PIV) analyzer is used to calculate the motion density of the tracer particles in the liquid phase, which is quantified as the average kinetic energy per unit volume. Under the new power configuration, the average kinetic energy of particles at the five monitoring points is 1.25×10 -5 J / mm 3 , 1.27×10 -5 J / mm 3 , 1.28×10 -5 J / mm 3 , 1.26×10 -5 J / mm 3 and 1.24×10 -5 J / mm 3The correlation between the change of particle kinetic energy in each region and the fluctuation of power of the corresponding shear cell is analyzed, i.e. the Pearson correlation coefficient of the two sets of time series data is calculated. For example, for monitoring point 3, the correlation coefficient of the change sequence of particle kinetic energy and the power fluctuation sequence of shear cell 2 is calculated as 0.98. Similarly, the correlation coefficients of other monitoring points and their corresponding unit power fluctuations are calculated, and the correlation coefficients of monitoring points 1, 2 and unit 1 are 0.97, and the correlation coefficients of monitoring points 4, 5 and unit 3 are 0.96. Finally, these correlation coefficients are weighted and averaged, and the weight is determined according to the number of monitoring points covered by each unit. The weight of unit 1 is 2 / 5, the weight of unit 2 is 1 / 5, and the weight of unit 3 is 2 / 5. Thus, the cracking energy distribution coordination parameter is obtained: 0.97×(2 / 5)+0.98×(1 / 5)+0.96×(2 / 5)=0.388+0.196+0.384=0.968.

[0073] Please refer to Figure 3 The acquisition step of the amplification reaction equilibrium information is specifically:

[0074] S201: Based on the cracking energy distribution coordination parameter, collect the intensity change data of the fluorescence signal in the amplification stage at equal time intervals, analyze the cumulative trend of the signal intensity with time, compare the average amplitude of the amplification curve with the amplitude consistency of the standard amplification curve, calculate the difference proportion between the two and record the cumulative signal sequence, and generate a fluorescence signal cumulative trend parameter;

[0075] Based on the cracking energy distribution coordination parameter 0.968, the subsequent metagenome amplification process is started. In the exponential phase of amplification, the fluorescence signal intensity data is collected every 30 seconds using a fluorescence detection system, and 20 cycles are continuously collected. For example, from the 15th cycle to the 34th cycle, the relative fluorescence unit (RFU) sequence recorded is: {150, 210, 295, 415, …, 16500}. The cumulative trend of the signal intensity with time is analyzed, i.e. an amplification curve is drawn. At the same time, a pre-set standard amplification curve database is retrieved, which contains amplification data for specific target points of Aspergillus under ideal conditions. Select a standard curve with an initial template concentration closest to the estimated concentration of the current sample as a reference, and its average amplitude (growth rate of RFU value per cycle) in the same cycle interval is 40%. Calculate the average amplitude of the current sample amplification curve during 15 to 34 cycles, if the calculation result is 35%, the difference proportion between the two is (40%-35%) / 40%=12.5%. The difference proportion and the collected original cumulative signal sequence {150, 210, …, 16500} are recorded together to generate a fluorescence signal cumulative trend parameter, which is a composite data structure containing the difference proportion and the signal sequence.

[0076] S202: According to the fluorescence signal accumulation trend parameters, the change direction of the product generation rate curve and the enzymatic reaction rate curve is analyzed, the continuity of the substrate consumption trajectory and the synchronization section of the product generation rate are compared, the phase shift degree between the enzymatic reaction and the substrate reaction is judged, the dynamic difference ratio between the two is calculated, and the enzyme-substrate reaction synchronization coefficient is generated;

