A method, device, equipment and medium for dynamically matching fecal flora transplant donors and recipients

Abstract

The embodiment of the application discloses a kind of fecal bacteria transplantation donor-recipient dynamic matching method, device, equipment and medium, in response to donor-recipient matching request, the fecal bacteria transplantation after monitoring data of current preset sampling time of to-be-matched recipient is acquired;Wherein, the fecal bacteria transplantation after monitoring data is determined based on the fecal bacteria group of any one target donor to to-be-matched recipient after completing fecal bacteria transplantation operation, and the target donor is determined based on the genetic feature information, immune feature information and clinical feature information of to-be-matched recipient and multiple candidate donors for adaptability matching;Fecal bacteria transplantation after monitoring data is input into transplantation strategy control model, and at least one of the donor switching strategy and dose adjustment strategy for adjusting to-be-matched recipient in next period is output, overall compatibility evaluation is realized by multi-modal data integration, and treatment scheme is continuously optimized by means of dynamic feedback mechanism, which significantly improves the precision and adaptability of fecal bacteria transplantation treatment.

Description

Technical Field

[0001] This invention relates to the field of intestinal flora transplantation technology, and in particular to a method, apparatus, equipment and medium for dynamic matching of donor and recipient in fecal microbiota transplantation. Background Technology

[0002] Fecal microbiota transplantation (FMT) is a treatment method that reconstructs the recipient's gut microbiota by transplanting gut microbiota from a healthy donor. It has shown significant efficacy in diseases such as recurrent Clostridium difficile infection and inflammatory bowel disease. Its core mechanism involves regulating the host's immune and metabolic functions and restoring gut microbiota balance through the implantation of the microbial community. With the development of microbiome, immunology, and artificial intelligence technologies, FMT is gradually evolving from empirical treatment towards precision and personalization. However, multi-dimensional data integration and dynamic optimization remain key challenges for technological breakthroughs.

[0003] Current FMT matching methods mainly rely on static matching strategies based on microbiome similarity analysis. That is, before fecal microbiota transplantation, the donor is determined once through a similarity assessment algorithm, thereby providing the recipient with a healthy microbiota through the donor's gut microbiota.

[0004] However, the aforementioned FMT matching methods have the following problems: Firstly, they rely excessively on microbiome data, resulting in a single data dimension and difficulty in comprehensively assessing the biological compatibility of donors and recipients. Secondly, due to the static matching strategy, they cannot respond to the dynamic changes in the recipient's microbiome and the evolution of their clinical condition after transplantation, leading to treatment lag. Therefore, existing FMT matching methods suffer from the dual limitations of static matching strategies failing to adapt to dynamic changes after transplantation and the single data dimension resulting in incomplete assessment of biological compatibility. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for dynamic matching of donor and recipient in fecal microbiota transplantation (FMT), which achieves comprehensive compatibility assessment through multimodal data integration and continuously optimizes treatment plans through a dynamic feedback mechanism, significantly improving the accuracy and adaptability of FMT treatment.

[0006] In a first aspect, embodiments of the present invention provide a method for dynamic matching of donor and recipient in fecal microbiota transplantation, the method comprising:

[0007] In response to a donor-recipient matching request, the system acquires fecal microbiota transplantation (FMT) post-transplantation monitoring data of the recipient at the current preset sampling time. The FMT post-transplantation monitoring data includes recipient symptom scores and microbiota colonization rates. This data is determined based on the fecal microbiota of any target donor after the FMT surgery is performed on the recipient. The target donor is one of multiple donors identified through compatibility matching based on the recipient's genetic characteristics, immune characteristics, and clinical characteristics with multiple candidate donors.

[0008] The fecal microbiota transplantation monitoring data is input into the transplantation strategy control model, which outputs a transplantation adjustment strategy for adjusting the recipient to be matched in the next time period; wherein, the transplantation adjustment strategy includes at least one of the following: donor switching strategy and dose adjustment strategy.

[0009] Secondly, embodiments of the present invention also provide a dynamic matching device for fecal microbiota transplantation donors and recipients, the device comprising:

[0010] The monitoring data acquisition module is used to acquire the fecal microbiota transplantation (FMT) post-transplantation monitoring data of the recipient to be matched at the current preset sampling time in response to the donor-recipient matching request. The FMT post-transplantation monitoring data includes the recipient's symptom score and microbiota colonization rate. The FMT post-transplantation monitoring data is determined after the fecal microbiota of any target donor is used to perform FMT surgery on the recipient to be matched. The target donor is a group of donors determined by adaptability matching based on the genetic characteristics, immune characteristics, and clinical characteristics of the recipient to be matched and multiple candidate donors.

[0011] The adjustment strategy determination module is used to input the fecal microbiota transplantation post-transplantation monitoring data into the transplantation strategy control model and output the transplantation adjustment strategy for adjusting the recipient to be matched in the next time period; wherein, the transplantation adjustment strategy includes at least one of the following: donor switching strategy and dose adjustment strategy.

[0012] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0013] One or more processors;

[0014] A storage device for storing one or more programs, which, when executed by one or more processors, enable the one or more processors to implement a fecal microbiota transplantation donor-recipient dynamic matching method as described in any embodiment of the present invention.

[0015] Fourthly, embodiments of the present invention also provide a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a fecal microbiota transplantation donor-recipient dynamic matching method as described in any of the embodiments of the present invention.

[0016] The technical solution of this invention, in response to a donor-recipient matching request, acquires fecal microbiota transplantation (FMT) monitoring data of the recipient at the current preset sampling time. The FMT monitoring data includes the recipient's symptom score and microbiota colonization rate. This data is determined based on the fecal microbiota of any target donor after the recipient has undergone FMT surgery. The target donor is determined by matching the recipient with multiple donors based on their genetic, immune, and clinical characteristics. Further, the FMT monitoring data is input into a transplantation strategy control model, which outputs a transplantation adjustment strategy for adjusting the recipient's transplantation in the next time period. This transplantation adjustment strategy includes at least one of a donor switching strategy and a dose adjustment strategy. The technical solution provided in this embodiment, based on a dynamic feedback mechanism of real-time post-transplant monitoring data, continuously optimizes donor selection and dosage adjustment through a transplantation strategy control model, solving the pain point that static matching cannot adapt to dynamic changes after transplantation. In addition, it integrates multimodal data such as genetic characteristics, immune characteristics, and clinical characteristics, breaking through the limitations of single microbiome data. Through multimodal data integration, it achieves comprehensive compatibility assessment and continuously optimizes treatment plans with the help of dynamic feedback mechanisms, significantly improving the accuracy and adaptability of FMT treatment. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of exemplary embodiments of the present invention, the accompanying drawings used in describing the embodiments are briefly introduced below. Obviously, the accompanying drawings described are only a portion of the drawings of the embodiments to be described in this invention, and not all of the drawings. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort.

[0018] Figure 1 This is a flowchart illustrating a dynamic matching method for fecal microbiota transplantation donors and recipients provided in an embodiment of the present invention.

[0019] Figure 2 This is a schematic flowchart of another dynamic matching method for fecal microbiota transplantation donors and recipients provided in an embodiment of the present invention;

[0020] Figure 3 This is a schematic flowchart of another method for dynamic matching of donor and recipient in fecal microbiota transplantation according to an embodiment of the present invention.

[0021] Figure 4 This is a schematic diagram of the structure of a fecal microbiota transplantation donor-recipient dynamic matching device provided in an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0023] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0024] Before introducing the embodiments of the present invention, the application scenarios of the present invention can be illustrated by example. The embodiments of the present invention can be applied to any situation that requires dynamic monitoring after fecal microbiota transplantation and real-time adjustment of the donor or transplantation strategy based on the recipient's symptoms and microbiota colonization. Fecal microbiota transplantation is a medical method that extracts and processes functional microbiota from the feces of a healthy donor using specific techniques, and then transplants them into the patient's intestines to rebuild the balance of the intestinal microecology. Its core purpose is to inhibit pathogenic bacteria and repair intestinal microbiota dysbiosis by introducing diverse beneficial microorganisms, thereby treating diseases related to intestinal microbial imbalance. Key steps of this technology include donor screening, fecal processing, and delivery of the microbiota to the patient's intestines via oral capsules, nasoenteric tubes, colonoscopy, etc. Its efficacy depends on the functional integrity of the donor microbiota and its compatibility with the recipient's intestinal environment. This embodiment aims to improve the accuracy and efficacy stability of fecal microbiota transplantation by dynamically monitoring the recipient's symptom scores and microbiota colonization rate after fecal microbiota transplantation, combining genetic, immune, and clinical characteristics information, and using a transplantation strategy control model to optimize donor selection and transplantation protocols in real time. This will reduce transplantation failure or adverse reactions caused by donor-recipient mismatch and achieve individualized, adaptive fecal microbiota transplantation treatment management.

[0025] Figure 1 This is a flowchart illustrating a dynamic matching method for fecal microbiota transplantation donors and recipients provided in an embodiment of the present invention. This embodiment is applicable to any situation where dynamic monitoring is required after fecal microbiota transplantation and the donor or transplantation strategy is adjusted in real time based on the recipient's symptoms and microbial colonization. This method can be executed by a dynamic matching device for fecal microbiota transplantation donors and recipients. This device can be implemented in the form of software and / or hardware. The hardware can be an electronic device, such as a mobile terminal, a PC, or a server.

[0026] like Figure 1 As shown, the dynamic matching method for fecal microbiota transplantation donors and recipients includes:

[0027] S110. In response to a donor-recipient matching request, acquire post-fecal microbiota transplantation monitoring data of the recipient to be matched at the current preset sampling time.

