A method for evaluating the metabolism of enteric bacteria by a polysaccharide carbon source

Abstract

The application discloses a polysaccharide carbon source intestinal bacteria metabolism evaluation method, and relates to the fields of microbial metabolism determination and stable isotope tracing technology. The same intestinal active flora is anaerobically cultured in parallel with carbon-13 labeled target polysaccharide and unlabeled target polysaccharide to establish a labeled carbon pool set; a fermentation phase is generated according to the cumulative conversion proportion of the target polysaccharide labeled carbon, and different culture batches are aligned; the distributable labeled carbon amount is determined according to the recovered labeled carbon amount; candidate coupling relationships are established according to the carbon-13 incorporation, abundance change, labeled metabolite change and phase sequence of the microbial classification units; the initial carbon flow contribution is projected to a feasible region defined by the distributable labeled carbon amount; and the candidate relationship weight is corrected according to the distribution residual feedback. When the carbon balance is effective and the distribution converges, the target polysaccharide metabolism rate, the microbial classification unit estimated contribution or the candidate metabolic path is output. The application can reduce the fermentation timing mismatch and the excessive attribution of the labeled carbon.

Description

Technical Field

[0001] This invention relates to the field of microbial metabolism assays and stable isotope tracing technology, specifically to a method for evaluating the metabolism of polysaccharide carbon source gut bacteria. Background Technology

[0002] Polysaccharide carbon sources include dietary fiber, resistant starch, pectin, inulin, arabinoxylan, β-glucan, and complex polysaccharides derived from plant or fungal cell walls. With the development of gut microbiota research and functional food development, the evaluation of polysaccharide degradation rates, metabolites, and utilization by gut microbiota has evolved from measuring gas production, pH, and short-chain fatty acids at fermentation endpoints to combining stable isotope tracing, microbiome, and metabolome analysis. Carbon-13 labeled polysaccharides can distinguish the carbon source of the target polysaccharide from the background carbon of the culture medium, and multi-timepoint microbiome and metabolome data can reflect the processes of microbial succession and metabolite formation.

[0003] Current technologies typically collect fermentation samples from different donors or polysaccharides at fixed culture times, then directly splice or input the bacterial abundance and metabolite concentration at the same clock moment into the analysis model. Fermentation batch synchronization technology can adjust the time axis based on process trajectories or key events, while stable isotope metabolic flux technology can apply material balance constraints based on known reaction networks. Both reduce batch misalignment and improve throughput physical consistency, respectively. However, the reaction networks in complex gut microbiota are difficult to obtain completely, and different donors exhibit different initiation times, cross-trophic processes, and product accumulation rates for the same polysaccharide. Aligning only with process curves may not yield positions corresponding to the same carbon conversion degree of the target polysaccharide; applying only metabolic flux constraints from single bacteria or known networks makes it difficult to handle unknown taxa and unresolved carbon flows.

[0004] In the aforementioned scenarios, the same culture time point may correspond to the initiation phase, rapid degradation phase, or plateau phase. Forcing data from these different biological stages to correspond can easily lead to misidentifying common changes caused by temporal misalignment as a coupling relationship between the microbial community and metabolites. On the other hand, using only metabolite concentration changes without verifying the recovery relationship of carbon-13 among parent polysaccharides, liquid products, gaseous products, and cell biomass can cause sampling losses, unmeasured products, and background carbon dilution to cause the model-assigned microbial community contribution to deviate from the actual recoverable target polysaccharide carbon content. Furthermore, phase sequence, abundance changes, and carbon-13 incorporation can only provide evidence of temporal consistency, population response, and substrate origin, respectively. No single piece of evidence is sufficient to directly prove the quantitative contribution of a specific taxonomic unit to the target polysaccharide carbon flux. Therefore, the core problem that existing technologies need to solve is: how to achieve fermentation stage alignment and contribution allocation constraints using the same measured benchmark independent of the complete reaction network, and output verifiable estimated contribution results with clear evidence thresholds, projection algorithms, error propagation, and control validation boundaries. Summary of the Invention

[0005] Technical Objective: The technical problem to be solved by this invention is to reduce the mismatch between microbiota and metabolites caused by fixed-time alignment, suppress the over-allocation of carbon from the source of the target polysaccharide by candidate relationships, and limit the applicable boundary of the estimated carbon flow contribution by using multiple evidence gating, deterministic projection, parallel replication and confidence intervals, so that the alignment of fermentation stages, contribution allocation, error evaluation and anomaly rejection use consistent experimental evidence.

[0006] Technical solution: To achieve the above technical objectives, the present invention adopts the following technical solution:

[0007] A method for assessing the metabolism of intestinal bacteria containing polysaccharide carbon sources is proposed, employing carbon-13 labeled culture and unlabeled control culture of the same sample. The labeled culture is used to obtain the isotopic carbon flux from the target polysaccharide source, while the control culture is used to correct for natural isotope abundance and unlabeled background. During the culture process, at least one fermentation response among gas production, pH, and redox potential is simultaneously acquired at multiple time points. The carbon-13 distribution in residual parent polysaccharide, liquid-phase metabolites, gas-phase products, and bacterial biomass is measured, and the abundance of microbial taxa is determined.

[0008] This invention does not directly use the number of cultivation hours as a common x-axis for different samples. Instead, it uses the same batch of measured carbon-13 data to calculate the cumulative proportion of the target polysaccharide being converted into liquid-phase metabolites, gas-phase products, and cell biomass. This cumulative proportion forms the fermentation phase coordinate. The raw time-series data of each cultivation batch is interpolated to the same fermentation phase anchor point, allowing comparisons between different cultivation batches at the same level of carbon conversion of the target polysaccharide. Therefore, cultivation batches with slower start-up will not be incorrectly correlated with those with faster start-up in terms of clock time.

[0009] After phase alignment, no single indicator of phase sequence, abundance change, or carbon-13 incorporation is directly equated to the contribution of a taxonomic unit. For each taxonomic unit and its relationship with a labeled metabolite, carbon-13 incorporation evidence, abundance response, labeled metabolite generation or consumption coupling, and phase compatibility are constructed. A candidate coupling relationship can only be formed if the carbon-13 incorporation evidence exceeds a background threshold determined by an unlabeled control or an inactivated negative control, and at least one of the other three types of evidence reaches its corresponding threshold. Therefore, "candidate relationship establishment" is limited to relationships with carbon evidence from the target polysaccharide source and supported by independent process evidence, rather than relationships solely based on correlation.

[0010] For gated candidate relations, the four types of evidence are normalized and weighted to form candidate evidence scores. These scores are then multiplied by a carbon scale coefficient determined by the corresponding labeled carbon pool amount or positive reference culture system to obtain the initial carbon flow contribution. All initial carbon flow contributions are projected onto a non-negative simplex feasible region with a sum not exceeding the allocable labeled carbon amount, with the objective of minimizing the Euclidean distance. The projection threshold is uniquely determined by the sorting and summing method. For candidate relations that are reduced by projection, their weights are updated exponentially according to the reduction ratio, and the process is repeated until the allocation residual and weight change simultaneously satisfy the threshold. Thus, candidate relation generation, initial contribution calculation, projection, and feedback updates all have executable inputs, formulas, and stopping conditions.