[0077] According to the fluorescence signal accumulation trend parameters, the intrinsic kinetics of the amplification reaction is further analyzed. First, by first-order difference of the cumulative signal sequence, the product generation rate curve is obtained, which reflects the increase of fluorescent product in each time interval, and the change direction is initially increasing and then tends to be flat. At the same time, based on the standard Michaelis equation model, combined with the initial substrate (dNTPs and primer) concentration (for example, the initial concentration of dNTPs is 200 μM), the theoretical enzymatic reaction rate curve is simulated, which also presents the trend of first increasing and then decreasing. Compare the peak time of the two curves, if the product generation rate peak appears in the 28th cycle, and the theoretical enzymatic reaction rate peak appears in the 26th cycle, it is judged that there is a phase shift between the two. Further, by monitoring the real-time concentration change of free dNTPs in the reaction system, the substrate consumption trajectory is drawn. Compare the continuity of the trajectory with the synchronization section of the product generation rate, it is found that during the 20th to 25th cycle, the substrate consumption rate and the product generation rate show high synchronization, the correlation coefficient is 0.99, and after the 26th cycle, the correlation coefficient decreases to 0.85, indicating that the activity of the enzyme is limited. Calculate the phase shift degree of the entire exponential period, calculate the distance between the two rate curves by dynamic time warping (DTW) algorithm, get a dynamic difference ratio. For example, the calculated dynamic difference ratio is 0.18, which is the enzyme-substrate reaction synchronization coefficient.

[0078] S203: Call the enzyme-substrate reaction synchronization coefficient, adjust the heating unit temperature control frequency and the substrate addition interval of the liquid supply unit, and generate amplification reaction balance information;

[0079] The enzyme substrate reaction synchronization coefficient is 0.18. The coefficient exceeds the preset stable reaction threshold 0.15, indicating that the reaction is unstable. Accordingly, the control system of the amplification instrument is adjusted. First, the temperature control frequency of the heating unit is adjusted, and the original temperature control precision of ± 0.1℃ is improved to ± 0.05℃, and the temperature rise and fall rate is optimized through the PID algorithm to reduce the temperature overshoot. Second, according to the mismatch phenomenon of the substrate consumption trajectory after the 26th cycle, the substrate addition interval of the liquid supply unit is adjusted, and 10 μL of dNTPs mixed solution with a concentration of 50 μM is additionally supplemented at the end of the 25th cycle. After the adjustment is completed, a verification amplification experiment is carried out, and the amplification curve data is re-collected. The new amplification curve is subjected to Savitzky-Golay smoothing filtering, and the sum of the absolute values of the second derivative is calculated to quantify the smoothness of the curve. For example, the smoothness index before adjustment is 25.6, and after adjustment it is reduced to 10.2. The smoothness index 10.2 is output as a reaction stability parameter, and together with the adjusted temperature control parameters and liquid supply strategy, it generates amplification reaction balance information for subsequent steps.

[0080] See Figure 4 , the sequence index mapping coefficient acquisition step is specifically:

[0081] S301: According to the amplification reaction balance information, the base arrangement data of the amplified sequence fragments is obtained, the composition proportion of the base types in each fragment is analyzed, the joint occurrence frequency of adjacent base pairs is calculated, the base joint distribution matrix is established, the numerical set of base joint probability in the fragment is extracted, and the base joint distribution parameter is generated;

[0082] According to the amplification reaction balance information, the amplification product is subjected to high-throughput sequencing. The base arrangement data of the sequence fragments (reads) generated by sequencing is obtained. 1000 sequence fragments with a length of 150 bp are randomly selected for analysis. For one of the fragments, for example “ATGCCG…TAG”, first analyze the base composition proportion, and count to get A: 28%, T: 30%, G: 22%, and C: 20%. Next, calculate the joint occurrence frequency of adjacent base pairs in the fragment, that is, count the number of occurrences of all 16 kinds of dinucleotides such as ‘AT’, ‘AG’, ‘AC’, ‘TA’, ‘TG’, … ‘CC’. For example, ‘AT’ appears 15 times, ‘GC’ appears 12 times, etc. Organize these 16 frequency values into a 4x4 matrix, i.e. a base joint distribution matrix. Perform this operation on the 1000 fragments to obtain 1000 base joint distribution matrices. Extract the numerical values from each matrix to form a 16-dimensional vector, which is the numerical set of base joint probability of the fragment. The 1000 vector sets are collected as the output of this step to generate the base joint distribution parameter.

