A method for predicting rhizosphere beneficial bacterial community interaction with pathogenic bacteria based on a genomic metabolic model

Abstract

The application discloses a method for predicting the interaction between rhizosphere beneficial bacteria and pathogenic bacteria based on a genomic metabolic model, comprising the following steps: A. performing whole genome sequencing on the separated bacterial strain to obtain whole genome data of the bacterial strain; B. using gapseq software to construct an initial single-bacterial genomic metabolic model by using the whole genome data of the bacterial strain, and predicting available resources; C. detecting the actual resource utilization characteristics of the bacterial strain, optimizing the single-bacterial genomic metabolic model by using the gapseq software according to the metabolic pathways missing in the bacterial strain, improving the metabolic prediction ability of the model, and obtaining a corrected single-bacterial genomic metabolic model; D. constructing a genomic metabolic model of the beneficial bacteria group based on the corrected single-bacterial genomic metabolic model according to the principle of diversity full coverage; and E. predicting the interaction characteristics between the beneficial bacteria group and the pathogenic bacteria by using the genomic metabolic model of the beneficial bacteria group.

Description

Technical Field

[0001] This invention relates to the fields of soil biology and bioinformatics, specifically to constructing metabolic models of beneficial bacteria and pathogens using genomic data, and predicting the interaction between rhizosphere beneficial bacteria and pathogens based on flux balance analysis. Background Technology

[0002] Soil-borne diseases caused by soil-borne pathogens severely restrict my country's food security and soil productivity. To address this issue, the application of beneficial bacteria is widely considered a highly efficient and environmentally friendly method that can precisely suppress pathogens. However, the interaction between beneficial bacteria and pathogens is influenced by the complex microorganisms and root exudates in the rhizosphere, thus affecting their ability to suppress soil pathogen invasion. Current research on microbial metabolic interactions and the suppression of soil-borne pathogen invasion relies heavily on experimental methods, which are labor-intensive and time-consuming, and have limited resources and the number of detectable microorganisms. Therefore, how to rapidly and effectively predict the effect of microbial community on pathogen suppression in complex rhizosphere environments is a major technical bottleneck currently faced. Summary of the Invention

[0003] This invention addresses the problems existing in the background technology by constructing a new method based on a genome metabolic model to predict the interaction between beneficial rhizosphere bacteria and pathogens. This method can quickly, efficiently, and accurately predict the effect of beneficial bacteria on inhibiting pathogens based on metabolic interactions, providing an important method and technology for constructing disease-suppressing functional bacteria communities applicable to different complex rhizosphere environments and reducing soil-borne diseases.

[0004] The technical solution proposed in this invention includes the following specific steps:

[0005] A. Perform whole-genome sequencing on the isolated strains to obtain whole-genome data after the sequencing process.

[0006] B. Using the gapseq software, an initial single-strain genome metabolic model was constructed using the strain's whole genome data, and the available resources were predicted.

[0007] C. The actual resource utilization characteristics of the strains were detected by microplate assay. The missing metabolic pathways of the strains were optimized using gapseq to improve the metabolic prediction ability of the model and obtain a corrected single-strain genome metabolic model.

[0008] D. Based on the calibrated single-strain genome metabolic model, construct a genome metabolic model of beneficial bacteria according to the principle of full diversity coverage.

[0009] E. Using flux balance analysis with the COBRApy package in Python, predict the interaction characteristics between beneficial bacteria and pathogens.

[0010] The present invention also provides a method for selecting pathogenic bacteria that inhibit disease, based on screening the bacteria with the strongest antibacterial effect as candidate representatives for inhibiting disease.

[0011] 1) Construction of genomic metabolic models of beneficial and pathogenic bacteria

[0012] Whole-genome sequencing was performed on the isolated beneficial and pathogenic bacteria. Initial single-bacterial metabolic models were then constructed using the gapeseq software. This software has a large and comprehensive database; after inputting the genome sequence, it can automatically align and predict the corresponding metabolic pathways, including pathway structure, key enzymes, and reaction stoichiometry. Therefore, the metabolic model includes all known metabolic reactions in the organism and the genes encoding each enzyme. Subsequently, flux balance analysis using the Python COBRApy package was employed. Using the initial metabolic models of beneficial and pathogenic bacteria, the resource metabolic flux of metabolites was calculated as the objective function. A resource metabolic flux (flux) less than 0 indicates that the substance can be utilized, thus predicting the resources that the strain can utilize.