[0011] The carbon balance residuals are used to determine whether the labeled carbon pool set is sufficient to support quantitative attribution; the allocation residuals are used to determine whether the contribution allocation of candidate relationships converges within the measured boundaries; the coefficient of variation of parallel culture replicates is used to determine experimental repeatability; and the resampling confidence interval is used to represent the uncertainty range of the contribution estimate. Only when the measured carbon balance residuals, allocation residuals, control cultures, repeatability, and confidence interval width all meet the validity conditions will the estimated carbon flow contribution of the core degrading microbial taxa be output and the corresponding batch be allowed to enter parameter updates; if the conditions are not met, only the cause of failure, descriptive indicators, and candidate evidence will be output, and they will not be expressed as validated direct degradation contributions.

[0012] To validate the detection and attribution chains, at least three independent culture replicates were set up for both labeled and unlabeled cultures, with a target polysaccharide-free blank and an inactivated bacterial flora negative control included; a positive reference culture system was set up when a strain with known target polysaccharide utilization was available. Measurement errors and candidate core taxonomic units were validated using at least one of the following methods: spiked recovery of labeled standards, coefficient of variation for technical repeatability, false positive rate of negative controls, detection ranking of positive reference strains, and taxonomic unit-specific quantitative nucleic acid detection, stable isotope probes, or single-cell isotope detection.

[0013] The “estimated carbon flow contribution” in this invention is a model estimate constrained by the measured carbon conservation boundary, multiple evidence gating, and error propagation. It is used to compare the relative contributions of different taxa to the carbon flow from the target polysaccharide under in vitro culture conditions, and does not replace the direct causal evidence obtained by single-cell isolation culture, gene function knockout, or single-cell direct isotope imaging.

[0014] Beneficial effects: The method for evaluating the metabolism of intestinal bacteria using polysaccharide carbon sources provided by this invention has the following beneficial effects:

[0015] First, this invention utilizes the carbon-13 conversion process of the target polysaccharide itself to establish fermentation phase coordinates, enabling culture batches with different fermentation start-up speeds and degradation rates to be compared at the same biological stage, thereby reducing pseudo-coupling relationships caused by clock time misalignment.

[0016] Second, the fermentation phase and contribution constraints are derived from the carbon-13 measurement results of the same set of parent polysaccharides, liquid products, gaseous products and cell biomass, avoiding the introduction of empirical indicators unrelated to the carbon flow of the target polysaccharide for time alignment, and improving the physical consistency between different data layers.

[0017] Third, by projecting the initial carbon flow contribution to the feasible region defined by the amount of labelable carbon that can be allocated, and by adjusting the candidate relationships that lead to over-allocation in reverse with the allocation residuals before and after the projection, instead of simply normalizing all the results at once, the contribution allocation can be gradually brought back to the range of measured labelable carbon support.

[0018] Fourth, by jointly screening candidate edges between target polysaccharides, primary degrading bacteria, labeled intermediate metabolites, and secondary utilizing bacteria through carbon-13 source conditions and phase sequence, the interpretability of metabolic pathways can be improved while maintaining the main line of multi-omics fusion and dynamic evaluation.

[0019] Fifth, by using the measured carbon balance residual to identify experimental carbon recovery failure and by using the allocation residual to identify contribution model non-convergence, it is possible to distinguish between abnormal detection process and abnormal candidate relationship allocation, and prevent the corresponding batch from updating the evaluation parameters.

[0020] Sixth, by using parallel culture replication, inactivated negative control, positive reference culture system, and resampling confidence interval, the detection error, false positives of candidate relationships, and stability of estimation results are controlled respectively, so that the output results simultaneously include point estimation, error range, and independent validation status, avoiding the presentation of correlation evidence as an unrestricted direct contribution conclusion. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0022] Figure 1 This is the overall flowchart of the polysaccharide carbon source gut microbiota metabolism assessment method of the present invention;

[0023] Figure 2 A schematic diagram illustrating the multidimensional dynamic data acquisition relationship between labeled culture, unlabeled control, blank culture, negative control, and positive reference culture;

[0024] Figure 3 A schematic diagram mapping the actual culture time to the fermentation phase of carbon conversion of the target polysaccharide;

[0025] Figure 4 A flowchart illustrating the multi-evidence gating, initial contribution calculation, simplex projection, and weight feedback processes;

[0026] Figure 5 A quality gating flowchart for measuring carbon balance, control validity, repeatability, allocation convergence, confidence interval, and result output. Detailed Implementation

[0027] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. The following embodiments are used to illustrate the technical principles and feasible paths of the present invention, and are not intended to limit the present invention to a certain polysaccharide, bacterial community source, detection instrument or specific model structure.

[0028] Target polysaccharides refer to polysaccharide carbon sources whose intestinal flora metabolic processes need to be evaluated, including homogeneous polysaccharides and complex polysaccharides composed of multiple glycan components. The structural characteristics of target polysaccharides can be represented by at least two of the following: monosaccharide composition, glycosidic bond type, degree of branching, and molecular weight distribution.

[0029] Carbon-13 labeled target polysaccharides refer to target polysaccharides in which all or part of the carbon atoms are enriched with carbon-13. Uniform labeling or site labeling capable of covering the main chain or branches of the target polysaccharide can be used. Labeled and unlabeled target polysaccharides should have comparable molecular weight distribution, monosaccharide composition, and glycosidic bond structure.

[0030] The labeled carbon pool set refers to the collection of labeled carbons from the parent target polysaccharide, liquid-phase metabolites, gas-phase products, bacterial biomass, and those carried over from the sample, all obtainable by measurement at the same sampling time. All labeled carbon pools are uniformly converted to carbon-13 moles.

[0031] Distributable labeled carbon refers to the cumulative 13 moles of carbon that have been transferred from the parent target polysaccharide to the liquid-phase labeled metabolite pool, gas-phase labeled product pool, and cell-labeled carbon pool in a fermentation phase and quantitatively recovered. This amount constitutes the upper limit of the cumulative carbon flow contribution of all microbial taxa within the corresponding phase, and the portion not explained by candidate relations is retained as unresolved carbon flow.

[0032] Fermentation phase coordinates are dimensionless coordinates determined by the cumulative proportion of target polysaccharide-labeled carbon converted from the parent polysaccharide pool to other labeled carbon pools, rather than the number of hours elapsed since the start of cultivation. The same fermentation phase indicates that the target polysaccharide carbon has reached a similar degree of conversion.

[0033] A microbial taxonomic unit refers to a group of microorganisms identified through nucleic acid sequences, specific probes, metagenomic genome assembly, or other reproducible identification methods. Its hierarchy can be amplicon sequence variants, operational taxonomic units, species, genera, or a group of strains with shared metabolic functions. The same taxonomic level is used for each comparison.