[0028] In this context, a donor-recipient matching request refers to an instruction or demand initiated by the healthcare system or relevant parties during fecal microbiota transplantation (FMT) to dynamically match a specific recipient (patient) with a potential donor (fecal microbiota provider). A recipient awaiting matching refers to an individual patient currently requiring donor matching assessment during FMT treatment. This recipient has already undergone FMT based on initial matching but still needs further optimization of donor selection or transplantation strategy based on their postoperative dynamic monitoring data. The current preset sampling time refers to a pre-defined key monitoring time point used to assess the recipient's postoperative status. This time point is determined based on clinical treatment guidelines or individualized protocols, aiming to systematically collect the recipient's post-transplantation monitoring data at that moment, providing timely input data for the dynamic matching model to support the phased assessment of transplantation outcomes and subsequent strategy adjustment decisions.

[0029] Post-fecal microbiota transplantation (FMT) monitoring data refers to a set of quantitative indicators collected at predetermined time points after FMT to assess the recipient's treatment efficacy. FMT monitoring data includes recipient symptom scores and microbiota colonization rate. The recipient symptom score is an indicator parameter that quantifies the degree of improvement in clinical symptoms of FMT recipients using a standardized assessment system. This scoring system is based on predefined clinical evaluation criteria and uses a specific algorithm to convert treatment efficacy into quantifiable values ​​by collecting data such as recipient physiological indicators and subjective symptom reports. The microbiota colonization rate is an indicator of the proportion of donor microbiota successfully colonizing the recipient's gut, determined using microbiome analysis technology. This parameter quantifies the implantation efficiency and stability of the transplanted microbiota in the recipient's gut microbiota ecosystem. As a core biological indicator for assessing the efficacy of FMT, it reflects the survival status and niche occupancy of the donor functional microbiota in the recipient's gut. Changes in its value are directly related to the microbiological efficacy of transplantation treatment, providing objective evidence of microbial colonization for clinical decision-making.

[0030] More specifically, post-fecal microbiota transplantation (FMT) monitoring data is determined based on the fecal microbiota of any target donor after the fecal microbiota transplantation surgery is completed on the recipient. Target donors are multiple donors identified through compatibility matching of the recipient and multiple candidate donors, considering genetic, immune, and clinical characteristics. A target donor is the fecal microbiota provider most compatible with the recipient, selected from a pool of candidate donors through multi-dimensional characteristic matching. This selection is based on a comprehensive evaluation of multiple parameters, including genetic, immune, and clinical characteristics of both donor and recipient. Its core characteristic is optimal biological compatibility with the recipient in terms of microbiome composition, function, and host interaction. Candidate donors refer to the set of potential fecal microbiota providers who have entered the final matching evaluation stage after initial screening. Genetic characteristic information refers to the set of genetic parameters used to assess biocompatibility, mainly including genetic matching data between the donor and recipient at specific functional gene loci, HLA genotype, and microbiome characteristic genes. Immune characteristic information refers to the set of key immunological parameters used to assess donor-recipient immunocompatibility, mainly including characteristic data of donor and recipient in terms of immune cell composition, cytokine profile, immune response pattern, and mucosal immune status. Clinical characteristic information refers to the set of medical parameters used to assess donor-recipient clinical suitability during fecal microbiota transplantation matching, mainly including clinical medical data directly related to predicting treatment efficacy, such as disease type, disease stage, medication history, and complications. This information is obtained through medical record collection and clinical evaluation, and is used to comprehensively determine the applicability and expected efficacy of transplantation treatment. It is an important clinical dimension for achieving precise matching by combining genetic and immune characteristics, and its core value lies in ensuring that donor selection matches the actual clinical needs of the recipient.

[0031] Before implementing this protocol, the relevant data of the recipient and the relevant data of each donor in the donor candidate pool can be compared and matched to identify multiple target donors suitable for the recipient. Specifically, the screening process for target donors involves comparing the degree of fit between the recipient to be matched and multiple candidate donors at the genetic level (such as microbiome similarity, host genetic background), immune characteristics (such as immune response matching degree, inflammatory factor compatibility), and clinical indicators (such as disease type, medication history, health status). The optimal donor combination is dynamically determined after evaluation using a matching algorithm, and the matching results may include multiple suitable donors to adapt to different treatment stages or individualized adjustment needs.

[0032] Specifically, the fecal microbiota transplantation (FMT) donor-recipient dynamic matching method provided in this embodiment can be integrated into a donor-recipient dynamic matching plugin. This plugin can be deployed on any server that needs to perform FMT donor-recipient matching. After completing FMT treatment, the recipient can determine the post-FMT monitoring data at a preset time period and store this data in a database. When the user clicks to respond to the donor-recipient matching request in the display interface of the donor-recipient dynamic matching plugin, the post-FMT monitoring data corresponding to the recipient can be retrieved from the database. This data reflects the current clinical status and microbial colonization effect of the recipient after receiving FMT, including key indicators such as the degree of symptom improvement and the survival rate of the transplanted microbiota, providing real-time basis for subsequent dynamic adjustment of the treatment plan.

[0033] S120. Input the fecal microbiota transplantation monitoring data into the transplantation strategy control model, and output the transplantation adjustment strategy for adjusting the recipient to be matched in the next period.

[0034] The transplantation strategy control model refers to an algorithm-based intelligent decision-making system. This model analyzes post-operative monitoring data of the recipient (such as symptom scores and microbial colonization rates), combines it with preset treatment rules or optimization strategies derived from machine learning, and automatically generates personalized adjustment plans for the next treatment stage. The output of the transplantation strategy control model can involve changes to key treatment parameters such as changing the donor source or adjusting the microbial transplantation dosage, thereby achieving dynamic optimization management of the fecal microbiota transplantation treatment process.

[0035] Transplantation adjustment strategies refer to the targeted treatment plan optimization instructions generated by the transplantation strategy control model. These strategies include at least one of donor switching strategies and dosage adjustment strategies. Donor switching strategies, during fecal microbiota transplantation (FMT), involve selecting a more suitable donor from a pool of alternative donors based on a pre-defined algorithm when the assessment reveals that the current donor's microbiota transplantation effect is not as expected. This strategy improves the success rate of microbiota colonization and clinical efficacy by dynamically optimizing the donor-recipient pairing and is one of the key intervention methods in transplantation adjustment strategies. Dosage adjustment strategies, during FMT, involve dynamically adjusting the amount of microbiota transplanted after the recipient's FMT, based on the assessment results of post-transplantation monitoring data, to optimize the treatment effect by increasing or decreasing the amount of microbiota transplanted subsequently. The aim is to improve microbiota colonization and the degree of clinical symptom relief by precisely adjusting the amount of microbial community input, and it is an important component of personalized transplantation protocol adjustments.

[0036] In this embodiment, the next time period refers to the next preset treatment cycle or observation window immediately following the current monitoring time point during fecal microbiota transplantation (FMT) treatment. This time interval is preset by the clinical treatment plan or model parameters and is used to implement the transplantation adjustment strategy based on the current assessment results. It is a key time node for implementing treatment intervention during dynamic matching.

[0037] Specifically, based on the post-fecal microbiota transplantation (FMT) monitoring data, the FMT monitoring data of the recipient to be matched can be input into the transplantation strategy control model. The model then uses algorithms to generate targeted transplantation adjustment strategies to guide the next stage of treatment. The output includes two core regulation methods: one is to change the source of fecal microbiota based on the donor microbiota fit, i.e., the donor switching strategy; the other is to adjust the amount of microbiota delivered in subsequent transplantations, i.e., the dosage adjustment strategy. The system will automatically select to implement one of the strategies or use both strategies simultaneously based on the model evaluation results to achieve personalized dynamic optimization of the treatment plan.

[0038] For example, after integrating this donor-recipient dynamic matching plugin into a hospital's fecal microbiota transplantation (FMT) system, the doctor can automatically record patient A's symptom scores and microbiota colonization rate to the hospital database at preset intervals (such as postoperative days 3, 5, 7, and 14). On postoperative day 4, when the attending physician clicks the "Smart Assessment" button in the electronic medical record system, the plugin retrieves patient A's data from week 3, i.e., postoperative FMT monitoring data. Subsequently, this fecal microbiota transplantation (FMT) monitoring data can be automatically input into the transplantation strategy control model. The model analysis revealed that the fecal microbiota functional genes of donor M matched the patient's data with only 68% (below the 75% threshold). The system immediately generated a strategy to "switch to alternative donor N with a matching rate of 82%". At the same time, it detected that the patient's intestinal bacterial load was insufficient, and a suggestion to "increase the transplantation dose by 20%" was added. After review by the medical team, the system automatically scheduled a 1.2 times dose transplantation treatment for donor N and set up a new round of postoperative monitoring plan, realizing closed-loop management of the entire process from data collection and intelligent analysis to treatment optimization.

[0039] The technical solution of this invention, in response to a donor-recipient matching request, acquires fecal microbiota transplantation (FMT) monitoring data of the recipient at the current preset sampling time. The FMT monitoring data includes the recipient's symptom score and microbiota colonization rate. This data is determined based on the fecal microbiota of any target donor after the recipient has undergone FMT surgery. The target donor is determined by matching the recipient with multiple donors based on their genetic, immune, and clinical characteristics. Further, the FMT monitoring data is input into a transplantation strategy control model, which outputs a transplantation adjustment strategy for adjusting the recipient's transplantation in the next time period. This transplantation adjustment strategy includes at least one of a donor switching strategy and a dose adjustment strategy. The technical solution provided in this embodiment, based on a dynamic feedback mechanism of real-time post-transplant monitoring data, continuously optimizes donor selection and dosage adjustment through a transplantation strategy control model, solving the pain point that static matching cannot adapt to dynamic changes after transplantation. In addition, it integrates multimodal data such as genetic characteristics, immune characteristics, and clinical characteristics, breaking through the limitations of single microbiome data. Through multimodal data integration, it achieves comprehensive compatibility assessment and continuously optimizes treatment plans with the help of dynamic feedback mechanisms, significantly improving the accuracy and adaptability of FMT treatment.