[0013] 2) Optimization of genomic metabolic models for beneficial and pathogenic bacteria

[0014] The resource utilization characteristics of beneficial and pathogenic bacteria were detected using a microplate system. The available resources predicted by the initial model were compared with experimental data from strain resource utilization. Gapseq software was used to fill gaps in the metabolic model, adding the relevant generation and transport reactions of resources predicted as unusable to the model, ensuring that the resource metabolic flux was less than 0, indicating that the resource could be utilized. The model was then optimized. Subsequently, based on the optimized metabolic model, a genomic metabolic model of the beneficial bacterial community was constructed using the Python COBRApy package.

[0015] 3) Prediction of the interaction characteristics between beneficial bacteria and pathogens

[0016] First, the interaction characteristics between beneficial bacteria and pathogens were predicted using the Python COBRApy package. For each bacterial community, the biomass flux of the pathogen growing alone was simulated first, followed by the biomass flux of the pathogen when co-cultured with beneficial bacteria. If the biomass flux of the pathogen during co-culture was lower than 90% of the biomass flux of the pathogen when cultured alone, the bacterial community was considered to effectively inhibit the pathogen. Conversely, if the biomass of the pathogen during co-culture was higher than 110% of the biomass of the pathogen when cultured alone, the bacterial community was considered to promote the growth of the pathogen. Furthermore, the ratio of the biomass flux of the pathogen in co-culture to log2 of the biomass flux when the pathogen was cultured alone was defined as the interaction fraction. If the absolute value of the interaction fraction exceeded 1, it was considered a strong inhibitory / promoting effect. To determine whether the disease-suppressing effect of a bacterial community is superior to that of a single bacterium, a similar comparative method can be used. If the biomass flux of the pathogen is less than 90% of that of the pathogen when the bacterial community is co-cultured with the pathogen, and the biomass flux of the pathogen is less than that when the single bacterium interacts with the pathogen, then the bacterial community is considered to have a stronger inhibitory effect than the single bacterium.

[0017] Beneficial effects of the present invention

[0018] 1. Traditional co-culture interaction detection mainly relies on plate experiments, which are time-consuming, cumbersome, and limited in the number of strains that can be co-cultured. This invention eliminates the need for plate experiments, directly constructing a corrected metabolic model based on the genomic DNA sequence and simple carbon source utilization characteristics. This significantly reduces the time required and enables large-scale, accurate prediction of co-culture interactions.

[0019] 2. Genomic metabolic models provide a comprehensive and systematic understanding of cellular metabolic networks, taking into account the complexity of intracellular physiological responses, including substrate and product generation, energy metabolism, and cell growth. This enables the models to more accurately predict the growth and metabolic characteristics of pathogens under different conditions compared to other strains.

[0020] 3. Based on a genomic metabolic model, this invention can realize large-scale, rapid, and accurate metabolic interaction characteristics between pathogens and beneficial bacteria, and can provide research techniques and strategies for constructing disease-suppressing bacterial communities under different complex environments and reducing pathogen pollution in soil-plant ecosystems. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the detection process of predicting the inhibition of pathogen invasion by beneficial rhizosphere bacteria based on a genomic metabolic model, as described in this invention.

[0022] Figure 2 This is a schematic diagram comparing the resource utilization prediction capabilities before and after model correction in the embodiment.

[0023] Figure 3 The results show the predicted interactions between beneficial bacteria and pathogens under different resource combinations in the examples. Detailed Implementation

[0024] The present invention will be further described below with reference to embodiments, but the scope of protection of the present invention is not limited thereto:

[0025] like Figure 1 As shown, the technical problem solved by this invention is that it can quickly and accurately predict the interaction between beneficial rhizosphere bacteria and pathogens using only genomic data and single-strain resource utilization experiments.