[0034] The estimated carbon flux contribution of a core degrading microbial taxa refers to the proportion of the target polysaccharide-derived carbon flux allocated to a specific taxa after multi-evidence gating and feasible region projection, under carbon conservation constraints, relative to the total allocated labeled carbon flux in the corresponding fermentation phase. This value is a relative contribution estimate under in vitro culture conditions and needs to be interpreted in conjunction with its confidence interval, candidate evidence level, and independent validation status. It is not directly equivalent to the taxa's actual contribution in vivo or direct causal evidence.

[0035] The carbon balance residual refers to the difference between the initially added labeled carbon and the measured labeled carbon in the parent target polysaccharide, liquid metabolites, gaseous products, bacterial biomass, and sample carry-out amounts. It is used to evaluate the effectiveness of experimental carbon recovery. The allocation residual refers to the difference between the initial carbon flow contribution of candidate coupling relationships and the carbon flow contribution obtained after projecting onto the feasible region defined by the amount of allocable labeled carbon. It is used to evaluate whether the contribution allocation has converged. The repeatability coefficient of variation is the ratio of the standard deviation to the mean of parallel culture replicates at the same phase anchor point. The contribution confidence interval is the contribution distribution interval obtained after resampling the measurement errors of culture replicates, the labeled carbon pool, and the amount of candidate evidence. The above indicators characterize the measurement layer error, allocation layer error, repeatability error, and estimation uncertainty, respectively, and are not interchangeable.

[0036] Candidate evidence levels include at least strong evidence, moderate evidence, and insufficient evidence. Carbon-13 incorporation exceeding the control background threshold is a necessary condition for the formation of a candidate relationship. In addition, strong evidence is defined as meeting two or more of the following criteria: abundance response, labeled metabolite coupling, and phase compatibility. Moderate evidence is defined as meeting one of these criteria. If none of these criteria are met, the relationship is defined as insufficient evidence and is excluded from the candidate relationship set.

[0037] like Figure 1 As shown, the present invention includes sample division and parallel culture, multi-time point dynamic acquisition, natural abundance correction and labeling carbon pool establishment, fermentation phase generation, phase sequence alignment, candidate coupling relationship generation, hierarchical fusion under carbon conservation constraints, quality gating, and result output.

[0038] Fresh fecal microbiota, artificial intestinal model microbiota, or pre-constructed synthetic microbiota were selected as samples for viable intestinal microbiota. The same thoroughly mixed sample from the same evaluation batch was aliquoted to minimize differences in initial microbiota composition. At least three independent replicates were set up for both labeled and unlabeled control cultures; a blank culture without the target polysaccharide and a negative control of microbiota inactivated by heat treatment, chemical treatment, or irradiation were also included. When a known target polysaccharide-utilizing strain was available, a positive reference culture system for that strain was established to confirm the initial range of labeled substrate availability, detection chain effectiveness, and candidate evidence weights.

[0039] In one embodiment, fresh feces are placed in an anaerobic environment, and a pre-reduction buffer is added at a mass-to-volume ratio of 1:5 to 1:20. After homogenization, large particles are removed by passing the solution through a filter with a pore size of 100 to 500 micrometers. The initial bacterial load is determined by the total bacterial 16S ribosomal RNA gene copy number, flow cytometry count, or bacterial dry weight.

[0040] The basal culture medium contains a nitrogen source, inorganic salts, vitamins, reducing agents, and buffer components, but does not contain fermentable polysaccharides that can replace the target polysaccharide as the main carbon source. The concentration of the target polysaccharide can range from 0.5 g / L to 10 g / L. The labeled culture system contains carbon-13 labeled target polysaccharide, the control culture system contains an equal mass of unlabeled target polysaccharide, and the blank culture system contains no target polysaccharide. The culture temperature is 35°C to 39°C, the initial pH is 6.2 to 7.2, and the culture is carried out under anaerobic conditions.

[0041] Before culturing, the labeled and unlabeled target polysaccharides underwent consistency checks on monosaccharide composition, molecular weight distribution, and glycosidic bond structure. Spiking recovery tests were also performed on the carbon-13 labeled standards. If the labeling process caused the weight-average molecular weight of the target polysaccharide to deviate by more than 20% from that of the unlabeled target polysaccharide, or if the spiked recovery rate was outside the preset range of 80% to 120%, the corresponding two polysaccharides were not considered as a single parallel evaluation pair.

[0042] like Figure 2 As shown, online or quasi-online sampling was used to collect gas production rate, cumulative gas production, pH, and redox potential. Offline sampling was used to determine the residual amount of target polysaccharides, metabolite concentrations and their isotopic isomer distribution, gaseous carbon-13 carbon dioxide, cell biomass carbon-13, and abundance of microbial taxonomic units.

[0043] The initial sampling time can be set to 0, 2, 4, 6, 8, 12, 18, 24, 36, and 48 hours. This timetable is used to ensure that the first experiment covers the fermentation initiation, rapid conversion, and plateau phases, and is not used as the final alignment coordinates between different donors. Sampling times can be increased based on the online fermentation response: when the rate of change in gas production, pH, or C-13 carbon dioxide release exceeds a preset threshold for two consecutive detection cycles, an additional offline sampling point is added before and after this threshold.

[0044] The residual amount of the parent target polysaccharide can be determined using size exclusion chromatography, high-performance anion exchange chromatography, or a validated colorimetric method. Liquid oligosaccharides, monosaccharides, organic acids, amino acids, and nucleotides can be determined using liquid chromatography-high-resolution mass spectrometry, gas chromatography-mass spectrometry, or nuclear magnetic resonance. Mass spectrometry data are corrected for natural isotopes, tracer purity, and isotope peak overlap. Gas-labeled carbon is obtained using gas chromatography-isotope ratio mass spectrometry or a detection method capable of distinguishing carbon-13 carbon dioxide. For each type of quantitative detection, the linear range, limit of quantitation, and intra-batch coefficient of variation are determined using a standard curve with at least five concentration points; preferably, the coefficient of determination of the standard curve is not less than 0.99, and the coefficient of variation for technical repeatability is not greater than 20%.

[0045] Bacterial pellets and culture supernatants were separated under low-temperature conditions. Biomass-labeled carbon in the bacterial cells could be determined using elemental analysis-isotope ratio mass spectrometry. The abundance of microbial taxa units could be obtained through 16S ribosomal RNA gene amplicon sequencing, metagenomic sequencing, or taxa unit-specific quantitative nucleic acid detection.

[0046] To distinguish between taxa that only show an increase in abundance and those that actually absorb the carbon from the target polysaccharide, stable nucleic acid isotope probes can be used. Nucleic acid was extracted from labeled and control bacterial cultures, centrifuged in a cesium chloride density gradient, and collected into multiple fractions according to buoyancy density. Each fraction was then subjected to 16S ribosomal RNA gene or metagenomic sequencing.