[0040] Figure 2 This diagram illustrates a dynamic matching method for fecal microbiota transplantation donors and recipients provided in an embodiment of the present invention. Based on the aforementioned embodiments, the method for determining the target donor is described in detail. Specific implementation methods can be found in the technical solution of this embodiment. Technical terms that are the same as or corresponding to those in the above embodiments will not be repeated here.

[0041] like Figure 2 As shown, the method specifically includes the following steps:

[0042] S210. Obtain the receptor sample data corresponding to the target receptor and the donor sample data corresponding to multiple candidate donors.

[0043] Recipient sample data refers to a multi-dimensional biomedical dataset collected from potential recipients to assess fecal microbiota transplantation suitability. This data includes recipient fecal microbiota genetic data, recipient immune status data, and recipient clinical metadata. Recipient fecal microbiota genetic data refers to the microbial genome information obtained from fecal samples of potential recipients using technologies such as high-throughput sequencing. This primarily includes genetic characteristics such as the species composition, functional gene distribution, and microbial diversity of the recipient's gut microbiota. These data objectively reflect the structural characteristics and metabolic potential of the recipient's gut microbiota ecosystem. Recipient immune status data refers to a comprehensive set of indicators obtained through laboratory testing and clinical assessment, reflecting the functional status and response characteristics of the potential recipient's immune system. This mainly includes immunological parameters such as immune cell composition and activity, cytokine levels, inflammatory marker concentrations, and autoantibody lineages. These data are used to assess the potential acceptance or rejection risk of the transplanted microbiota by the recipient's immune system. Receptor clinical metadata refers to the collection of clinical background and medical history information related to the candidate receptor, including but not limited to the patient's past medical history, medication records, allergy history, physiological and biochemical indicators, imaging examination results, and other structured clinical data. This metadata provides the necessary clinical background for assessing the patient's overall health status, disease progression, and treatment suitability.

[0044] In this context, "potential donors" refers to a group of potential fecal microbiota transplantation (FMT) donors who have passed preliminary screening and meet basic donation criteria but have not yet completed a matching assessment with a specific recipient. These donors have completed basic screening, including microbial gene testing, immune status assessment, and clinical health examinations, and their sample data has been included in a donor database as candidate matching targets. Donor sample data refers to a multi-dimensional biomedical data set collected from individual potential donors to assess their suitability as FMT donors. Donor sample data includes donor fecal microbiota gene data, donor immune status data, and recipient clinical metadata. Donor fecal microbiota gene data refers to the microbial genome information obtained from a fecal sample of a potential donor using technologies such as high-throughput sequencing. Donor immune status data refers to a comprehensive set of indicators reflecting the functional status and response characteristics of a potential donor's immune system, obtained through laboratory testing and clinical assessment. Donor clinical metadata refers to a set of clinical background and medical history information related to a potential donor. In this embodiment, each potential donor corresponds to a set of donor sample data.

[0045] In this embodiment, two types of key datasets can be retrieved from a medical database: one is the complete biological and clinical dataset of the recipient to be matched (including its gut microbiota genetic characteristics, immune system status indicators, and comprehensive clinical background information), i.e., recipient sample data; the other is the dataset of candidate donors of the same dimensions after preliminary screening, i.e., donor sample data corresponding to each candidate donor. This data collection step establishes an analytical foundation for subsequent precise matching of donors and recipients, which includes multi-dimensional information from microbiology, immunology, and clinical medicine, and is a prerequisite for achieving intelligent matching.

[0046] S220. Perform feature extraction processing on the receptor sample data to obtain receptor gene feature information, receptor immune feature information, and receptor clinical feature information.

[0047] In this embodiment, receptor genetic feature information refers to a set of key genetic indicators characterizing the composition and functional characteristics of the receptor's gut microbiota, obtained through feature extraction processing of the receptor's fecal microbiota gene data. Receptor immune feature information refers to a set of key immunological indicators characterizing the functional characteristics and response patterns of the receptor's immune system, obtained through feature extraction processing of the receptor's immune status data. Receptor clinical feature information refers to a set of key clinical indicators characterizing the receptor's overall health status and disease characteristics, obtained through feature extraction processing of the receptor's clinical metadata.

[0048] Specifically, the recipient fecal microbial genetic data includes: metagenomic sequencing data, species composition types of recipient fecal samples, and relative abundance data of each species. The recipient genetic feature information includes: relative abundance features of microbial species, microbial co-occurrence network diagram, and functional gene prediction feature information.

[0049] Metagenomic sequencing data refers to the genetic information obtained by high-throughput sequencing of bacterial 16S ribosomal RNA genes in recipient fecal samples. This gene serves as a marker gene for bacterial classification, and its sequence variations accurately reflect the composition and abundance distribution of different bacterial species within the microbial community. Species composition type refers to the taxonomic information of microbial species present in recipient or donor fecal samples and their community structure characteristics obtained through bioinformatics analysis of metagenomic sequencing data, including the distribution of identifiable bacterial phyla, classes, orders, families, genera, and species at different taxonomic levels. Relative abundance data for each species refers to the quantitative indicator of the proportion of each microbial species in the overall microbial community obtained after processing the metagenomic sequencing data through bioinformatics analysis. This data accurately characterizes the distribution of microorganisms at different taxonomic levels (such as genus, species, etc.) within the community in percentage or proportion form.

[0050] Based on this, the specific steps for extracting features from the recipient fecal microbial gene data to obtain recipient gene feature information may include:

[0051] S01. The species composition types and relative abundance data of each species in the recipient fecal samples are processed by feature vectorization to obtain the relative abundance characteristics of microbial species.

[0052] Among them, the relative abundance characteristics of microbial species refer to the quantitative characterization obtained by structuring or dimensionality reduction of the relative abundance data of different microbial species in fecal samples through mathematical or statistical methods. It reflects the composition and distribution pattern of the microbial community and is usually presented in the form of numerical vectors or matrices.

[0053] In this embodiment, all species composition types and their relative abundance information detected in the recipient fecal sample can be converted into a structured numerical representation. The original species abundance data can then be integrated into a standardized feature vector that can be used for calculation or modeling through mathematical transformation or statistical methods, thus obtaining the relative abundance characteristics of microbial species. In this way, by converting complex microbial community composition information into low-dimensional numerical features that reflect its overall distribution pattern, subsequent machine learning analysis or pattern recognition is facilitated.

[0054] S02. Using co-occurrence network analysis tools, network graphs were constructed from the species composition types and relative abundance data of each species in the recipient fecal samples to obtain a microbial co-occurrence network graph.

[0055] Co-occurrence network analysis tools refer to a type of computational software or algorithm platform specifically designed to identify and visualize co-occurrence relationships among microbial species. They use statistical methods (such as correlation analysis and mutual information calculation) to detect co-occurrence patterns of different species in multiple samples, and construct species interaction networks based on preset significance thresholds, ultimately generating a visualized network graph containing topological parameters.

[0056] Microbial co-occurrence network diagrams are visual network structures used to illustrate the coexistence patterns among different species in a microbial community. Nodes in a microbial co-occurrence network diagram represent different species, and the edges connecting nodes indicate significant correlations between species. This can be understood as the edges connecting nodes indicating a statistically significant positive correlation (symbiosis) or negative correlation (competition) between species, with the edge weights typically reflecting the strength of the correlation. Such network diagrams can intuitively present the ecological interaction characteristics and community structural stability of a microbial community.

[0057] In this embodiment, a specialized co-occurrence network analysis tool can be used to screen species pairs with significant co-occurrence relationships by calculating statistical correlations between species (such as Spearman rank correlation, SparCC, etc.) based on the species composition types of microbial species in multiple recipient fecal samples and the relative abundance data of each species. These relationships are then transformed into a network topology structure composed of nodes (representing species) and edges (representing significant correlations), thereby forming a microbial co-occurrence network diagram that can intuitively display the interaction relationships between species within the microbial community.

[0058] S03. Gene function prediction analysis is performed on metagenomic sequencing data using the PICRUSt2 functional prediction tool to obtain functional gene prediction results. Feature vectorization is then performed on the functional gene prediction results to obtain functional gene prediction feature information.

[0059] Among them, PICRUSt2 is a bioinformatics software for predicting the function of microbial communities based on metagenomic sequencing data. Functional gene prediction results refer to the functional characteristics of microbial communities obtained through indirect inference from metagenomic sequencing data; essentially, they are based on the correspondence between species classification and functional genes in known microbial genome databases. Functional gene prediction feature information refers to the quantitative representation data formed after mathematical transformation and structuring of the functional gene prediction results. It transforms the original prediction results into computable numerical features through dimensionality reduction or feature extraction methods.

[0060] In this embodiment, the PICRUSt2 tool can be used to infer the potential functional gene composition of a microbial community based on 16S rRNA gene sequence data, obtaining functional gene prediction results. Furthermore, these functional gene prediction results are transformed into structured, quantifiable numerical features using mathematical methods, thus obtaining functional gene prediction feature information. In this way, by converting complex gene function prediction data into standardized feature representations suitable for machine learning algorithms, quantitative analysis and comparison of microbial community functional characteristics can be achieved.