[0026] Based on the whole genome data of Ralstonia solanacearum and five beneficial bacteria, as well as the resource utilization characteristics of single bacteria, this invention constructs and optimizes a genome metabolic model. Based on flux balance analysis, it can quickly and accurately predict the interaction between beneficial bacteria and Ralstonia solanacearum, providing research techniques and strategies for constructing disease-suppressing bacteria communities under different complex environments and reducing pathogen pollution in soil-plant ecosystems.

[0027] The invention is described in detail below through examples. Experimental methods not specifically described herein are generally performed under conventional conditions or as recommended by the instrument manufacturer. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those familiar to those skilled in the art.

[0028] NA liquid culture medium: 3 g / L beef extract, 0.5 g / L yeast extract, 5 g / L peptone, 10 g / L glucose, deionized water to a final volume of 1 L, sterilized at 115°C for 30 min.

[0029] 48 types of tomato rhizosphere resources:

[0030] 1. Acetic acid (fluorochem, catalog number: 044721)

[0031] 2. L-Alanine (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: D800003)

[0032] 3. β-Alanine (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: D800003)

[0033] 4. L-Arginine (VETEC, catalog number: V900303)

[0034] 5. Ascorbic acid (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: A800295)

[0035] 6. L-Asparagine (Fluorochem, catalog number: M02999)

[0036] 7. γ-aminobutyric acid (TRC, catalog number: A602920)

[0037] 8. Citric acid (source leaf, catalog number: 12242)

[0038] 9. Citrulline (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: L805041)

[0039] 10. Ethanolamine (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 81007928)

[0040] 11. Formic acid (Shanghai Jingchun Biochemical Technology Co., Ltd., Aladdin, Catalog No.: F112038)

[0041] 12. Methyl sugar (Santa Cruz, catalog number: sc-221456)

[0042] 13. Galacturonic acid (Solebio Technology Co., Ltd., Solebio, Catalog No.: G8120-5)

[0043] 14. Glucose (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 63005518)

[0044] 15. L-Glutamine (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 62010838)

[0045] 16. Glutaric acid (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: g822471-5g)

[0046] 17. L-Glycine (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: G800883-500g)

[0047] 18. Glycolic acid (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: g810374-100ml)

[0048] 19. L-Histidine (Aladdin, Shanghai Jingchun Biochemical Technology Co., Ltd., Catalog No.: H111055-25g)

[0049] 20. Isoleucine (j&k, catalog number: 559486)

[0050] 21. Lactic acid (Aladdin, Shanghai Jingchun Biochemical Technology Co., Ltd., catalog number: L108841-500g)

[0051] 22. L-Lysine (Phytotech, catalog number: L594)

[0052] 23. L-Leucine (Caisson, catalog number: L005-100GM)

[0053] 24. Maleic acid (BBI, catalog number: A610333-0250)

[0054] 25. Malic acid (Phytotech, catalog number: M5536)

[0055] 26. Malonic acid (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 30111926)

[0056] 27. L-Methionine (J&K, Catalog No.: 945776)

[0057] 28. Inositol (Shanghai Jingchun Biochemical Technology Co., Ltd., Aladdin, catalog number: i108336-25g)

[0058] 29.2-Ketoglutaric acid (source leaf, catalog number: S64994-5g)

[0059] 30. L-Phenylalanine (J&K, Catalog No.: 162943)

[0060] 31. L-Proline (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 62019936)

[0061] 32. Pyruvic acid (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, catalog number: p815676-25g)

[0062] 33. L-Serine (Shanghai Maclean Biochemical Technology Co., Ltd., Maclean, Catalog No.: d817775-25g)

[0063] 34. Succinic acid (j&k, catalog number: 106413)

[0064] 35. Sucrose (Shanghai Jingchun Biochemical Technology Co., Ltd., Aladdin, catalog number: s112224-5g)

[0065] 36. Tartaric acid (Santa Cruz, catalog number: SC-252654)

[0066] 37. L-Threonine (Shanghai Jingchun Biochemical Technology Co., Ltd., Aladdin, catalog number: t108222-25g)

[0067] 38. L-Tryptophan (LGC, catalog number: MM1267.01)

[0068] 39. L-Valine (Source leaf, catalog number: S67607-5g)