[0047] For taxonomic unit i, the weighted abundance of the high-density fraction in the labeled culture is calculated based on the nucleic acid content of each density fraction and the relative abundance of that taxonomic unit. And calculate the heavy density fractional weighted abundance in the unlabeled control. Evidence of carbon-13 doping According to Confirmed. (In the negative control) The mean plus 2 to 4 times the standard deviation is used as the background threshold. Only when Greater than At this point, evidence of carbon incorporation from the target polysaccharide source is considered to exist for the taxonomic unit. The relative incorporation amount of each taxonomic unit can also be scaled using the total cell carbon-13 content, ensuring that the sum of the incorporations of each taxonomic unit does not exceed the measured molar amount of carbon-13 in the cell-labeled carbon pool.

[0048] The labeled carbon molar amount of each metabolite is calculated from the metabolite concentration, sample volume, number of carbon atoms per molecule, and isotopic isomer ratio corrected for natural abundance. For metabolites with multiple carbon-13 isotopic isomers, the molar amount of each isotopic isomer is multiplied by the number of carbon-13 atoms in it and then summed.

[0049] After the kth sampling, the cumulative sample carry-out amount is calculated based on the culture volume before sampling, the sampling volume, and the measured concentrations of each labeled carbon cell. The carbon pool volume at subsequent time points was converted to the baseline culture volume after sampling volume correction to avoid a systematic decrease in carbon recovery caused by multiple sampling.

[0050] The initial amount of labeled carbon, C0, was determined by the mass of the target polysaccharide, the total carbon mass fraction, and the carbon-13 atomic abundance. The parent target polysaccharide pool, the liquid-phase labeled metabolite pool, the gas-phase labeled product pool, and the cell-labeled carbon pool were denoted as follows: , , and The cumulative amount of labeled carbon carried out by the sampling is recorded as follows: All labeled carbon pool quantities use the same C-13 molar unit, and the standard uncertainty determined by standard curves, spiked recoveries, and technical repeatability is preserved.

[0051] like Figure 3 As shown, within the same culture batch, the actual culture time is still retained in chronological order, but the horizontal axis used for cross-batch comparisons has been changed to the carbon conversion phase of the target polysaccharide. Fermentation Phase Calculate according to the following formula:

[0052]

[0053] When there are labeled metabolites that are not included in the quantification list but can be identified by non-targeted mass spectrometry, the labeled carbon content of the metabolite is included. If a low-abundance marker can only be detected qualitatively but not reliably quantitatively, its occurrence phase is recorded, but it is not added to the denominator or numerator of the phase.

[0054] In one implementation, select Four phase anchor points, 0.10, 0.30, 0.60, and 0.85, were used to represent the initial conversion period, accelerated conversion period, major conversion period, and late conversion period, respectively. Additional phase anchor points can be added based on the degradation characteristics of the target polysaccharide. For slowly degrading polysaccharides that did not reach 0.85 within the observation period, the phase anchor points already reached were retained, and high-phase-missing markers were written into the quality record. Phase interpolation simultaneously propagates the measurement uncertainty of adjacent measured points; interpolated values ​​are not considered independent measured values.

[0055] For any original feature x, piecewise cubic interpolation is performed between two adjacent measured phases to maintain monotonicity; when the feature itself is not monotonic, piecewise linear interpolation is used. Extrapolation of data outside the measured phase range is prohibited. This yields the characteristics of the microbial community, metabolites, fermentation response, and polysaccharide residues for each culture batch at the same phase anchor point.

[0056] Carbon balance residual measured at each sampling time Determined according to the following formula:

[0057]

[0058] In the formula, For the k-th sampling time The measurement of carbon balance residuals, This refers to the initial molar amount of carbon-13 added to the culture system. , , and Sampling time The amount of carbon-13 moles in the parent target polysaccharide pool, liquid-phase labeled metabolite pool, gas-phase labeled product pool, and cell-labeled carbon pool. up to the sampling time The cumulative carbon-13 moles carried over from the sampling and corrected for culture volume. The carbon balance residual is measured to evaluate whether each labeled carbon pool and the cumulative sample carryover can explain the initially added labeled carbon; preferably, a single fermentation phase anchor point. Not greater than 0.20, when two consecutive fermentation phase anchor points correspond to When all values ​​are greater than 0.30, it is considered that the labeled carbon recovery is insufficient to support the quantitative allocation of the estimated carbon flow contribution of the taxonomic unit, and the estimated carbon flow contribution is not output for the corresponding culture batch.

[0059] If the relative deviation between the unlabeled control culture and the labeled culture in any of the following indicators—cumulative gas production, pH change, and residual target polysaccharide—is greater than 30%, then the structural consistency between the labeled and unlabeled target polysaccharides and the batch operation should be checked. If persistent labeled carbon transfer exceeding the limit of quantitation occurs in the inactivated bacterial flora negative control, or if the expected substrate consumption and carbon-13 incorporation are not detected in the positive reference culture system, then the batch of test chains should be marked as invalid. If differences in labeled substrate structure, non-biodegradation, or culture abnormalities cannot be ruled out, the estimated carbon flow contribution of taxa units should not be output.

[0060] Target polysaccharides, microbial taxa, and labeled metabolites are represented as directed nodes. Edges from target polysaccharides to microbial taxa represent direct or primary utilization candidate relationships; edges from microbial taxa to labeled metabolites indicate that the taxa may participate in labeled metabolite generation; edges from labeled metabolites to microbial taxa indicate that the taxa may utilize labeled intermediates. Each candidate edge stores the amount of carbon-13 incorporation evidence, abundance response, labeled metabolite coupling, phase difference, and corresponding threshold, allowing for tracing the reasons for its retention or exclusion.

[0061] Abundance response A of taxon i in fermentation phase q i,q The coupling amount M of the labeled metabolite m was determined by the logarithmic fold change of the labeled culture relative to the unlabeled control and the initial phase. i,m,q The phase compatibility T was determined by multiplying the phase difference of the taxonomic unit carbon-13 incorporation with the rate of formation of the labeled metabolite in the same direction, or by multiplying it with the rate of consumption of the labeled metabolite in the opposite direction; i,m,q according to Confirmed, among which To introduce the initial phase for the carbon-13 taxonomic unit, To mark the phase of metabolites, The phase bandwidth is from 0.05 to 0.20. A i,q and M i,m,q The threshold was determined by the 95th to 99th percentiles of blank culture and negative control.

[0062] Primary degradation candidate taxa must satisfy E i,q >θ E and satisfy A i,q >θ A M i,m,q >θ M and T i,m,q >θ T At least one of the following conditions must be met: Secondary utilization candidate taxa must satisfy the following conditions: the labeled intermediate metabolite appears first, followed by an increase in Ei,q of the taxa, and the corresponding rate of generation of the labeled intermediate metabolite decreases or the rate of consumption increases. Phase sequence and abundance response are only considered as supplementary evidence and do not form candidate relationships independently.