[0061] Optionally, a specific implementation method for extracting features from receptor immune status data to obtain receptor immune feature information includes: performing feature vectorization processing on receptor immune status data to obtain receptor immune feature information.

[0062] In this embodiment, raw receptor immune status data (such as cytokine levels, immune cell counts, or antibody titers) can be transformed into structured numerical vectors through mathematical transformation or standardization methods to obtain receptor immune feature information. This processing can integrate multi-dimensional immune indicators into machine-readable quantitative features, thereby converting complex immune status information into a digital representation that can be used for computational analysis and modeling, facilitating subsequent immune feature analysis and pattern recognition.

[0063] Optionally, a specific implementation method for extracting features from receptor clinical metadata to obtain receptor clinical feature information may include: performing feature vectorization processing on receptor clinical metadata based on a bidirectional encoder representation transformation model to obtain clinical data feature vectors.

[0064] Among them, the bidirectional encoder representation transformation model is a natural language processing architecture based on deep learning. Its core feature is that it can simultaneously capture the contextual and semantic information of each word in the text data through the bidirectional Transformer structure. The model learns to transform the input text sequence into a dense vector representation containing rich semantic features through two stages: pre-training and fine-tuning. This enables deep semantic encoding and feature extraction of text data, and is particularly suitable for handling text feature representation tasks that require understanding global contextual relationships.

[0065] The recipient clinical metadata includes at least one of the following: recipient attribute information, clinical symptom description information, recipient disease type, past medical history information, family medical history information, recent medication use, lifestyle information, human leukocyte antigen (HLA) typing, and blood type information. Receptor attribute information refers to the patient's basic demographic characteristics; clinical symptom description information is a textual record of the patient's subjective discomfort and objective signs; recipient disease type indicates the patient's diagnosed disease classification; past medical history information includes the patient's previously diagnosed diseases and treatments; family medical history information records the history of hereditary diseases in immediate family members; recent medication use reflects the patient's current or recent medication information; lifestyle information covers daily habits such as diet and exercise; HLA typing is the result of testing for immune-related genotypes; and blood type information refers to the results of ABO and Rh red blood cell antigen typing.

[0066] In this embodiment, a bidirectional encoder representation transformation model can be used to transform clinical metadata (such as symptom descriptions and medical history records) containing both textual and numerical data into low-dimensional dense numerical vector representations through a deep learning model, thereby obtaining clinical data feature vectors. The semantic understanding capabilities acquired through pre-training of this model can capture contextual relationships and potential medical features in clinical texts, thus generating feature vectors rich in clinical semantic information, providing structured numerical representations for subsequent medical data analysis.

[0067] S230. Perform feature extraction processing on the donor sample data of each group to obtain donor gene feature information, donor immune feature information and donor clinical feature information corresponding to each candidate donor.

[0068] In this embodiment, for each group of donor sample data, the same feature extraction method as that used for recipient sample data can be used for feature extraction processing, thereby obtaining donor gene feature information, donor immune feature information, and donor clinical feature information corresponding to each group of donor sample data.

[0069] Based on the recipient's genetic characteristics, immune characteristics, and clinical characteristics, as well as the donor's genetic characteristics, immune characteristics, and clinical characteristics corresponding to each donor, this data can be stored in a pre-defined storage space.

[0070] S240. Obtain the receptor sample feature data corresponding to the recipient to be matched and the donor sample feature data corresponding to multiple donors.

[0071] The recipient sample characteristic data includes: disease type and severity identifiers, recipient genetic characteristics, recipient immune characteristics, and recipient clinical characteristics. The donor sample characteristic data includes: donor genetic characteristics, donor immune characteristics, and donor clinical characteristics. The disease type and severity identifiers refer to the standardized classification codes of the patient's disease determined through medical diagnosis, as well as quantitative assessment indicators of the current stage of disease development and the severity of clinical symptoms.

[0072] In this embodiment, when it is necessary to determine multiple target donors that are compatible with the receptor, the receptor sample feature data corresponding to the receptor to be matched and the donor sample feature data corresponding to multiple donors can be retrieved from the preset storage space.

[0073] S250. For each candidate donor, the matching data of the recipient sample and the donor sample corresponding to the current candidate donor are calculated based on the pre-trained donor-recipient matching model to obtain the microbial compatibility attribute and clinical suitability attribute.

[0074] Among them, the donor-recipient matching model refers to an intelligent matching system built through machine learning algorithms. Based on multi-dimensional biomedical characteristic data of the recipient and the donor (including genetic characteristics, immune characteristics and clinical characteristics, etc.), the model calculates the matching degree between the two parties in terms of microbial compatibility and clinical suitability, thereby quantitatively assessing the transplantation suitability between the recipient and each potential donor, and providing a predictive analysis model for optimal donor screening decision support for fecal microbiota transplantation.

[0075] Among them, microbial compatibility attribute refers to a quantitative indicator that characterizes the degree of matching between the microbial communities of the recipient and the donor. It comprehensively assesses the similarity and complementarity of the fecal microorganisms of both parties in terms of species composition, functional genes and ecological networks, and reflects the potential adaptive level of the donor microbial community in the recipient's body to colonize and integrate functions.

[0076] Clinical compatibility attribute refers to a quantitative indicator that characterizes the degree of matching between the recipient and the donor in terms of clinical characteristics. It comprehensively assesses the compatibility and complementarity of the two parties in terms of clinical medical characteristics such as disease type, physiological parameters, medication history, and lifestyle habits, and reflects the potential therapeutic compatibility of the donor microbiota in improving specific clinical symptoms of the recipient.

[0077] For each candidate donor, the matching process between the recipient and each candidate donor is consistent. To clearly illustrate this technical solution, we will use one candidate donor as an example in the following explanation. The current candidate donor refers to the candidate donor currently being matched. In practical applications, a trained and optimized donor-recipient matching model can be used to comprehensively analyze multidimensional data such as microbiome characteristics, immune characteristics, and clinical characteristics of the recipient to be matched and the current candidate donor through specific algorithms. This results in the output of two key matching indicators: microbial compatibility attributes reflecting the microbial community structure and functional compatibility of both parties, and clinical fit attributes assessing the degree of fit between the clinical characteristics and treatment needs of both parties.

[0078] More specifically, the donor-recipient matching model may include: a microbial compatibility assessment model and a clinical suitability assessment model; the recipient gene characteristic information includes: a microbial co-occurrence network diagram and functional gene prediction characteristic information. Optionally, the specific steps for performing matching calculations on the recipient sample characteristic data and the donor sample characteristic data corresponding to the current candidate donor based on the pre-trained donor-recipient matching model to obtain microbial compatibility attributes and clinical suitability attributes may include:

[0079] S2501. Based on the relative abundance characteristics of recipient microbial species, recipient microbial co-occurrence network characteristics in the recipient sample characteristic data, the relative abundance characteristics of donor microbial species, donor microbial co-occurrence network characteristics in the donor sample characteristic data, and the microbial compatibility assessment model, determine the microbial compatibility attributes between the recipient to be treated and the current candidate donor.

[0080] Among them, the microbial compatibility assessment model refers to a computational model used to quantitatively assess the compatibility of the microbial communities between the recipient and the donor. This model analyzes data such as the composition, abundance distribution, and ecological network characteristics of the microbial species of both parties, and uses a specific compatibility assessment algorithm to calculate a matching index that reflects the colonization potential and functional integration ability of the donor microbiota in the recipient's intestinal environment, providing a scientific basis for the selection of fecal microbiota transplantation donors.

[0081] In this embodiment, a pre-established microbial compatibility assessment model can be used to comprehensively evaluate the matching degree of the microbial communities of the two parties in terms of structural characteristics and functional relationships by calculating the similarity and complementarity between the recipient and the donor in two dimensions: microbial species composition (relative abundance characteristics) and ecological interaction relationship (co-occurrence network characteristics). Finally, a numerical index that quantitatively characterizes the compatibility of the microbial communities of the recipient and the donor is output, which is the microbial compatibility attribute between the recipient to be treated and the current candidate donor.

[0082] Optionally, the specific implementation of determining the microbial compatibility attributes between the recipient to be treated and the current candidate donor through a microbial compatibility assessment model may include:

[0083] S11. The relative abundance data and evolutionary relationship encoding information of each microbial species in the receptor sample data are configured as the node parameter information of the corresponding species in the receptor microbial co-occurrence network diagram to obtain the receptor characteristic microbial co-occurrence network diagram.

[0084] Among them, evolutionary relationship coding information refers to the data form that transforms the evolutionary distance and kinship between microbial species into numerical representations. It can be based on multiple alignments of 16S rRNA gene sequences or other conserved gene sequences and phylogenetic tree construction, and uses specific coding methods to quantify the relative positions and genetic associations of different species in the evolutionary lineage.

[0085] Among them, the receptor-characteristic microbial co-occurrence network graph refers to an enhanced network structure representation that integrates the relative abundance data of microbial species and the encoding information of evolutionary relationships in the receptor sample. The nodes represent specific microbial species and carry the abundance characteristics and phylogenetic position information of the species, while the edges represent statistically verified ecological interaction relationships between species. This network graph, by fusing multi-dimensional microbial characteristic data, provides a comprehensive topological representation that includes quantitative species composition and evolutionary background for subsequent graph neural network processing.