[0083] S302: Constructing a feature vector for each sequence segment based on the base joint distribution parameters, comparing the numerical difference of the feature vectors between adjacent segments, judging the continuity breakpoint of the weight change direction, delimiting the mutation point as the demarcation line of the mapping section, calculating the distribution trend of the segment weight difference in the demarcation section, and generating the structural mutation demarcation coefficient;

[0084] Based on the base joint distribution parameters (1000 16-dimensional vectors), a feature vector is constructed for each sequence segment. The 1000 segments are sorted according to their original positions on the genome. Then, the numerical difference of the feature vectors between adjacent segments is compared. Here, the comparison uses the Euclidean distance between the two vectors. For example, the feature vectors of segment i and segment i+1 are V i and V i+1 , and the numerical difference between them is D(i,i+1) = |V i -V i+1 |. The distance between all adjacent segments is calculated to obtain a distance sequence {D(1,2), D(2,3), …, D(999,1000)}. Assuming that the sequence has a sudden increase at D(500,501), for example, from an average of 0.2 to 0.8, it is determined that this is a continuity breakpoint of the weight change direction, i.e. a structural mutation point. With this mutation point as the boundary, the sequence is divided into two mapping sections (segment 1-500 and segment 501-1000). The distribution trend of the segment weight difference in each demarcation section is calculated, for example, the standard deviation of the distance values in the section. If the standard deviation of the first section is 0.05 and the second section is 0.06, the two values and the position information of the mutation point (segment 501) are combined to generate the structural mutation demarcation coefficient, represented as {position:501, section standard deviation:[0.05,0.06]}.

[0085] S303: Calling the structural mutation demarcation coefficient, analyzing the segment aggregation tightness of each mapping section, reconstructing the hierarchical distribution of the index structure according to the aggregation degree, comparing the correlation of the segment feature vectors between multiple levels, calculating the correlation coefficient of the vectors between adjacent segments, establishing the corresponding relationship between the index level and the correlation coefficient, and generating the sequence index mapping coefficient;

[0086] The mutation structure mutation boundary coefficient divides the sequence index structure into two main levels according to the mutation point position 501. Within each level, the aggregation tightness of the fragments is further analyzed. This process uses the K-means clustering algorithm to cluster the fragment feature vectors within each mapping section. For example, in the first mapping section (fragments 1-500), set K=3, and aggregate the 500 fragments into 3 sub-clusters. These 3 sub-clusters constitute the next level of the index structure. The level distribution of the index structure is reconstructed according to the compactness of the vectors within each cluster (e.g., within-cluster sum of squares). Then, the correlation of the fragment feature vectors between different levels is compared, specifically the cosine similarity between the centroid vectors of different clusters is calculated. For example, the cosine similarity between the centroid vectors of cluster 1 and cluster 2 is 0.75. Finally, the average value of the correlation coefficient between the vectors of adjacent fragments within each cluster is calculated. For example, the average correlation coefficient of adjacent fragments within cluster 1 is 0.95. A corresponding relationship between the index level (e.g., "main level 1-cluster 1") and the correlation coefficient (0.95) is established. All these corresponding relationships are integrated to generate the sequence index mapping coefficient.

[0087] Referring to Figure 5 The acquisition steps of the Aspergillus feature recognition fragment set are as follows:

[0088] S401: Based on the sequence index mapping coefficient, a fixed-length sliding window is constructed, the joint frequency distribution of consecutive adjacent base combinations within each window is counted, the distribution set of base combinations within each window is obtained, and a base joint frequency parameter is generated;

[0089] Based on the sequence index mapping coefficient, one of the clusters with high aggregation is selected for analysis. A sliding window with a fixed length of 100 bases (bp) is constructed, and the window is slid on the representative sequence of the cluster with a step size of 10 bp. For each window, the joint frequency distribution of all consecutive adjacent base combinations (di-base) is counted. For example, for a 100 bp sequence within a window, the number of occurrences of 'AA', 'AC', 'AG', 'AT', …, 'TT' is counted, and the frequency is calculated. If 'AT' occurs 10 times, its frequency is 10 / 99 ≈ 0.101. This operation is performed for each window to obtain a frequency distribution vector {P(AA), P(AC), …, P(TT)}. After the sliding window traverses the entire sequence, a series of window corresponding base combination distribution sets are obtained, which is the base joint frequency parameter.