[0063] like Figure 4 As shown, for E i,q A i,q and M i,m,q Linear normalization was performed according to the corresponding control threshold and positive reference value, respectively, converting them to a range of 0 to 1. , and ;T i,m,q It is itself in the range of 0 to 1. For candidate relation e, the candidate evidence score is calculated in the r-th iteration according to the following formula:

[0064] , where β E β A β M and β T All values ​​are not less than 0 and their sum is 1. When there is no positive reference data, 0.40, 0.20, 0.25 and 0.15 can be used as initial weights; when there are positive and negative reference data, under the constraint that the sum is 1, the weight combination that maximizes the distinction between positive and negative reference relationships should be selected.

[0065] The initial carbon flux contribution of candidate relation e in phase q is according to Sure. The carbon scaling factor is used as follows: the relationship between the target polysaccharide and the primary degradation taxa is determined by the sum of the cell-labeled carbon pool and the early oligosaccharide-labeled carbon in the liquid phase; the relationship between the taxa and the labeled metabolites is determined by the corresponding labeled metabolite pool; and the relationship between the labeled metabolites and the secondary utilization taxa is determined by the inter-phase carbon consumption of the labeled metabolites. Each candidate relationship independently forms an initial estimate, therefore their sum is allowed to temporarily exceed the allocatable labeled carbon.

[0066] For phase q, establish the feasible region. And solve for the minimum Euclidean distance projection shown in the following equation:

[0067] .

[0068] when season ;when At that time, Sort by size from largest to smallest and determine a unique projection threshold. , so that:

[0069] ,and .

[0070] Unassigned labeled carbons after projection are retained as unresolved carbon flows and are not forcibly assigned to any microbial taxonomic unit. Compared to one-time scaling, the minimum distance projection prioritizes retaining candidate relationships with higher evidence scores and minimizes the total squared deviation between all contributions and the initial estimate.

[0071] Distributed residuals Reduction ratio of single candidate relationships Determine them according to the following formulas respectively:

[0072]

[0073]

[0074] for Candidate relationships >0, according to Update its relation weights, and with Multiply by S e,q (r) Redetermine the initial carbon flow contribution for the next round; where λ is the feedback step size from 0.10 to 1.00, w min The minimum retention weight is between 0.01 and 0.20. For Candidate relationships with a value of 0 retain their weights unchanged.

[0075] when A phase assignment is considered convergent when the preset convergence threshold is no greater than 0.05 to 0.15 and the maximum relative change in the weights of all candidate relations is no greater than 0.01 to 0.05; the preset number of iterations is 5 to 100. If convergence is not achieved after the preset number of iterations, the phase data, candidate evidence, and unresolved carbon flow are retained, but the estimated carbon flow contribution of the core degrading microbial taxa of that phase is not output.

[0076] The carbon balance residuals and the allocation residuals are processed using different paths: the carbon balance residual exceeding the limit indicates that the measured carbon pool is insufficient, triggering the rejection of the culture batch; the allocation residual exceeding the limit indicates that the candidate relation causes over-allocation, triggering feedback and weight reduction for the candidate relation; the coefficient of variation of parallel replication exceeding the limit indicates that the experimental repeatability is insufficient, triggering the addition of replication or re-testing; the confidence interval is too wide, indicating that the estimation is unstable, and only low-confidence candidate relations are output.

[0077] For each valid culture batch, each independent culture replicate is used as a primary resampling unit, and in each resampling, the standard uncertainty perturbation C for each detection item is adjusted. P C M C G C B E i,q A i,q and M i,m,qThe phase calculation, candidate gating, projection, and feedback update are repeated 500 to 2000 times. The calculation is performed for each resampling result of classification unit i. and with The 2.5 percentile and 97.5 percentile of the distribution are used as the 95% confidence interval.

[0078] When a classification unit has a 95% confidence interval lower limit greater than 0, an interval width not greater than a preset threshold, and its candidate evidence level is strong evidence, it is marked as a high-confidence core degradation candidate classification unit; when the interval lower limit is greater than 0 but the interval is wide or only has general evidence, it is marked as a general-confidence candidate classification unit; when the interval contains 0, the negative control shows the same signal, or the independent verification is inconsistent, it is not listed as a core degradation candidate classification unit.

[0079] The monosaccharide molar ratio, glycosidic bond type, degree of branching, weight-average molecular weight, and dispersion factor of the target polysaccharide are encoded as structural features. The weighted distance between these structural features can be used to determine the similarity between the polysaccharide to be evaluated and existing polysaccharides in the database.

[0080] When the structural similarity is higher than a preset threshold, the evidence normalization upper limit and initial weight values ​​of similar polysaccharides can be transferred, but the estimated carbon flow contribution is still re-determined using the fermentation phase of the current culture batch, the control threshold, and the carbon conservation constraint. When the structural similarity is lower than the preset threshold, existing contribution labels are not transferred; the threshold is determined only using the control data of the current batch, and the estimated carbon flow contribution of the core degrading bacteria is output only after the current polysaccharide covers at least three fermentation phase anchor points.

[0081] like Figure 5 As shown, only batches with valid carbon balance residuals, valid control and reference cultures, parallel culture repeatability dispersion meeting requirements, convergent allocation residuals, complete phase anchors, and sufficient confidence interval width are updated in the database. Abnormal batches retain their original data, candidate evidence, and failure causes, but the model weights and judgment thresholds are not updated.

[0082] Example 1: Metabolic assessment of microbial community inulin labeled with carbon-13

[0083] This embodiment illustrates the complete implementation process using carbon-13 labeled inulin. A 10% (w / v) bacterial suspension was prepared from the same sample of fresh fecal microbiota, mixed thoroughly, and inoculated into three labeled culture replicates, three unlabeled control culture replicates, three inulin-free blank culture replicates, and three inactivated bacterial negative control replicates. 2 g / L of carbon-13 labeled inulin was added to the labeled culture, and 2 g / L of unlabeled inulin was added to the unlabeled control culture. A positive reference culture system was also established when a pre-validated reference strain capable of utilizing inulin was available. All culture systems were cultured anaerobically at 37°C and an initial pH of 6.8.

[0084] Gas production and pH were continuously recorded. Gas, culture supernatant, and cell precipitate were collected at 0, 2, 4, 6, 8, 12, 18, 24, 36, and 48 hours. Size exclusion chromatography was used to determine the molecular weight distribution and residual amount of inulin; liquid chromatography-high resolution mass spectrometry was used to determine labeled fructose oligomers, fructose, lactic acid, and succinic acid; gas chromatography-mass spectrometry was used to determine the concentrations and carbon-13 isotope isomers of acetic acid, propionic acid, and butyric acid; and isotope ratio mass spectrometry was used to determine carbon-13 carbon dioxide and total carbon-13 in the cells.