[0086] In this embodiment, the relative abundance data of each microbial species detected in the recipient sample and the evolutionary relationship encoding information of each species can be jointly encoded into the feature vector of the recipient microbial co-occurrence network graph. These enhanced feature vectors are assigned to the corresponding species nodes in the co-occurrence network, thereby constructing a network representation that includes both species ecological associations and integrates quantitative composition and evolutionary characteristics, thus obtaining the recipient characteristic microbial co-occurrence network graph. This processing realizes the structured integration and networked representation of multi-dimensional data of microbial communities.

[0087] S12. The relative abundance data and evolutionary relationship encoding information of each microbial species in the donor sample data are configured as the node parameter information of the corresponding species in the donor microbial co-occurrence network diagram to obtain the donor characteristic microbial co-occurrence network diagram.

[0088] Specifically, the method for determining the co-occurrence network diagram of donor characteristic microorganisms is the same as the method for determining the co-occurrence network diagram of recipient characteristic microorganisms. The specific implementation method can be found in S11, and will not be repeated here.

[0089] S13. Input the recipient-featured microbial co-occurrence network diagram and the donor-featured microbial co-occurrence network diagram into the pre-trained graph neural network to obtain the recipient microbial feature embedding and the donor microbial feature embedding, respectively.

[0090] Microbial feature embedding refers to the low-dimensional vector representation obtained after deep feature learning of the co-occurrence network graph of characteristic microorganisms. This vector captures comprehensive features of the microbial community in terms of species composition, abundance distribution, evolutionary relationships, and ecological interactions through the nonlinear transformation of the neural network. It compresses the complex network topology and node attribute information into a dense numerical vector with rich semantics, providing a standardized feature representation for subsequent compatibility comparisons. Recipient microbial feature embedding refers to the microbial feature embedding obtained by feature extraction from the co-occurrence network graph of recipient characteristic microorganisms; donor microbial feature embedding refers to the microbial feature embedding obtained by feature extraction from the co-occurrence network graph of donor characteristic microorganisms.

[0091] In this embodiment, a trained graph neural network can be used to perform deep feature extraction and dimensionality reduction on the co-occurrence network graph of recipient-characteristic microorganisms to obtain recipient microorganism feature embeddings. Similarly, a graph neural network can be used to perform deep feature extraction and dimensionality reduction on the co-occurrence network graph of donor-characteristic microorganisms to obtain donor microorganism feature embeddings. In this way, through the hierarchical information transmission and aggregation mechanism of the network, the complex network structure data is transformed into a low-dimensional dense vector representation that can comprehensively characterize the features of each microbial community.

[0092] S14. Based on the similarity attribute between the recipient microbial feature embedding and the donor microbial feature embedding, determine the microbial compatibility attribute between the recipient to be matched and the current candidate donor.

[0093] In this embodiment, the degree of matching between the microbial communities of the recipient and donor can be quantitatively assessed by calculating the distance or similarity index (such as cosine similarity or Euclidean distance) between the recipient microbial feature embedding and the donor microbial feature embedding in the embedding space. This results in a microbial compatibility attribute that can objectively reflect the potential of the current candidate donor microbial community to colonize and integrate functions in the intestinal environment of the recipient to be matched.

[0094] S2502. Based on the receptor functional gene prediction feature information, receptor immune feature information, receptor clinical feature information in the receptor sample feature data, the donor functional gene prediction feature information, donor immune feature information, donor clinical feature information in the donor sample feature data, and the clinical fit assessment model, determine the clinical fit attributes between the receptor to be treated and the current candidate donor.

[0095] Among them, the clinical suitability assessment model refers to a computational model used to assess the degree of matching between the recipient and the donor in terms of clinical characteristics. This model analyzes multidimensional clinical data of both parties in terms of disease characteristics, immune status, medication history, and lifestyle habits, and uses a specific matching algorithm to calculate a suitability index that reflects the potential of the donor's microbiota to improve specific clinical symptoms of the recipient, thus providing a basis for clinical suitability assessment for fecal microbiota transplantation.

[0096] In this embodiment, a clinical compatibility assessment model can be used to comprehensively analyze the degree of matching between the recipient to be matched and the current candidate donor in multiple dimensions such as functional gene composition, immune status and clinical characteristics. The potential improvement effect and therapeutic compatibility of the donor microbiota on the specific disease symptoms of the recipient can be quantitatively assessed. Finally, a quantitative index reflecting the clinical treatment suitability of both parties is output, which is the clinical compatibility attribute between the recipient to be treated and the current candidate donor.

[0097] Optionally, the specific implementation methods for determining the clinical compatibility attributes between the recipient to be treated and the current candidate donor through a clinical fit assessment model may include:

[0098] S21. Based on the receptor functional gene prediction feature information, receptor immune feature information and receptor clinical feature information in the receptor sample data, construct the receptor to be processed vector.

[0099] Among them, the receptor vector to be processed refers to the structured numerical vector formed by standardizing and feature fusion of multi-dimensional medical indicators such as functional gene prediction features, immune features and clinical features in receptor sample data.

[0100] In this embodiment, the feature data of the recipient in three dimensions—receptor functional gene prediction features, recipient immune features, and recipient clinical features—can be standardized and fused to form a unified numerical vector representation, thus obtaining the recipient processing vector. This vector, through specific mathematical transformations and dimensionality reduction methods, converts multi-source heterogeneous clinical and biological features into standardized numerical features that can be processed by machines, providing structured input data for subsequent transplant suitability prediction.

[0101] S22. Based on the donor functional gene prediction feature information, donor immune feature information, and donor clinical feature information in the donor sample data, construct the donor to be processed vector.

[0102] Among them, the donor vector to be processed refers to the structured feature vector formed by transforming multi-dimensional biomedical indicators such as functional gene prediction features, immune features and clinical features in donor sample data through feature engineering and numerical standardization.

[0103] Specifically, the method for determining the donor vector to be processed is the same as the method for determining the recipient vector to be processed. The specific implementation method can be found in S21, and will not be repeated here.

[0104] S23. Input the recipient vector to be processed and the donor vector to be processed into the pre-trained Transformer model to obtain the postoperative infection risk probability and rejection risk probability.

[0105] Postoperative infection risk probability refers to a numerical indicator used to quantify the likelihood of a recipient developing a pathogenic microbial infection after undergoing fecal microbiota transplantation. Rejection risk probability refers to a numerical indicator used to quantify the likelihood of the recipient's immune system producing an abnormal immune response to the transplanted fecal microbiota.

[0106] In this embodiment, a trained Transformer deep learning model can be used to perform joint analysis and deep feature extraction on the recipient vector to be processed and the donor vector to be processed. The model’s self-attention mechanism captures the complex interaction between the two in multiple dimensions such as functional genes, immune features and clinical features. Finally, two probabilities are output, which respectively quantify the probability of postoperative infection and the probability of immune rejection.

[0107] S24. Based on the postoperative infection risk probability and rejection risk probability, determine the clinical compatibility attributes between the recipient to be treated and the current candidate donor.

[0108] In this embodiment, by comprehensively evaluating two key clinical indicators—the probability of postoperative infection and the probability of rejection—predicted by the evaluation model, and employing a specific decision-making algorithm or scoring rule, these two risk probability values ​​are transformed into a comprehensive clinical fit score, which represents the clinical fit attribute between the recipient to be treated and the current candidate donor. This score quantitatively reflects the overall balance between the potential clinical therapeutic effect and safety that can be achieved after the current donor microbiota is transplanted into the recipient.

[0109] S260. Based on the disease type and severity label information, microbial compatibility attributes, and clinical suitability attributes of the recipient to be matched, determine the matching attributes between the recipient to be matched and the current candidate donor.

[0110] Among them, the matching attribute refers to the quantitative index obtained by comprehensively evaluating the microbial compatibility attribute, clinical suitability attribute and disease severity weight. This index reflects the overall matching degree between the recipient to be matched and the specific donor in three dimensions: microbial community compatibility, clinical treatment suitability and disease-specific needs. It is the final comprehensive evaluation result that determines the priority of donor selection.

[0111] In this embodiment, the evaluation weights of microbial compatibility and clinical suitability can be dynamically adjusted based on disease type information and severity label information. The evaluation indicators of the two dimensions are comprehensively analyzed through a specific weighted fusion algorithm, and finally a quantitative comprehensive matching score is generated, which is the matching attribute between the recipient to be matched and the current candidate donor.

[0112] Optionally, the specific implementation steps of S260 may include:

[0113] S2601. Based on the disease type information and severity label information of the receptor to be matched, determine the microbial compatibility weight and clinical fit weight.

[0114] Microbial compatibility weight refers to the relative importance coefficient assigned to the microbial compatibility attribute. Clinical adaptability weight refers to the relative importance coefficient assigned to the clinical adaptability attribute.

[0115] In this embodiment, based on the disease type and severity label information of the recipient to be matched, the relative importance ratio of the two key dimensions, microbial compatibility assessment and clinical suitability assessment, in the final matching decision can be dynamically calculated and determined through preset weight allocation rules. These are the microbial compatibility weight and clinical suitability weight, thereby achieving differentiated control of the contribution of each assessment factor under different disease states.

[0116] Optionally, the specific implementation methods for determining microbial compatibility weights and clinical fit weights based on the disease type and severity label information of the receptors to be matched may include:

[0117] If the recipient to be matched is determined to be a newly diagnosed patient based on the disease type and severity label information, then the microbial compatibility weight is determined from the first preset value range, and the clinical fit weight is determined from the second preset value range; if the recipient to be matched is determined to be a refractory patient based on the disease type and severity label information, then the microbial compatibility weight is determined from the third preset value range, and the clinical fit weight is determined from the fourth preset value range.

[0118] Among them, the values ​​in the first preset value range are all greater than the values ​​in the second preset value range, and the values ​​in the third preset value range are all less than the values ​​in the fourth preset value range.