[0090] S402: According to the base joint frequency parameter, the information entropy value of each window is calculated, the information entropy values between adjacent windows are compared, the entropy value change direction is judged, the position of the information entropy change direction reversal is identified, the entropy change reversal fragments are screened, and the information entropy reversal fragment interval is obtained;

[0091] According to the base joint frequency parameters, the information entropy value of each window is calculated. The calculation method of information entropy is H = -∑ i p i log2(p i ), wherein p i is the frequency of the ith di-base. For example, for a certain window, the information entropy value calculated by the frequency distribution is 3.2 bits. Calculate all the windows to obtain an information entropy sequence {H1, H2, H3, …}. Then, the information entropy sequence is continuously differentiated to calculate ΔH j = H j+1 -H j . Assuming that the difference sequence obtained is {0.1, 0.15, 0.08, -0.05, -0.12, …}, the difference value changes from positive to negative between ΔH3 and ΔH4, so the position of the fourth window is marked as an information entropy inversion candidate point. The information entropy values of the two windows before and after the candidate point are H2, H3, H4, H5, respectively, which are {3.1, 3.18, 3.13, 3.01}. Calculate the absolute value of the mean difference of the information entropy of the two windows: | (3.1+3.18) / 2- (3.13+3.01) / 2 | = |3.14-3.07| = 0.07. To set the information entropy inversion threshold, experimental verification is required. Select 30 cases of patients with confirmed pulmonary aspergillosis and 30 cases of healthy controls of bronchial lavage fluid samples, use this method for sequencing and analysis, calculate the average (for example, 0.01) and standard deviation (for example, 0.02) of the information entropy difference value set of all samples. The information entropy inversion threshold is set to the average plus twice the standard deviation, that is, 0.01+2×0.02=0.05. Since the calculated mean difference 0.07 is greater than the threshold 0.05, the inversion candidate point is marked as an effective inversion point. Take the effective inversion point as the center, expand 5 windows (11 windows in total) to both sides, and the interval {window-1, …, window9} is an information entropy inversion fragment interval.

[0092] S403: According to the information entropy inversion fragment interval, compare the stability of the information entropy difference value on both sides of the inversion point, adjust the window span by considering the continuity of the data distribution in the fragment, calculate the segment information increment index, and generate a set of aspergillus feature recognition fragments;

[0093] According to the multiple information entropy inversion fragment intervals, the stability of each interval is evaluated. The stability of the information entropy difference value on both sides of each effective inversion point (for example, 5 windows respectively) is compared by calculating the standard deviation of the information entropy sequence of each side. If the standard deviation of one side is 0.02 and the standard deviation of the other side is 0.08, it indicates that the data distribution is discontinuous and the fluctuation is large. At this time, the span of the sliding window is adjusted, for example, the window length is increased from 100bp to 120bp, the step is unchanged, and the information entropy sequence of the interval and the stability of the two sides of the inversion point are recalculated. After adjustment, the standard deviations of the two sides are 0.03 and 0.04, and the continuity of the data distribution is optimized. In this optimized fragment interval, the information increment index is calculated. The index is obtained by calculating the KL divergence (Kullback-Leibler divergence) of the segment relative to the whole genome background. If the KL divergence value of a segment is 1.8, which exceeds the preset feature threshold 1.5, the segment is confirmed as a Aspergillus feature fragment. All fragments passing this screening are collected to generate an Aspergillus feature recognition fragment set.