[0085] Bacterial DNA was extracted and 16S ribosomal RNA gene sequencing was performed. Samples near the 0.30 and 0.60 phases were further sequenced using stable isotope probes. Sampling volume correction and carbon-13 molar conversion were performed as described above to calculate the fermentation phase at each measured time point. Then, community abundance, taxonomic unit carbon-13 incorporation, metabolites, and fermentation response were interpolated to the 0.10, 0.30, 0.60, and 0.85 phases. All calculations were performed separately for the three culture replicates, without averaging the raw data to preserve interculture variability.

[0086] During feature fusion, carbon-13 incorporation, abundance response, and metabolite coupling thresholds were first determined using unlabeled controls and inactivated bacterial colony negative controls. Only relationships with carbon-13 incorporation exceeding the background threshold and having at least one piece of supporting evidence were included in the candidate set. The initial carbon flow contribution was calculated based on the candidate evidence score, and simplex projection and exponential feedback updates were performed. After allocation convergence, culture repeatability and measurement errors were resampled, and the following were output: inulin phase metabolic rate, estimated carbon flow contribution of candidate core degradation taxa, 95% confidence interval, major labeled intermediates, proportion of unresolved carbon flow, and independent validation status.

[0087] To clearly illustrate the computational chain and parameter boundaries of this invention, a complete set of data is provided below. These values ​​do not constitute a limitation on the scope of protection. In a certain culture replicate, the initial amount of added carbon-13 (C0) was measured at a sampling point close to the 0.60 phase as 100.0 μmol, and the parent polysaccharide pool... 40.0 μmol, liquid-phase labeled metabolite pool 30.0 micromolar, gas-phase labeled product cell The concentration was 8.0 micromolar, and the cell-labeled carbon pool was... The cumulative sample carry-out amount is 18.0 micromoles. It is 2.0 micromoles.

[0088] according to Calculate φ(t) k=(30.0+8.0+18.0) / (40.0+30.0+8.0+18.0)=0.5833; mapped to the 0.60 phase anchor point through piecewise interpolation of adjacent measured points. Assignable labeled carbon content. = + + =56.0 micromoles. Measurement of carbon balance residuals. =|100.0-40.0-30.0-8.0-18.0-2.0| / 100.0=0.020, which satisfies the validity condition.

[0089] Four candidate relations e1 to e4, gated by multiple evidence, are defined with normalized carbon-13 incorporation values ​​of 0.90, 0.75, 0.60, and 0.45, abundance response values ​​of 0.80, 0.60, 0.50, and 0.35, labeled metabolite coupling values ​​of 0.85, 0.70, 0.65, and 0.55, and phase compatibility values ​​of 1.00, 0.90, 0.80, and 0.70, respectively. β-coupling is used. E =0.40, β A =0.20, β M =0.25, β T =0.15, and the candidate evidence scores were 0.8825, 0.7300, 0.6225 and 0.4925 respectively.

[0090] When the carbon scaling factor for all four candidate relationships is 30.0 μmol, the initial carbon flow contribution G in the first round is 26.475, 21.900, 18.675, and 14.775 μmol, respectively, totaling 81.825 μmol, exceeding C. R According to the minimum projection distance, τ = (81.825 - 56.000) / 4 = 6.45625 micromoles. After the first round of projection, F are 20.019, 15.444, 12.219 and 8.319 micromoles respectively, with a total of 56.000 micromoles. The residual of the first round of allocation is D = (81.825 - 56.000) / 56.000 = 0.461.

[0091] Using a feedback step size λ=0.50, exponential weighting was performed and recalculated based on the reduction ratio of each candidate relation. The allocation residuals for rounds 2 to 5 were 0.249, 0.131, 0.067, and 0.034, respectively. Round 5 satisfied the convergence threshold of D≤0.05, yielding projected contributions of approximately 19.919, 15.402, 12.236, and 8.443 micromoles, corresponding to estimated contribution percentages of approximately 35.6%, 27.5%, 21.9%, and 15.1%. This example demonstrates that the projection threshold, reduction amount per round, feedback step size, convergence rounds, and final contribution percentage can all be recalculated using explicit formulas.

[0092] For three independent culture replicates, the above calculations were performed separately, and 1000 resampling cycles were conducted, with each culture replicate as the first-level resampling unit. Each resampling cycle perturbed the labeled carbon pool amount and candidate evidence amount according to the measurement standard deviation determined by the standard curve and technical replicates, thereby obtaining the 95% confidence interval for the estimated contribution of each taxonomic unit.

[0093] Under the above culture conditions, three independent labeled culture replicates were analyzed, with two technical replicates for each labeled carbon pool. The coefficients of determination for the standard curves of C-13 quantification of liquid metabolites, gaseous products, and cell biomass were 0.998, 0.996, and 0.997, respectively; the corresponding intra-batch coefficients of variation were no higher than 4.2%, 5.1%, and 5.8%, respectively. The following data are the mean ± standard deviation of the three independent culture replicates, used to further illustrate the complete process of phase calculation, measurement error, contribution allocation, and independent validation.

[0094] Table 1. Data on the labeled carbon pool during the carbon-13 labeled inulin culture process.

[0095]

[0096] Table 1 shows that the three culture replicates all covered phase anchor points of 0.10, 0.30, 0.60, and 0.85, and the measured carbon balance residual E at each measured time point was... B (t k All values ​​were no greater than 0.015, and no batch rejection was triggered. Taking 24-hour data as an example, the average fermentation phase was 0.594. After interpolation to a phase of 0.60 by adjacent sampling points, the amount of labeled carbon C that can be allocated is... R The value was 56.0 μmol. The background increments of carbon-13 in the heavy-density nucleic acid fraction were 0.014 ± 0.006, 0.011 ± 0.005, and 0.009 ± 0.004 for the unlabeled control, inulin-free blank, and inactivated bacterial flora negative control, respectively. The carbon-13 incorporation evidence threshold θ was determined by adding three standard deviations to the mean of the negative controls. E It is 0.021.

[0097] Table 2. Multi-evidence data for 0.60 phase candidate classification units

[0098]

[0099] Evidence for carbon-13 incorporation in taxa U1 through U4 all exceeds θ. E Each of these relationships demonstrated evidence of abundance response, labeled metabolite coupling, or phase compatibility, thus entering the candidate relation set. β-methods were used. E =0.40, β A =0.20, β M =0.25, β T=0.15, carbon scale coefficient κ=30.0 micromoles, feedback step size λ=0.50, allocation residual threshold 0.05, maximum relative change threshold of weight 0.02, and maximum number of iterations 20.

[0100] Table 3. Contribution distribution and independent validation results of candidate taxonomic units

[0101]

[0102] The initial carbon flow contribution of the four candidate relationships in the first round totaled 81.825 μmol, exceeding the CR proportion in 46.1%. After simplex projection and five rounds of feedback updates, the allocation residual decreased to 0.034, and the final total projected contribution was 56.000 μmol, no longer exceeding the measured labeled carbon boundary. The 95% confidence intervals obtained from 1000 stratified resamplings did not cross zero. The specific quantitative PCR enrichment folds of taxonomic units U1, U2, U3, and U4 in the high-density nucleic acid fraction decreased sequentially, consistent with the estimated contribution ranking, with a Spearman rank correlation coefficient of 0.95.