[0119] In this embodiment, a differentiated weighting strategy can be adopted, that is, the importance weights of microbial compatibility and clinical suitability can be dynamically adjusted according to disease type information and severity label information (new onset or refractory): for newly diagnosed patients, microbial compatibility can be given a higher weight (the value of the first preset numerical range > the value of the second preset numerical range), emphasizing the fundamental role of microbial matching; while for refractory patients, more emphasis is placed on clinical suitability (the value of the fourth preset numerical range > the value of the third preset numerical range), highlighting the precision of treatment effect. This differentiated design reflects a personalized matching strategy for patients at different stages of disease.

[0120] S2602. Based on microbial compatibility weight and clinical fit weight, perform weighted calculation on microbial compatibility attributes and clinical fit attributes to determine the matching attributes between the recipient to be matched and the current candidate donor.

[0121] In this embodiment, the microbial compatibility score and the clinical suitability score can be multiplied by their respective weight coefficients and then added together. Through this weighted summation calculation method, the evaluation results of the two dimensions are integrated according to their importance ratio, and finally a comprehensive matching attribute score is generated. This score quantifies the overall suitability of the donor and recipient in terms of both microbial community and clinical characteristics.

[0122] S270. Based on the matching attributes between the recipient to be matched and each donor, identify multiple target donors that are compatible with the recipient to be matched.

[0123] In this embodiment, the matching attributes between all candidate donors and recipients can be screened or sorted according to a pre-set matching threshold or ranking rules. Several candidate donors that meet the required matching standards or rank highly are selected as the final recommended transplant recipients. These recipients are the multiple target donors that are compatible with the recipient. These target donors demonstrate the best matching degree with the recipient in terms of microbial compatibility and clinical suitability.

[0124] S280. In response to a donor-recipient matching request, acquire post-transplant monitoring data of the recipient to be matched at the current preset sampling time.

[0125] The fecal microbiota transplantation (FMT) monitoring data includes recipient symptom scores and microbiota colonization rates. The FMT monitoring data is determined based on the fecal microbiota of any target donor after the fecal microbiota transplantation surgery is performed on the target recipient.

[0126] S290. Input the fecal microbiota transplantation monitoring data into the transplantation strategy control model, and output the transplantation adjustment strategy for the next period of time to adjust the matching recipient.

[0127] The transplantation adjustment strategy includes at least one of the following: donor switching strategy and dose adjustment strategy.

[0128] The technical solution of this invention, when determining multiple target donors, acquires receptor sample feature data corresponding to the recipient to be matched and donor sample feature data corresponding to multiple candidate donors. The receptor sample feature data includes: disease type and severity identifiers, receptor gene feature information, receptor immune feature information, and receptor clinical feature information. The donor sample feature data includes: donor gene feature information, donor immune feature information, and donor clinical feature information. For each candidate donor, a matching calculation is performed on the receptor sample feature data and the donor sample feature data corresponding to the current candidate donor based on a pre-trained donor-recipient matching model to obtain microbial compatibility attributes and clinical suitability attributes. Based on the disease type and severity label information, microbial compatibility attributes, and clinical suitability attributes of the recipient to be matched, the matching attributes between the recipient to be matched and the current candidate donor are determined. Based on the matching attributes between the recipient to be matched and each candidate donor, multiple target donors suitable for the recipient to be matched are determined. The technical solution provided in this embodiment achieves scientific and objective donor screening by integrating quantitative assessments of both microbial compatibility and clinical suitability, combined with disease-specific weight adjustments. The multi-parameter comprehensive scoring mechanism ensures the comprehensiveness and accuracy of the matching results. At the same time, it retains multiple suitable donor selection options, providing flexible and diverse transplantation choices for clinical treatment, and dynamically adjusting the donor strategy according to the patient's postoperative response, significantly improving the personalization and success rate of fecal microbiota transplantation therapy.

[0129] Figure 3 This is a schematic diagram of a dynamic matching method for fecal microbiota transplantation donors and recipients provided in an embodiment of the present invention. Based on the foregoing embodiments, the specific implementation process of S120 is further refined, and the specific implementation method can be found in the technical solution of this embodiment. Technical terms that are the same as or corresponding to those in the above embodiments will not be repeated here.

[0130] like Figure 3 As shown, the method specifically includes the following steps:

[0131] S310. In response to a donor-recipient matching request, acquire post-transplantation monitoring data of the recipient to be matched at the current preset sampling time.

[0132] The fecal microbiota transplantation (FMT) monitoring data includes recipient symptom scores and microbiota colonization rates. The FMT monitoring data is determined based on the fecal microbiota of any target donor after the FMT surgery is performed on the candidate recipient. The target donor is determined by matching multiple donors based on the genetic, immune, and clinical characteristics of the candidate recipient and multiple candidate donors.

[0133] In this embodiment, the transplantation strategy control model includes an agent decision network and a virtual environment network. The agent decision network includes an action value mapping layer and an action decision layer. Specifically, the transplantation strategy control model can be a decision system built using a reinforcement learning framework. The agent decision network evaluates the value parameters of each potential transplantation strategy through the action value mapping layer, and the action decision layer selects the optimal action to execute. The virtual environment network is a digital simulation system that simulates the physiological response after fecal microbiota transplantation. By receiving action instructions from the decision network, it simulates and calculates clinical effects and provides feedback on state updates and reward signals. The two work together to achieve iterative optimization of the transplantation strategy.

[0134] S320. Use the monitoring data after fecal microbiota transplantation as the initial state information, and use the initial state information as the current state information.

[0135] Among them, the current state information refers to the dynamic data representation of the receptor's comprehensive clinical status and microbial colonization at a certain decision moment during the strategy optimization iteration process.

[0136] In this embodiment, the monitoring data after fecal microbiota transplantation is standardized to form a structured data representation. This information can be used as the starting input for the reinforcement learning process, fully characterizing the recipient's immediate clinical status and microbial colonization after receiving fecal microbiota transplantation from the current donor, and providing a basic state benchmark for subsequent strategy optimization.

[0137] S330. Input the current state information into the action value mapping layer to obtain at least one preset execution action corresponding to the value parameter.

[0138] In this context, the pre-defined execution actions refer to a set of optional operations predefined during the transplantation strategy optimization process. More specifically, at least one pre-defined execution action includes: switching to each of the multiple target donors, and a random combination with at least one dose adjustment method. During the decision-making process, the pre-set set of optional treatment options includes the basic operation of switching to a specific target donor, as well as composite treatment options formed by combining donor switching with different dose adjustment strategies (such as incremental, decremental, or maintenance). This design ensures that the decision space covers all possible clinical intervention methods. The value parameter refers to a numerical indicator calculated through the action value mapping layer, which quantitatively characterizes the expected effect of each pre-defined execution action in the current state. This parameter comprehensively evaluates the potential benefits of candidate treatment options in improving clinical symptoms, maintaining microbiome balance, and reducing adverse reactions.

[0139] In this embodiment, the initial state information representing the current clinical state of the recipient is used as the current state information and input into the intelligent agent decision network, including the action value mapping layer. The built-in algorithm calculates the value parameters of each preset action in the current state. These value parameters quantitatively reflect the balance between the therapeutic effect and risk that each candidate action may bring.

[0140] S340. Input each value parameter into the action decision layer to obtain the target value parameter that meets the preset conditions, and determine the preset execution action corresponding to the target value parameter as the current execution action.

[0141] The current action refers to the optimal treatment plan selected from a pre-set set of actions by the agent decision-making network during the decision-making process. It represents the donor switching or dosage adjustment strategy that best optimizes the balance between treatment efficacy and safety in the current state, and is the immediate clinical intervention recommended by the system.

[0142] In this embodiment, each value parameter can be input to the action decision layer. The action decision layer filters and compares the value parameters of all preset execution actions, and selects the optimal score that meets the optimization criteria (such as the highest value or the lowest risk). This score is the target value parameter. Thus, the preset execution action corresponding to the target value parameter can be determined as the current optimal decision and used as the current execution action.

[0143] S350. After the current action is executed in the virtual environment network simulation, output the action reward value and the next state information corresponding to the current action.

[0144] The action reward value is a numerical evaluation index that quantifies the expected effect of the current action (donor switching or dose adjustment strategy). The action reward value is determined based on the clinical remission prediction rate, adverse reaction prediction rate, and gut microbiota diversity index prediction value. The action reward value comprehensively considers the combined benefits of the current action in improving clinical symptoms (clinical remission prediction rate), reducing adverse reactions (adverse reaction prediction rate), and optimizing gut microbiota balance (guest microbiota diversity index prediction value), serving as a key feedback signal for strategy optimization during reinforcement learning.

[0145] The next state information refers to the clinical state data representation of the recipient that is predicted to appear in the next time period after the current action is performed through virtual environment network simulation. This information includes key indicators such as the updated symptom score and microbial colonization rate calculated by simulation, reflecting the expected changes in the recipient's state after the current transplant adjustment strategy is adopted.

[0146] In this embodiment, after the current action (i.e., donor switching or dose adjustment) is virtually simulated through a virtual environment network, two key outputs are generated: an action reward value that quantifies the overall benefits of the strategy in terms of symptom improvement, safety, and microbiome balance; and next state information that predicts the new clinical state that the recipient will reach after implementing the strategy. Together, these two outputs constitute the feedback signal for iterative optimization of the strategy.

[0147] S360. Update the next state information to the current state information, and repeatedly execute the steps of determining the value parameter corresponding to at least one preset execution action, determining it as the current execution action, and determining the action reward value and next state information corresponding to the current execution action, until the action reward value obtained by the loop multiple times meets the preset conditions, and use the current execution action as the transplantation adjustment strategy for adjusting the recipient to be matched in the next time period.