[0094] See Figure 6 The classification confidence convergence index is obtained by the following steps:

[0095] S501: According to the Aspergillus feature recognition fragment set, analyze the coverage depth distribution sequence of the pathogenic region, count the sequencing coverage of each fragment, extract the coverage depth change trend and sequencing repetition frequency, and generate the sequencing depth coverage parameter;

[0096] According to the Aspergillus feature recognition fragment set, these fragments are aligned to the Aspergillus reference genome database to determine their specific regions on the pathogen genome. Analyze the coverage depth distribution sequence of these pathogenic regions. For example, for a feature region with a length of 500bp, count the number of sequencing reads (reads) aligned to each base position in the region. Thus, a coverage depth array with a length of 500 is obtained, such as {35, 38, 40, 36, …, 50}. Extract the coverage depth change trend of the array, for example, calculate the average coverage depth of 42X and the standard deviation of 5X. At the same time, the sequencing repetition frequency of the region is counted, that is, the proportion of completely identical reads due to PCR amplification bias, if the proportion is 8%, it is recorded. The average coverage depth, depth standard deviation, and repetition frequency are integrated to generate the sequencing depth coverage parameter, represented as {average depth: 42, depth standard deviation: 5, repetition frequency: 0.08}.

[0097] S502: Based on the sequencing depth coverage parameter, compare the coverage gradient change amplitude of adjacent fragments, judge the matching relationship between sequencing uniformity and fragment stability, screen the coverage fluctuation section, adjust the weight calculation method, and obtain the pathogenic fragment weight coefficient;

[0098] Based on the sequencing depth coverage parameters of each pathogen fragment, the stability of the fragment is evaluated. The coverage gradient change amplitude of adjacent fragments is compared, that is, the difference between the average coverage depth of one fragment and the average coverage depth of the next fragment. If the average coverage depths of two adjacent fragments are 42X and 25X respectively, the gradient change amplitude is large, indicating that the sequencing uniformity is poor. The matching relationship between the sequencing uniformity and the stability of the fragment is judged, and a rule is set: if the depth standard deviation of the fragment is greater than 20% of the average depth, or the depth difference with the adjacent fragment exceeds 50%, the fragment is considered to be of low stability. All segments with large coverage fluctuations are screened out and are given a weight reduction in the weight calculation. The weight calculation method is as follows: the initial weight is 1, and for the low-stability fragment, the weight is multiplied by a penalty factor. The penalty factor is set according to the severity of the fluctuation, for example, for a fragment with a depth standard deviation of 25% of the average depth, the penalty factor is 0.8. After adjustment, each pathogen fragment obtains a weight coefficient, for example, the weight of a stable fragment is 1.0, and the weight of an unstable fragment is 0.8, and a list of pathogen fragment weight coefficients of all fragments is obtained.

[0099] S503: According to the pathogen fragment weight coefficient, the deviation direction of the pathogen region classification probability is corrected, the convergence process of the probability distribution is analyzed, the change trend of the classification probability with time is calculated, the convergence criterion is established, and the classification confidence convergence index is generated;

[0100] According to the list of pathogen fragment weight coefficients, the initial classification result is corrected. In the initial classification, each feature fragment may be given the same weight to determine whether Aspergillus exists in the sample. Now, the overall classification probability of the pathogen region is corrected by weighted average. For example, assuming that there are two fragments, the initial classification probability of fragment A is 0.99 (tending to Aspergillus), and the weight is 1.0; the initial classification probability of fragment B is 0.95, but the unstable weight is adjusted to 0.7. The corrected classification probability is (0.99x1.0+0.95x0.7) / (1.0+0.7)=0.974. The convergence process of this probability distribution is analyzed as more weighted fragments are included in the calculation. The change trend of the total classification probability is calculated after each fragment is added. The convergence criterion is established: when the total classification probability changes by less than 0.001 after 5 fragments are added in succession, it is considered that the probability has converged. The final probability value at this time, for example 0.985, is the classification confidence. The confidence 0.985 is output as the final classification confidence convergence index. The result shows that the confidence of the existence of Aspergillus pathogen in the sample is 98.5%.