[0103] Table 4 Comparison of Fixed-Time Unconstrained Analysis and the Method of the Present Invention

[0104]

[0105] The above comparison results show that, in this set of example data, after replacing the fixed culture time with the carbon conversion phase of the target polysaccharide, the dispersion of the cross-replication feature decreased; after using carbon-13 source gating, the false inclusion rate of negative control candidate relationships decreased; and after using allocable labeled carbon amount constraints and feedback projection, the total contribution of candidate relationships was limited to the measured labeled carbon boundary.

[0106] Example 2: Cross-substrate phase comparison of pectin and resistant starch

[0107] Independent culture batches were established for the same donor bacterial population, using carbon-13 labeled pectin and carbon-13 labeled resistant starch as target polysaccharides. Labeled cultures, unlabeled control cultures, and blank cultures without the target polysaccharide were included for each polysaccharide. For the pectin batch, galacturonic acid oligomers, acetic acid, and succinic acid were the primary detection components; for the resistant starch batch, maltodextrin, glucose, lactic acid, acetic acid, and butyric acid were the primary detection components.

[0108] Because pectin and resistant starch degrade at different rates, the contribution of the microbial community was not directly compared at fixed time points of 12 or 24 hours. Instead, the φ value for each polysaccharide culture batch was calculated separately, and the taxonomic unit carbon-13 incorporation, labeled metabolite generation, and fermentation response were compared at the same 0.10, 0.30, and 0.60 phases. For resistant starch batches that did not reach the 0.85 phase during the observation period, no 0.85 phase extrapolation was performed.

[0109] The structural features of the two polysaccharides, along with the phase-aligned microbial community-metabolite features, are jointly input into the multi-evidence gating and projection process. If a taxonomic unit only shows similarities at the same clock time, but differs in carbon-13 incorporation, labeled metabolite coupling, or estimated carbon flow contribution at the same fermentation phase, it is not identified as a cross-substrate common core degradation candidate taxonomic unit.

[0110] Example 3: Phase Alignment of Different Donor Microbial Communities

[0111] Fecal microbiota from at least three donors were selected and cultured separately for the same carbon-13 labeled target polysaccharide. All donors used the same culture medium, initial bacterial count, target polysaccharide concentration, and sampling and detection methods. Different initiation and transformation rates were permitted between donors over the actual culture time.

[0112] The fermentation phase for each donor culture batch is calculated separately, and phase interpolation is performed on the multidimensional features of each donor. The estimated carbon flow contribution and 95% confidence interval of core degradation taxa at the same fermentation phase are compared, rather than directly comparing data from the same number of hours. The database simultaneously stores the actual culture time, fermentation phase, candidate evidence, estimated contribution, and confidence interval, ensuring that the output represents both the time required for each donor to reach the same phase and the stability of the estimated contribution of each taxa within the same phase.

[0113] For any donor, if the measured carbon balance residuals at two consecutive phase anchor points are both greater than 0.30, or if the allocation residuals do not converge within a preset number of iterations, then the culture batch of that donor will not participate in the cross-donor contribution ranking, but descriptive results of the target polysaccharide residuals, gas production, and measured labeled metabolites can still be output.

[0114] Example 4: Controlled Experiment and Independent Verification

[0115] High-confidence candidate taxa identified by this method at phases 0.30 or 0.60 were selected. Their absolute copy numbers in the heavy-density nucleic acid fraction were determined using taxa-specific quantitative PCR, or their single-cell carbon-13 enrichment was determined using fluorescence in situ hybridization combined with single-cell isotope detection. The results of these independent detections were compared with the candidate evidence level and estimated carbon flux contribution ranking of this method.

[0116] Negative validation includes an inactivated bacterial colony control, a blank strain without the target polysaccharide, and a reference strain that cannot utilize the target polysaccharide; positive validation includes a reference strain confirmed by single-cell culture to be able to utilize the target polysaccharide. Taxonomic units with carbon-13 inclusion exceeding the background threshold in negative validation are excluded from the core candidate set; if a positive reference strain fails the candidate gating, the labeled substrate, detection limit, and threshold settings are re-examined. For high-confidence candidate taxonomic units, it is preferable to require that the direction of independently detected carbon-13 enrichment is consistent with the estimated contribution ranking.

[0117] Control experiments are not used to simplify all carbon flux in complex microbial communities to contributions from known strains, but rather to verify the availability of labeled substrates, detection chain background, candidate gating false positives, and directional consistency of core candidates. If independent validation results are inconsistent with the estimation results, candidate evidence and inconsistency markers for that taxonomic unit are retained, but its confidence level is lowered, and it is not output as a validated core degradation taxonomic unit.

[0118] Causal chain verification

[0119] By converting the carbon-13 distribution of the same batch of parent polysaccharide, liquid metabolites, gaseous products, and cell biomass into fermentation phase coordinates, and applying them to the time-series data alignment process of different culture batches, the microbial community-metabolite characteristics can be compared under different fermentation rates and polysaccharide structures according to the same target polysaccharide carbon conversion stage, thereby reducing misjudgments of coupling relationships caused by fixed time misalignment.

[0120] The cumulative conversion ratio of carbon-13 can be used for time alignment because this ratio directly characterizes the extent to which the carbon of the target polysaccharide leaves the parent polysaccharide pool and enters metabolites, gases, and cells. After alignment according to this ratio, the same phase corresponds to similar substrate conversion states. Therefore, the comparison of community abundance and metabolite changes no longer depends on the start-up time of different donors.

[0121] By setting the allocatable labeled carbon amount determined by the same set of labeled carbon pools as the upper limit of contribution allocation for taxonomic units, and selectively reducing the weight of candidate relationships that lead to over-allocation based on the reduction ratio generated by the minimum distance projection, this method is applied to the multi-evidence fusion and contribution calculation stages. This allows the estimated carbon flow contribution of core degradation taxonomic units to gradually converge to the measured labeled carbon boundary in scenarios where complex bacterial communities lack a complete reaction network, thereby solving the problem of inconsistency between the total output of the model and the actual target polysaccharide carbon amount.

[0122] The allocatable carbon amount can limit the contribution of taxa because the carbon from the target polysaccharide source absorbed, generated, or transferred by the taxa cannot exceed the carbon from the target polysaccharide source actually detected in the corresponding phase; the allocation residual can identify and correct overallocation because it directly quantifies the amount of carbon reduced after the initial contribution of the candidate relationship exceeds the feasible region; the multi-evidence gating requires that carbon-13 source evidence and at least one independent process evidence be established together to avoid mistaking simple abundance correlation or time sequence as direct degradation; when the measured carbon balance residual is too large, quantitative attribution is rejected, which can avoid misallocating the unmeasured carbon pool or sampling loss to a certain taxa.