[0148] Specifically, through continuous state updates and action evaluation cycles—that is, using the predicted next state as the new current state, recalculating action values, and selecting the optimal action—the effects of different treatment strategies are simulated multiple times in a virtual environment. The iteration terminates when the accumulated action reward value reaches a preset optimal standard (such as convergence or maximization), and the final output is used to adjust the transplantation strategy for the matched recipient in the next time period. This process achieves dynamic optimization and validation of the treatment plan in a simulated environment.

[0149] The technical solution of this invention, when determining the transplantation adjustment strategy for the matched recipient in the next time period, uses post-fecal microbiota transplantation monitoring data as initial state information and uses the initial state information as the current state information. Then, the current state information is input to the action value mapping layer to obtain value parameters corresponding to at least one preset execution action. The at least one preset execution action includes: switching to each of multiple target donors, and a random combination with at least one dose adjustment method. Further, each value parameter is input to the action decision layer to obtain target value parameters that meet preset conditions. The preset execution action corresponding to the target value parameter is determined as the current execution action. Subsequently, after the current action is executed in a virtual network simulation, the corresponding action reward value and next state information are output. The action reward value is determined based on the clinical remission prediction rate, adverse reaction prediction rate, and microbial diversity index prediction value. Then, the next state information is updated to the current state information, and the steps of determining the value parameter corresponding to at least one preset action, determining it as the current action, and determining the corresponding action reward value and next state information are repeated until the action reward value obtained from multiple iterations meets the preset conditions. At this point, the current action is used as the transplantation adjustment strategy for the matched recipient in the next time period. The technical solution provided in this embodiment firstly verifies the clinical effects of different treatment plans in advance through multiple rounds of virtual simulation, significantly reducing actual treatment risks; secondly, dynamic weighting integrates multi-dimensional indicators such as symptom improvement, safety, and microbial balance to ensure the scientific and comprehensive nature of strategy selection; and finally, the iterative optimization mechanism can adaptively adjust donor selection and dosage regimens, providing personalized treatment pathways for complex cases and improving the overall accuracy and efficacy stability of fecal microbiota transplantation. This method deeply integrates the predictive capabilities of artificial intelligence with clinical decision-making needs, pioneering a new paradigm for intelligent fecal microbiota transplantation therapy.

[0150] Figure 4 This is a schematic diagram of a dynamic matching device for fecal microbiota transplantation donor and recipient provided in an embodiment of the present invention. The device includes: a monitoring data acquisition module 410 and an adjustment strategy determination module 420.

[0151] The monitoring data acquisition module 410 is used to acquire the fecal microbiota transplantation (FMT) monitoring data of the recipient at the current preset sampling time in response to a donor-recipient matching request. The FMT monitoring data includes the recipient's symptom score and microbiota colonization rate. The FMT monitoring data is determined after the recipient has undergone FMT surgery based on the fecal microbiota of any target donor. The target donor is a group of donors determined by adaptability matching based on the genetic characteristics, immune characteristics, and clinical characteristics of the recipient and multiple candidate donors.

[0152] The adjustment strategy determination module 420 is used to input the fecal microbiota transplantation post-transplantation monitoring data into the transplantation strategy control model and output the transplantation adjustment strategy for adjusting the recipient to be matched in the next time period; wherein, the transplantation adjustment strategy includes at least one of the following: donor switching strategy and dose adjustment strategy.

[0153] The technical solution of this invention, in response to a donor-recipient matching request, acquires fecal microbiota transplantation (FMT) monitoring data of the recipient at the current preset sampling time. The FMT monitoring data includes the recipient's symptom score and microbiota colonization rate. This data is determined based on the fecal microbiota of any target donor after the recipient has undergone FMT surgery. The target donor is determined by matching the recipient with multiple donors based on their genetic, immune, and clinical characteristics. Further, the FMT monitoring data is input into a transplantation strategy control model, which outputs a transplantation adjustment strategy for adjusting the recipient's transplantation in the next time period. This transplantation adjustment strategy includes at least one of a donor switching strategy and a dose adjustment strategy. The technical solution provided in this embodiment, based on a dynamic feedback mechanism of real-time post-transplant monitoring data, continuously optimizes donor selection and dosage adjustment through a transplantation strategy control model, solving the pain point that static matching cannot adapt to dynamic changes after transplantation. In addition, it integrates multimodal data such as genetic characteristics, immune characteristics, and clinical characteristics, breaking through the limitations of single microbiome data. Through multimodal data integration, it achieves comprehensive compatibility assessment and continuously optimizes treatment plans with the help of dynamic feedback mechanisms, significantly improving the accuracy and adaptability of FMT treatment.

[0154] The fecal microbiota transplantation donor-recipient dynamic matching device provided in the embodiments of the present invention can execute the fecal microbiota transplantation donor-recipient dynamic matching method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.

[0155] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Figure 5 The electronic device 50 shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the present invention. Figure 5As shown, electronic device 50 is represented in the form of a general-purpose computing device. Components of electronic device 50 may include, but are not limited to: one or more processors or processing units 501, system memory 502, and a bus 503 connecting different system components (including system memory 502 and processing unit 501). Bus 503 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus. Electronic device 50 typically includes various computer system readable media. These media can be any available media that can be accessed by electronic device 50, including volatile and non-volatile media, removable and non-removable media. System memory 502 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 504 and / or cache memory 505. Electronic device 50 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 506 can be used to read and write non-removable, non-volatile magnetic media (…). Figure 5 Not shown; usually referred to as a "hard drive"). Although Figure 5Not shown, disk drives for reading and writing to removable non-volatile disks (e.g., "floppy disks") and optical disc drives for reading and writing to removable non-volatile optical discs (e.g., CD-ROMs, DVD-ROMs, or other optical media) may be provided. In these cases, each drive may be connected to bus 503 via one or more data media interfaces. Memory 502 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention. A program / utility 508 having a set (at least one) of program modules 507 may be stored, for example, in memory 502. Such program modules 507 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 507 generally perform the functions and / or methods described in the embodiments of the present invention. Electronic device 50 can also communicate with one or more external devices 509 (e.g., keyboard, pointing device, display 810, etc.), and with one or more devices that enable a user to interact with electronic device 50, and / or with any device that enables electronic device 50 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 511. Furthermore, electronic device 50 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 512. As shown, network adapter 512 communicates with other modules of electronic device 50 via bus 503. It should be understood that, although... Figure 5 As not shown, other hardware and / or software modules can be used in conjunction with electronic device 50, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems. Processing unit 501 executes various functional applications and page processing by running programs stored in system memory 502, such as implementing the fecal microbiota transplantation donor-recipient dynamic matching method provided in embodiments of the present invention. In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via I / O interface 511, or installed from storage system 506. When the computer program is executed by processing unit 501, it performs the functions defined in the methods of embodiments of the present invention.

[0156] This invention also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the aforementioned dynamic matching method for fecal microbiota transplantation donors and recipients.

Claims

1. A method for dynamic matching of donor and recipient in fecal microbiota transplantation, characterized in that, include: In response to a donor-recipient matching request, the system acquires fecal microbiota transplantation (FMT) post-transplantation monitoring data of the recipient at the current preset sampling time. The FMT post-transplantation monitoring data includes recipient symptom scores and microbiota colonization rates. This data is determined based on the fecal microbiota of any target donor after the FMT surgery is performed on the recipient. The target donor is one of multiple donors identified through compatibility matching based on the recipient's genetic characteristics, immune characteristics, and clinical characteristics with multiple candidate donors. The fecal microbiota transplantation post-transplantation monitoring data is input into the transplantation strategy control model, which outputs a transplantation adjustment strategy for adjusting the matching recipient in the next time period; wherein, the transplantation adjustment strategy includes at least one of: donor switching strategy and dose adjustment strategy; The transplantation strategy control model includes an agent decision network and a virtual environment network. The agent decision network includes an action value mapping layer and an action decision layer. The process of inputting the fecal microbiota transplantation monitoring data into the transplantation strategy control model and outputting a transplantation adjustment strategy for adjusting the recipient in the next time period includes: The monitoring data after fecal microbiota transplantation is used as the initial state information, and the initial state information is used as the current state information; The current state information is input to the action value mapping layer to obtain value parameters corresponding to at least one preset execution action; wherein, the at least one preset execution action includes: switching to each of the multiple target donors, and a random combination with at least one dose adjustment method; Each of the value parameters is input to the action decision layer to obtain the target value parameters that meet the preset conditions, and the preset execution action corresponding to the target value parameters is determined as the current execution action; After the current action is executed in the virtual environment network simulation, the action reward value and the next state information corresponding to the current action are output; wherein, the action reward value is determined based on the clinical remission prediction rate, adverse reaction prediction rate and the microbial diversity index prediction value; The next state information is updated to the current state information, and the steps of determining the value parameter corresponding to the at least one preset execution action, determining the current execution action, and determining the action reward value and next state information corresponding to the current execution action are repeated until the action reward value obtained by multiple iterations meets the preset conditions. Then, the current execution action is used as the transplantation adjustment strategy for the matching recipient in the next time period.