[0069] 40. Maltose (phytotech, catalog number: m588)

[0070] 41. L-Arabinose (Absin Biotechnology Co., Ltd., catalog number: abs47018938-20mg)

[0071] 42. D-Galactose (TargetMOL, catalog number: T0591-25mg)

[0072] 43. D-Mannose (j&k, catalog number: 196857)

[0073] 44. D-xylose (j&k, catalog number: 203390)

[0074] 45. D-ribose (Shanghai Maclean Biotechnology Co., Ltd., Maclean, Catalog No.: d817221-25g)

[0075] 46. ​​D-Mannitol (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 63008816)

[0076] 47. Inosine (Sinopharm Chemical Reagent Co., Ltd., Shanghai Test, Catalog No.: 65009734)

[0077] 48. Oxalic acid (J&K, catalog number: 391404)

[0078] This implementation example is based on one pathogenic Ralstonia bacillus QL-Rs1115 and five beneficial bacteria isolated from soil in fields with a high incidence of bacterial wilt in tomatoes in the laboratory. These are: *Ralstonia pickettii* NJQL-A6 (deposited at the China General Microbiological Culture Collection Center, CGMCC No. 6628 on September 26, 2012), *Lysinibacillus sphaericus* HR92 (deposited at the China General Microbiological Culture Collection Center, CGMCC No. 6629 on September 26, 2012), and *Streptomycestorulosus* SHS (deposited at the China General Microbiological Culture Collection Center, CGMCC No. 11, 2022). No. 25281), *Streptomyces rishiriensis* SNP, and *C. rishiriensis* CHR6 (classified as *Streptomyces rishiriensis*, deposited on July 11, 2022 at the China General Microbiological Culture Collection Center, accession number CGMCC No. 25282) are used to illustrate a new method for predicting metabolic networks and disease-suppressing effects, including the following specific steps:

[0079] Construction of the initial single-strain genome metabolic model

[0080] Five beneficial bacteria and *Ralstonia solanacearum* isolated from the organism were subjected to whole-genome sequencing. Using the gapeseq software, the whole-genome data of the strains were input to generate preliminary single-strain genome metabolic models. This software has a large and complete self-built database; after inputting the genome sequence, it can perform alignment annotation and predict corresponding metabolic pathways, including pathway structure, key enzymes, and reaction stoichiometry. Therefore, the metabolic model includes all known metabolic reactions in the organism and the genes encoding each enzyme. Furthermore, gapeseq can simulate the growth environment of the strains; therefore, we set the growth environment of the strains to the same OS inorganic salt medium as in the resource utilization experiment, supplemented with 48 tomato rhizosphere resources. Subsequently, based on the Python COBRApy package, flux balance analysis was used to predict the resource utilization capacity of the strains. The resource metabolic flux of metabolites was calculated using resource metabolic flux as the objective function. If the resource metabolic flux (flux) is less than 0, it indicates that the substance can be utilized, thus predicting the resources that the strain can utilize.

[0081] Optimization of genomic metabolic models

[0082] Forty-eight resource utilization characteristics of five beneficial bacteria and *Ralstonia solanacearum* were detected using a microplate system. The resource utilization characteristics predicted by the initial model were compared with the microplate resource utilization experimental data. Gapseq software was used to fill gaps in the metabolic model, adding the relevant generation and transport reactions of resources that were predicted to be unusable into the model, ensuring that the resource metabolic flux was less than 0, indicating that the resource could be utilized. The model was then optimized to obtain a corrected single-bacterial genome metabolic model.