[0123] The aforementioned chain of action is connected by a common set of labeled carbon pools and explicit quality gating: the set of labeled carbon pools determines both the biological phase in which different culture batches should be compared and the maximum amount of target polysaccharide source carbon that can be allocated in that phase; multi-evidence gating determines which taxa relationships have the minimum evidence to enter quantitative allocation; projection residuals return candidate relationship weights; parallel replication and resampling further provide the error range of contribution estimation, thus forming a closed loop from dynamic acquisition, phase alignment, candidate relationship generation, feasible region projection, feedback correction, error propagation to result output.

[0124] The output should include at least the target polysaccharide identifier, culture batch identifier, correspondence between actual culture time and fermentation phase, target polysaccharide metabolic rate, recovery of each labeled carbon pool, measured carbon balance residual, allocation residual, estimated carbon flow contribution and 95% confidence interval of core degrading microbial taxa, candidate evidence level, labeled intermediate metabolites, candidate metabolic pathways, proportion of unresolved carbon flow, coefficient of variation for parallel replications, independent validation status, and culture batch validity identifier.

[0125] This invention can be used for screening prebiotic raw materials, evaluating polysaccharides in functional foods, evaluating polysaccharides in animal feed, studying the function of gut microbiota in fermented foods, researching artificial intestinal models, and establishing a database of gut microbiota functional phenotypes. This method outputs the metabolic assessment results and estimated uncertainties of the target polysaccharide by in vitro microbiota, without directly outputting disease diagnostic conclusions or replacing estimated carbon flow contributions with direct causal conclusions not supported by independent experiments.

[0126] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for evaluating the metabolism of intestinal bacteria containing polysaccharide carbon sources, characterized in that, Specifically, the following steps are included: S1. The same intestinal active bacteria sample was placed in a labeled culture system containing carbon-13 labeled target polysaccharide and a control culture system containing unlabeled target polysaccharide for in vitro anaerobic culture, and the residual amount of target polysaccharide, the amount of labeled carbon pool, the abundance of microbial taxa and carbon-13 incorporation information were obtained at multiple sampling times. S2. Using the control culture system to correct the abundance of natural isotopes, a set of labeled carbon pools is established according to the parent target polysaccharide pool, liquid phase labeled metabolite pool, gas phase labeled product pool and cell labeled carbon pool. The fermentation phase is determined according to the cumulative conversion ratio of the target polysaccharide labeled carbon to the other labeled carbon pools, and the amount of labeled carbon that can be allocated in the corresponding fermentation phase is determined according to the amount of labeled carbon recovered from the other labeled carbon pools. S3. Align the detection data of different culture batches according to the fermentation phase, establish candidate coupling relationships based on the carbon-13 incorporation, abundance changes, label metabolite changes and phase sequence of microbial taxonomic units, and determine the initial carbon flow contribution of each candidate coupling relationship. S4. Project the initial carbon flow contribution of the same fermentation phase to the feasible region defined by the non-negative condition and the amount of carbon that can be assigned. Adjust the weights of the candidate coupling relationships according to the difference in contribution before and after projection and repeat the projection until the assignment residuals meet the convergence condition. When the carbon balance measurement is effective, output at least one of the target polysaccharide metabolic rate, the estimated contribution of microbial taxonomic units, and the candidate metabolic pathway.

2. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, The labeled culture system and the control culture system also include a blank culture system that does not contain the target polysaccharide. Each culture system uses the same basal culture medium, initial bacterial count, culture volume and culture conditions.

3. The method for evaluating the metabolism of polysaccharide carbon sources in gut microbiota according to claim 1, characterized in that, It also acquires fermentation response data reflecting the culture process and uses the fermentation response data to help determine the fermentation stage or candidate coupling relationship; the labeled metabolites include target polysaccharide degradation intermediates and fermentation products.

4. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, The normalized values ​​of carbon-13 incorporation, abundance response, labeled metabolite coupling, and phase compatibility of candidate coupling relationship e in fermentation phase q are denoted as follows: , , and ,according to Determine the scores of candidate evidence, among which , , and The corresponding weights are given for each item, and each weight is non-negative and their sum is 1; according to Determine the initial carbon flow contribution, where The carbon scale factor is determined by the corresponding labeled carbon pool amount or the positive reference culture system; when the sum of the initial carbon flow contributions exceeds the distributable labeled carbon amount, it is calculated according to... Projection is performed, where In order to make each The sum equals the projection threshold of the amount of assignable labeled carbon.

5. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, The fermentation phase coordinates are determined according to the following formula: ； In the formula, Sampling time Fermentation phase coordinates, The amount of carbon-13 in the target polysaccharide pool of the parent compound. This represents the amount of carbon-13 moles in the liquid-phase labeled metabolite pool. This refers to the molar amount of carbon-13 in the gas-labeled product pool. The amount of carbon-13 moles in the carbon pool is used to label the bacterial cells.

6. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, Phase anchor points are set for multiple fermentation phases covering different transformation stages of the target polysaccharide. Interpolation is performed between adjacent measured fermentation phases, and data outside the range of measured fermentation phases are not extrapolated.

7. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 6, characterized in that, The measured carbon balance residual is determined according to the following formula: ； In the formula, For the k-th sampling time The measurement of carbon balance residuals, This refers to the amount of carbon-13 added to the culture system. up to the sampling time The amount of carbon-13 carried out from the sample and corrected for volume is as follows: The amount of carbon-13 in the target polysaccharide pool of the parent compound. This represents the amount of carbon-13 moles in the liquid-phase labeled metabolite pool. This refers to the molar amount of carbon-13 in the gas-labeled product pool. The amount of carbon-13 moles in the carbon pool for labeling bacterial cells; When the measured carbon balance residuals at two consecutive fermentation phase anchor points are both greater than the preset threshold, the output of the estimated contribution of the microbial taxa unit for the corresponding culture batch is stopped.

8. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, The bacterial nucleic acids of the labeled culture system and the control culture system were graded according to buoyancy density. The nucleic acid sequence of each density grade was detected, and the carbon-13 incorporation information was determined based on the relative enrichment of the same microbial taxonomic unit in the high density grade.

9. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, The candidate coupling relationships include the relationship between the target polysaccharide and the primary degrading microbial taxa, the relationship between the primary degrading microbial taxa and the labeled intermediate metabolite, and the relationship between the labeled intermediate metabolite and the secondary utilizing microbial taxa; the candidate coupling relationships must at least satisfy the carbon-13 source condition and the corresponding phase sequence condition.

10. The method for evaluating the metabolism of polysaccharide carbon source intestinal bacteria according to claim 1, characterized in that, Quantitative contribution results are output only for culture batches that simultaneously meet the requirements of valid carbon balance measurement, convergent distribution residuals, complete phase data, and valid control culture, along with a culture batch validity identifier.

Images