2. The method according to claim 1, characterized in that, The target donor was determined in the following manner: The method involves acquiring receptor sample feature data corresponding to the target receptor and donor sample feature data corresponding to multiple candidate donors; wherein, the receptor sample feature data includes: disease type and severity identifier, receptor gene feature information, receptor immune feature information, and receptor clinical feature information; the donor sample feature data includes: donor gene feature information, donor immune feature information, and donor clinical feature information. For each of the candidate donors, a matching calculation is performed on the recipient sample feature data and the donor sample feature data corresponding to the current candidate donor based on a pre-trained donor-recipient matching model to obtain microbial compatibility attributes and clinical suitability attributes. Based on the disease type and severity identifier of the candidate receptor, the microbial compatibility attribute, and the clinical fit attribute, the matching attributes between the candidate receptor and the current candidate donor are determined; Based on the matching attributes between the recipient to be matched and each candidate donor, multiple target donors that are compatible with the recipient to be matched are identified.

3. The method according to claim 2, characterized in that, The method further includes: Obtain receptor sample data corresponding to the candidate receptor and donor sample data corresponding to the multiple candidate donors; wherein, the receptor sample data includes: receptor fecal microbial gene data, receptor immune status data and receptor clinical metadata; the donor sample data includes donor fecal microbial gene data, donor immune status data and recipient clinical metadata. The receptor sample data were processed by feature extraction to obtain receptor gene feature information, receptor immune feature information and receptor clinical feature information; Feature extraction processing is performed on the data of each donor sample to obtain donor gene feature information, donor immune feature information and donor clinical feature information corresponding to each candidate donor.

4. The method according to claim 3, characterized in that, The recipient fecal microbial gene data includes: metagenomic sequencing data, species composition types of recipient fecal samples, and relative abundance data of each species. The recipient gene feature information includes: relative abundance features of microbial species, microbial co-occurrence network diagram, and functional gene prediction feature information. Feature extraction processing is performed on the receptor sample data to obtain receptor gene feature information, including: The species composition types and relative abundance data of each species in the recipient fecal samples were processed by feature vectorization to obtain the relative abundance features of microbial species. The species composition types and relative abundance data of each species in the recipient fecal samples were processed by a co-occurrence network analysis tool to construct a microbial co-occurrence network diagram. In the microbial co-occurrence network diagram, the nodes represent different species, and the connecting edges between nodes represent significant correlations between species. Gene function prediction analysis was performed on the metagenomic sequencing data using the PICRUSt2 functional prediction tool to obtain functional gene prediction results. The functional gene prediction results were then processed into feature vectors to obtain functional gene prediction feature information.

5. The method according to claim 3, characterized in that, The receptor sample data is subjected to feature extraction processing to obtain receptor clinical feature information, including: The receptor clinical metadata is processed by feature vectorization based on a bidirectional encoder representation transformation model to obtain clinical data feature vectors. The receptor clinical metadata includes at least one of the following: receptor attribute information, clinical symptom description information, receptor disease type, past medical history information, family medical history information, recent medication information, lifestyle information, human leukocyte antigen typing, and blood type information.

6. The method according to claim 2, characterized in that, The donor-recipient matching model includes a microbial compatibility assessment model and a clinical suitability assessment model, and the recipient gene feature information includes a microbial co-occurrence network diagram and functional gene prediction feature information. Based on a pre-trained donor-recipient matching model, the matching calculation is performed on the recipient sample feature data and the donor sample feature data corresponding to the current candidate donor to obtain microbial compatibility attributes and clinical suitability attributes, including: Based on the relative abundance characteristics of recipient microbial species and recipient microbial co-occurrence network characteristics in the recipient sample feature data, the relative abundance characteristics of donor microbial species and donor microbial co-occurrence network characteristics in the donor sample feature data, and the microbial compatibility assessment model, the microbial compatibility attributes between the recipient to be treated and the current candidate donor are determined. Based on the receptor functional gene prediction feature information, receptor immune feature information, and receptor clinical feature information in the receptor sample feature data, the donor functional gene prediction feature information, donor immune feature information, and donor clinical feature information in the donor sample feature data, as well as the clinical fit assessment model, the clinical fit attributes between the receptor to be treated and the current candidate donor are determined.

7. The method according to claim 6, characterized in that, The determination of microbial compatibility attributes between the recipient and the current candidate donor is based on the relative abundance characteristics of recipient microbial species, recipient microbial co-occurrence network characteristics in the recipient sample feature data, the relative abundance characteristics of donor microbial species, donor microbial co-occurrence network characteristics in the donor sample feature data, and the microbial compatibility assessment model. This includes: The relative abundance data and evolutionary relationship encoding information of each microbial species in the receptor sample data are configured as the node parameter information of the species corresponding to the receptor microbial co-occurrence network graph feature information to obtain the receptor feature microbial co-occurrence network graph. The relative abundance data and evolutionary relationship encoding information of each microbial species in the donor sample data are configured as the node parameter information of the corresponding species in the donor microbial co-occurrence network diagram to obtain the donor characteristic microbial co-occurrence network diagram. The recipient-featured microorganism co-occurrence network diagram and the donor-featured microorganism co-occurrence network diagram are respectively input into a pre-trained graph neural network to obtain the recipient microorganism feature embedding and the donor microorganism feature embedding; Based on the similarity attribute between the recipient microbial feature embedding and the donor microbial feature embedding, the microbial compatibility attribute between the recipient to be matched and the current candidate donor is determined.

8. The method according to claim 6, characterized in that, The determination of the clinical compatibility attributes between the recipient to be treated and the current candidate donor, based on the recipient functional gene prediction feature information, recipient immune feature information, and recipient clinical feature information in the recipient sample feature data, the donor functional gene prediction feature information, donor immune feature information, and donor clinical feature information in the donor sample feature data, and the clinical suitability assessment model, includes: Based on the receptor functional gene prediction feature information, receptor immune feature information and receptor clinical feature information in the receptor sample data, a receptor vector to be processed is constructed. Based on the donor functional gene prediction feature information, donor immune feature information, and donor clinical feature information in the donor sample data, a donor vector to be processed is constructed. The recipient vector to be processed and the donor vector to be processed are input into a pre-trained Transformer model to obtain the postoperative infection risk probability and rejection risk probability. Based on the postoperative infection risk probability and the rejection risk probability, the clinical compatibility attributes between the recipient to be treated and the current candidate donor are determined.

9. The method according to claim 2, characterized in that, The determination of matching attributes between the candidate receptor and the current candidate donor based on the disease type and severity label information, the microbial compatibility attributes, and the clinical suitability attributes includes: Based on the disease type and severity label information of the receptors to be matched, the microbial compatibility weight and clinical fit weight are determined. Based on the microbial compatibility weight and the clinical fit weight, a weighted calculation is performed on the microbial compatibility attribute and the clinical fit attribute to determine the matching attribute between the recipient to be matched and the current candidate donor.

10. The method according to claim 9, characterized in that, The determination of microbial compatibility weights and clinical fit weights based on the disease type and severity label information of the receptors to be matched includes: If, based on the disease type information and severity label information of the recipient to be matched, it is determined that the recipient to be matched is a newly diagnosed patient, then a microbial compatibility weight is determined from a first preset value range, and a clinical fit weight is determined from a second preset value range; wherein, the values ​​in the first preset value range are all greater than the values ​​in the second preset value range. If, based on the disease type information and severity label information of the recipient to be matched, it is determined that the recipient to be matched is a refractory patient, then a microbial compatibility weight is determined from a third preset value range, and a clinical fit weight is determined from a fourth preset value range; wherein, the values ​​in the third preset value range are all the same as the values ​​in the fourth preset value range.

11. A dynamic matching device for fecal microbiota transplantation donors and recipients, characterized in that, The device includes: The monitoring data acquisition module is used to acquire post-fecal microbiota transplantation monitoring data of the recipient at the current preset sampling time in response to a donor-recipient matching request. The post-fecal microbiota transplantation monitoring data includes recipient symptom scores and microbiota colonization rate. This data is determined after the fecal microbiota of any target donor is used to perform fecal microbiota transplantation on the recipient. The target donor is one of multiple donors determined through compatibility matching based on the genetic, immune, and clinical characteristics of the recipient and multiple candidate donors. The adjustment strategy determination module is used to input the post-fecal microbiota transplantation monitoring data into the transplantation strategy control model and output the transplantation adjustment strategy for adjusting the recipient to be matched in the next time period; wherein, the transplantation adjustment strategy includes at least one of the following: donor switching strategy and dose adjustment strategy; The transplantation strategy control model includes an agent decision network and a virtual environment network. The agent decision network includes an action value mapping layer and an action decision layer. The adjustment strategy determination module is used for: The monitoring data after fecal microbiota transplantation is used as the initial state information, and the initial state information is used as the current state information; The current state information is input to the action value mapping layer to obtain value parameters corresponding to at least one preset execution action; wherein, the at least one preset execution action includes: switching to each of the multiple target donors, and a random combination with at least one dose adjustment method; Each of the value parameters is input to the action decision layer to obtain the target value parameters that meet the preset conditions, and the preset execution action corresponding to the target value parameters is determined as the current execution action; After the current action is executed in the virtual environment network simulation, the action reward value and the next state information corresponding to the current action are output; wherein, the action reward value is determined based on the clinical remission prediction rate, adverse reaction prediction rate and the microbial diversity index prediction value; The next state information is updated to the current state information, and the steps of determining the value parameter corresponding to the at least one preset execution action, determining the current execution action, and determining the action reward value and next state information corresponding to the current execution action are repeated until the action reward value obtained by multiple iterations meets the preset conditions. Then, the current execution action is used as the transplantation adjustment strategy for the matching recipient in the next time period.

12. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, which enables the at least one processor to perform the fecal microbiota transplantation donor-recipient dynamic matching method according to any one of claims 1-10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the fecal microbiota transplantation donor-recipient dynamic matching method as described in any one of claims 1-10.

Images