[0083] Microplate system for detecting the utilization characteristics of bacterial strain resources

[0084] The bacterial culture was removed from the -80℃ freezer and activated on NA solid medium. After single colonies grew on the plate, they were transferred to liquid NA medium and incubated overnight at 30℃ and 170 rpm with shaking until the logarithmic growth phase. The fermented bacterial solution was washed three times with 0.85% sterile physiological saline to remove residual culture medium from the surface. Specifically, the cells were collected by centrifugation at 6000 rpm for 6 min, and the cells were resuspended in 0.85% sterile physiological saline. This process was repeated three times. Finally, the bacterial concentration was adjusted to OD600 = 0.5 (approximately 107 CFU·ml⁻¹) with 0.85% sterile physiological saline. The tomato rhizosphere resource bank contains 48 carbon sources, including amino acids, organic acids, and sugars. 10 μL of bacterial solution was inoculated into 96-well plates containing different resources. The final concentration of the resources in the wells was 10 mM. Finally, the system was brought up to 200 μL with OS inorganic salt medium. The 96-well plate was placed in a shaker and cultured at 30°C and 170 rpm for 48 hours. The absorbance was measured using a microplate reader. If the absorbance OD600 > 0.05, the microorganisms were considered to be able to utilize the carbon source in the well.

[0085] Construction of a beneficial microbial community genome metabolic model

[0086] Based on the optimized single-strain genome metabolic model, and following the principle of full diversity coverage, the beneficial bacterial community diversity was constructed from 1 to 5 using the Python COBRApy package. These were arranged in pairs, groups of three, groups of four, and full coverage to form 31 beneficial bacterial communities, with each microorganism present in the same quantity in all communities.

[0087] Prediction and Validation of the Interaction Characteristics between Beneficial Bacteria and Ralstonia solanacearum. Twenty resources were randomly selected from 48 resources as resource combinations, repeated 20 times to simulate the rhizosphere environment. Microplate experiments were used to detect the interaction ability of beneficial bacteria with Ralstonia solanacearum under the 20 resource combinations. Subsequently, using the Python COBRApy package, the same growth environment as the 20 resource combinations was set in the model, and flux balance analysis was used to predict the interaction between beneficial bacteria and Ralstonia solanacearum. If the biomass flux of Ralstonia solanacearum during co-culture was lower than that of 90% of Ralstonia solanacearum cultured alone, the bacteria were considered to effectively inhibit Ralstonia solanacearum. Conversely, if the biomass of Ralstonia solanacearum during co-culture was higher than that of 110% of that of Ralstonia solanacearum cultured alone, the bacteria were considered to promote Ralstonia solanacearum growth. Furthermore, the ratio of the log2 of the biomass flux of Ralstonia solanacearum during co-culture to that of Ralstonia solanacearum cultured alone was defined as the interaction fraction. If the absolute value of the interaction fraction exceeded 1, it was considered a strong inhibitory / promoting effect. To determine whether the inhibitory effect of a bacterial community is superior to that of a single bacterium, a similar comparative method can be used. If the biomass flux of Ralstonia solanacearum (RSS) when co-cultured with the bacterial community is lower than 90% of the biomass flux of RSS when co-cultured with a single bacterium, then the bacterial community is considered to have a stronger inhibitory effect than the single bacterium. Finally, the predicted accuracy is compared with the interaction characteristics of the microplate experiment, the accuracy is calculated, and the results are plotted.

[0088] Microplate system for detecting the ability of beneficial bacteria to inhibit Ralstonia solanacearum invasion

[0089] All bacterial strains were cultured separately in NA liquid medium at 30℃ and 170 rpm for 12 h in shake flasks to construct 31 synthetic bacterial communities with an initial density of 10⁷ CFU·ml⁻¹. These communities were then inoculated into different culture environments to investigate the inhibitory effect of beneficial bacteria on *Ralstonia solanacearum* (Ralstonia bacterialis) under OS inorganic salt medium supplemented with 20 randomly combined rhizosphere resources. The final concentration of beneficial bacteria was consistent with the final concentration of individual bacteria, and individual bacteria within the communities were mixed in equal proportions. The following treatments were set up: 1) co-culture of the 31 synthetic bacterial communities with red fluorescently labeled *Ralstonia solanacearum* (10⁵ CFU·ml⁻¹), 2) 31 synthetic bacterial communities, and 3) *Ralstonia solanacearum* alone. The cultures were incubated at 30℃ and 170 rpm for 48 h with shaking. OD₆₀ and mCherry fluorescence signals (excitation: 587 nm, emission: 610 nm, expressed as RFP) were recorded. Each treatment was repeated in 6 replicates. The relative abundance of *Ralstonia solanacearum* was expressed as [log₁₀(OD₆₀ / RFP)].