[0101] Please refer to Figure 7 , the metagenomic sequencing diagnosis system for pulmonary aspergillosis, comprising:

[0102] The lysis energy coordination module acquires the lavage liquid lysis reaction monitoring data, analyzes the flow characteristics and the particle dispersion state of the lysis liquid, compares the synchronism of the liquid flow rate change trend and the shear equipment output power, adjusts the output curve gradient of the shear power, and generates the lysis energy distribution coordination parameter;

[0103] The amplification reaction regulation module analyzes the fluorescence signal intensity accumulation trend based on the lysis energy distribution coordination parameter, compares the consistency of the amplification rate slope and the standard curve, judges the synchronous relationship of the reaction rate and the substrate trajectory, adjusts the temperature control frequency and the substrate addition interval, and generates the amplification reaction balance information.

[0104] The sequence mapping reconstruction module analyzes the amplification fragment base composition difference according to the amplification reaction balance information, constructs a feature vector and identifies a mutation point, draws a mapping boundary line, reconstructs an index structure level, calculates the inter-fragment vector correlation coefficient, and generates a sequence index mapping coefficient.

[0105] The feature entropy recognition module constructs a fixed-length sliding window based on the sequence index mapping coefficient, calculates the information entropy value of the joint frequency distribution of the continuous adjacent base combination in the window, selects the reverse fragment according to the adjacent window entropy value difference, adjusts the window span, calculates the segment information increment index, and generates a set of Aspergillus feature recognition fragments.

[0106] The classification confidence calculation module analyzes the coverage depth sequence and the repetition frequency according to the set of Aspergillus feature recognition fragments, judges the matching of the uniformity and the stability, adjusts the weight calculation structure, corrects the deviation direction of the rule classification probability, analyzes and adjusts the convergence process of the probability distribution, calculates the classification probability change trend of the pathogenic area, and generates a classification confidence convergence index.

[0107] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A metagenomic sequencing diagnostic method for pulmonary aspergillosis, characterized in that, Includes the following steps: S1: Acquire monitoring data of the pyrolysis reaction of the rinsing fluid, analyze the flow characteristics and particle dispersion state of the pyrolysis fluid, compare the synchronicity between the liquid flow rate change trend and the output power of the shearing device, adjust the output curve gradient of the shearing power, and generate pyrolysis energy distribution coordination parameters. S2: Based on the aforementioned lysis energy distribution coordination parameters, analyze the cumulative trend of fluorescence signal intensity, compare the consistency between the amplification rate slope and the standard curve, determine the synchronization relationship between the reaction rate and the substrate trajectory, adjust the temperature control frequency and substrate addition interval, and generate amplification reaction equilibrium information. S3: Based on the amplification reaction equilibrium information, analyze the differences in base composition of the amplified fragments, construct feature vectors and identify mutation points, delineate mapping boundaries, reconstruct the index structure hierarchy, calculate the vector correlation coefficient between fragments, and generate sequence index mapping coefficients; S4: Based on the sequence index mapping coefficient, a fixed-length sliding window is constructed, the joint frequency distribution of consecutive adjacent base combinations within the window is statistically analyzed and the information entropy value is calculated. Based on the entropy value difference between adjacent windows, inverted fragments are selected, the window span is adjusted, the segment information increment index is calculated, and a set of Aspergillus genus feature recognition fragments is generated.

2. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 1, characterized in that, The lysis energy distribution coordination parameters include shear energy gradient coverage, liquid phase particle motion density matching value, and power output response continuity. The amplification reaction equilibrium information specifically includes the consistency of fluorescence signal accumulation rate, the synchronous trajectory of enzyme-catalyzed reaction and substrate consumption, and the amplification curve smoothness index. The sequence index mapping coefficients include the base joint distribution weight vector, the aggregation density of mutation points, and the correlation coefficient between index levels. The Aspergillus genus feature recognition fragment set specifically includes entropy change reversal fragments, information increment comparison results between windows, and continuous distribution structure mapping fragments.

3. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 1, characterized in that, The specific steps for obtaining the pyrolysis energy distribution coordination parameters are as follows: S101: Acquire monitoring data of the pyrolysis reaction of the rinsing fluid, analyze the flow characteristics and particle dispersion state of the pyrolysis fluid, compare the synchronicity between the liquid flow rate change trend and the output power of the shearing device, determine the stability of the shear energy distribution in the sample, and generate shear energy distribution parameters. S102: Based on the shear energy distribution parameters, identify the location and direction of change of local energy offset regions, adjust the gradient of the shear power output curve, optimize the power coverage ratio of each region of the fracture cavity, and establish a power ratio coordination configuration. S103: Based on the power ratio coordination configuration, calculate the matching relationship between the liquid phase particle motion density and the shear power change trend, analyze the correspondence between the particle motion density change and the shear power fluctuation in each region, and obtain the pyrolysis energy distribution coordination parameters.

4. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 3, characterized in that, The specific steps for obtaining the amplification reaction equilibrium information are as follows: S201: Based on the lysis energy distribution coordination parameter, collect the intensity change data of the fluorescence signal during the amplification stage within equal time intervals, analyze the cumulative trend of signal intensity over time, compare the consistency between the average increase of the amplification curve and the increase of the standard amplification curve, calculate the difference ratio between the two and record the cumulative signal sequence, and generate fluorescence signal cumulative trend parameters. S202: Based on the fluorescence signal accumulation trend parameter, analyze the changing direction of the product generation rate curve and the enzyme-catalyzed reaction rate curve, compare the continuity of the substrate consumption trajectory with the synchronous segment of the product generation rate, determine the phase shift between the enzyme-catalyzed reaction and the substrate reaction, calculate the dynamic difference ratio between the two, and generate the enzyme-substrate reaction synchronization coefficient. S203: Call the enzyme substrate reaction synchronization coefficient, adjust the temperature control frequency of the heating unit and the substrate addition interval of the liquid supply unit, and generate amplification reaction equilibrium information.

5. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 4, characterized in that, The steps for obtaining the sequence index mapping coefficients are as follows: S301: Based on the amplification reaction equilibrium information, obtain the base arrangement data of the amplified sequence fragment, analyze the composition ratio of base types in each fragment, calculate the frequency of joint occurrence of adjacent base pairs, establish a base joint distribution matrix, extract the numerical set of base joint probabilities in the fragment, and generate base joint distribution parameters. S302: Based on the base combination distribution parameters, construct the feature vector of each sequence segment, compare the numerical differences of the feature vectors between adjacent segments, determine the continuity interruption point of the weight change direction, delineate the mutation point as the boundary line of the mapping segment, calculate the distribution trend of the segment weight difference within the boundary segment, and generate the structural mutation boundary coefficient. S303: Call the structural mutation boundary coefficient to analyze the fragment aggregation density of each mapping segment, reconstruct the hierarchical distribution of the index structure based on the degree of aggregation, compare the correlation of fragment feature vectors between multiple levels, calculate the correlation coefficient of vectors between adjacent fragments, establish the correspondence between index levels and correlation coefficients, and generate sequence index mapping coefficients.

6. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 5, characterized in that, The specific steps for obtaining the Aspergillus genus feature recognition fragment set are as follows: S401: Based on the sequence index mapping coefficient, a fixed-length sliding window is constructed, the joint frequency distribution of consecutive adjacent base combinations within each window is statistically analyzed, the distribution set of base combinations within each window is obtained, and base joint frequency parameters are generated. S402: Based on the base combination frequency parameter, calculate the information entropy value of each window, compare the information entropy values ​​between adjacent windows, determine the direction of entropy change, identify the position where the direction of information entropy change is reversed, filter the entropy change reversal segment, and obtain the information entropy reversal segment interval. S403: Based on the information entropy reversal segment interval, compare the stability of the information entropy difference on both sides of the reversal point, adjust the window span by considering the continuity of data distribution within the segment, calculate the segment information increment index, and generate a set of Aspergillus genus feature recognition segments.

7. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 6, characterized in that, The process of identifying the location where the entropy change direction is reversed and filtering entropy change reversal segments is as follows: The information entropy sequence constructed by the base combination frequency parameters is continuously differentiald, the direction of increase or decrease of information entropy between adjacent sliding windows is calculated, and the positive and negative change points in the continuous differential results are trend-judged. The position that changes from positive differential to negative differential or from negative differential to positive differential is selected as the information entropy reversal candidate point. Obtain the information entropy values ​​of the two windows before and after each information entropy reversal candidate point, calculate the absolute value of the difference between the average information entropy values ​​of the windows on both sides of the target reversal candidate point, compare the difference with the set information entropy reversal threshold, mark the reversal candidate points that satisfy the information entropy average difference being greater than the information entropy reversal threshold as valid reversal points, and expand the window to both sides with each valid reversal point as the center to obtain the information entropy reversal segment interval; The information entropy inversion threshold is obtained by calculating the average of the information entropy differences between all windows in the information entropy sequence and then adding the standard deviation of the information entropy difference set.

8. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 1, characterized in that, The method further includes: S5: Based on the Aspergillus genus characteristic identification fragment set, analyze the coverage depth sequence and repetition frequency, determine the balance and stability matching, adjust the weight calculation structure, correct the offset direction of the rule classification probability, analyze and adjust the convergence process of the probability distribution, calculate the classification probability change trend of the pathogen region, and generate the classification confidence convergence index. The classification confidence convergence indices include sequencing coverage balance coefficient, probability offset correction parameter, and pathogen distribution convergence trend.

9. The metagenomic sequencing diagnostic method for pulmonary aspergillosis according to claim 8, characterized in that, The specific steps for obtaining the classification confidence convergence index are as follows: S501: Based on the Aspergillus genus characteristic identification fragment set, analyze the coverage depth distribution sequence of the pathogen region, count the sequencing coverage of each fragment, extract the coverage depth change trend and sequencing repetition frequency, and generate sequencing depth coverage parameters; S502: Based on the sequencing depth coverage parameters, compare the coverage gradient change amplitude of adjacent fragments, determine the matching relationship between sequencing uniformity and fragment stability, screen coverage fluctuation segments, adjust the weight calculation method, and obtain pathogen fragment weight coefficients. S503: Based on the pathogen fragment weight coefficient, correct the offset direction of the pathogen region classification probability, analyze the convergence process of the probability distribution, calculate the change trend of classification probability over time, establish convergence criteria, and generate a classification confidence convergence index.

10. A metagenomic sequencing diagnostic system for pulmonary aspergillosis, characterized in that, The system is used to implement the metagenomic sequencing diagnostic method for pulmonary aspergillosis according to any one of claims 1-9, the system comprising: The pyrolysis energy coordination module acquires monitoring data of the pyrolysis reaction of the flushing fluid, analyzes the flow characteristics and particle dispersion state of the pyrolysis fluid, compares the synchronicity between the liquid flow rate change trend and the output power of the shearing device, adjusts the output curve gradient of the shearing power, and generates pyrolysis energy distribution coordination parameters. Based on the lysis energy distribution coordination parameters, the amplification reaction regulation module analyzes the cumulative trend of fluorescence signal intensity, compares the consistency between the amplification rate slope and the standard curve, determines the synchronization relationship between the reaction rate and the substrate trajectory, adjusts the temperature control frequency and substrate addition interval, and generates amplification reaction equilibrium information. The sequence mapping reconstruction module analyzes the differences in base composition of amplified fragments based on the amplification reaction equilibrium information, constructs feature vectors and identifies mutation points, delineates mapping boundaries, reconstructs the index structure hierarchy, calculates the vector correlation coefficient between fragments, and generates sequence index mapping coefficients. The feature entropy recognition module constructs a fixed-length sliding window based on the sequence index mapping coefficient, statistically analyzes the joint frequency distribution of consecutive adjacent base combinations within the window and calculates the information entropy value, filters inverted fragments based on the entropy value difference between adjacent windows, adjusts the window span, calculates the segment information increment index, and generates a set of Aspergillus feature recognition fragments. The classification confidence calculation module identifies the fragment set based on the Aspergillus genus characteristics, analyzes the coverage depth sequence and repetition frequency, judges the balance and stability matching, adjusts the weight calculation structure, corrects the offset direction of the rule classification probability, analyzes and adjusts the convergence process of the probability distribution, calculates the classification probability change trend of the pathogen region, and generates a classification confidence convergence index.

Images