[0090] The predicted resource utilization results before and after model correction in the example are as follows: Figure 2 The results showed that compared with the uncorrected model, the model correction significantly improved the predictive ability of 48 resource utilizations, reaching as high as 95% (P<0.0001). The prediction results of the interaction between beneficial bacteria and *Ralstonia solanacearum* under different resource combinations in the example are as follows: Figure 3The results showed that out of 620 interactions, the genomic metabolism model failed to predict interactions in only 10 instances, achieving a prediction accuracy of 98%, demonstrating the advantages of this method.

[0091] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for predicting the interaction between beneficial rhizosphere bacteria and pathogens based on a genome metabolic model, characterized in that... It includes the following steps: A. Perform whole-genome sequencing on the isolated strains to obtain whole-genome data; B. Using the GAPSESeq software, an initial single-strain genome metabolic model was constructed using the strain's whole genome data, and the available resources were predicted; the initial single-strain genome metabolic model was constructed through the following steps: Initial single-bacterial genome metabolic models were constructed using the gapseq software based on the genome sequences of isolated beneficial and pathogenic bacteria. The software automatically compares and predicts the corresponding metabolic pathways after inputting the genome sequences, including pathway structure, key enzymes, and reaction stoichiometry information. The initial single-bacterial genome metabolic model contains all known metabolic reactions in the organism and the genes encoding each enzyme. The predicted available resources include: Flux balance analysis in Python's COBRApy package was used to calculate the flux of beneficial bacteria and pathogens producing or utilizing metabolites by using initial single-strain genome metabolic models. If the flux is less than 0, it means that the substance can be utilized, and the resources that the strain can utilize can be predicted accordingly. C. The actual resource utilization characteristics of the strain were detected using a microplate assay. The missing metabolic pathways in the strain were optimized using Gapseq software to improve the metabolic prediction ability of the initial single-strain genome metabolic model, thereby obtaining a corrected single-strain genome metabolic model. The corrected single-strain genome metabolic model was obtained through the following steps: The available resources predicted by the initial model were compared with the microplate resource utilization test data. Gapseq software was used to fill the gaps in the initial single-strain genome metabolic model. The relevant generation and transport reactions of the unusable resources predicted in the model were added to the model so that the resource metabolic flux was less than 0, which indicated that the resource could be utilized, and a corrected single-strain genome metabolic model was obtained. D. Based on the calibrated single-strain genome metabolic model, and following the principle of full diversity coverage, construct a genome metabolic model of beneficial bacteria. The genome metabolic model of beneficial bacteria is obtained through the following steps: Based on the calibrated single-strain genome metabolic model, construct a genome metabolic model of beneficial bacteria using the Python COBRApy package. E. Using flux balance analysis with the COBRApy package in Python, predict the interaction characteristics between beneficial bacteria and pathogens, specifically including: 1) For each bacterial community, first simulate the biomass flux when the pathogen grows alone, and then simulate the biomass flux of the pathogen when the beneficial bacteria are co-cultured with the pathogen. If the biomass flux of the pathogen during co-culture is lower than 90% of the biomass flux when the pathogen is cultured alone, then the bacterial community is considered to effectively inhibit the pathogen. Conversely, if the biomass of the pathogen during co-culture is higher than 110% of the biomass when the pathogen is cultured alone, then the bacterial community is considered to promote the growth of the pathogen. 2) The ratio of the biomass flux of pathogens in co-culture to log2 of the biomass flux of pathogens in single-culture is defined as the interaction fraction; if the absolute value of the interaction fraction exceeds 1, it is considered a strong inhibitory or strong promoting effect. 3) If the biomass flux of the pathogen is less than 90% of the biomass flux of the pathogen when the microbial community is co-cultured with the pathogen, then the microbial community is considered to have a stronger inhibitory effect than the single bacteria.

2. A method for selecting pathogenic bacteria to suppress disease populations, characterized in that, Using the method described in claim 1 for predicting the interaction between rhizosphere beneficial bacteria and pathogens based on a genomic metabolic model, the bacterial groups with the strongest antibacterial effect are screened as candidate representatives for suppressing diseases.

Images