A system and method for monitoring recovery of gastrointestinal function after gastrectomy
By establishing an antibiotic impact model and optimizing the microbial distribution data of patients after gastrectomy, the usability problem of microbial detection caused by antibiotic interference was solved, and the accuracy of monitoring the recovery of gastrointestinal function was improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING FRIENDSHIP HOSPITAL CAPITAL MEDICAL UNIV
- Filing Date
- 2026-03-05
- Publication Date
- 2026-06-02
AI Technical Summary
In existing technologies, the bacterial flora data of patients after gastrectomy are affected by antibiotic interference, resulting in low usability of the bacterial flora data detected by sequencing technology, making it difficult to provide effective reference for medical staff.
By establishing an antibiotic impact model, using DNA testing information, antibiotic usage time and dosage, and sample moisture content, we can optimize microbial distribution data and provide it to medical staff for reference.
It reduces the interference of antibiotics on gut microbiota testing, improves the reference value and usability of gut microbiota data, and helps medical staff to more accurately assess the recovery of patients' gastrointestinal function.
Smart Images
Figure CN122136022A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of gastrointestinal care technology, specifically a system and method for monitoring the recovery of gastrointestinal function after gastrectomy. Background Technology
[0002] Gastrectomy is an important treatment for serious stomach diseases. After gastrectomy, patients need to undergo digestive tract reconstruction and functional adaptation, facing challenges in both short-term recovery and long-term nutritional management. When using gut microbiota testing to assess the postoperative recovery of patients, prophylactic antibiotics used after gastrectomy kill bacteria. The DNA fragments of these killed bacteria can remain in the intestines for a considerable period and be detected by sequencing technology, creating a false impression of "normal gut microbiota composition." In reality, the number of viable bacteria has been significantly reduced, affecting the patient's digestive and absorptive functions. This makes the usability of gut microbiota data detected by sequencing technology low, and difficult to provide effective reference for medical staff. Therefore, how to reduce the interference of antibiotics and improve the usability of gut microbiota data detected by sequencing technology to provide reference for medical staff has become an urgent problem to be solved. Summary of the Invention
[0003] The purpose of this invention is to provide a system and method for monitoring the recovery of gastrointestinal function after gastrectomy, in order to solve the problems raised in the prior art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: a monitoring system for gastrointestinal function recovery after gastrectomy, comprising a detection module, a data storage module, a sample analysis module, and a status analysis module; the detection module is used to perform DNA testing on samples from target patients to obtain the microbial community distribution; the data storage module is used to store sample testing data from historical patients; the sample analysis module is used to optimize the DNA testing data of the target patient samples and provide the optimized microbial community distribution in the samples to medical personnel for reference; the status analysis module is used to analyze the moisture content of the samples to determine the inhibitory effect of antibiotics.
[0005] Specifically, the sample analysis module further includes a population analysis unit and a first model training unit. The population analysis unit acquires the preoperative microbiota data and gastrectomy characteristics of the target patient, obtains the baseline status of the target patient based on the preoperative microbiota data, and obtains the population information of the target patient based on the baseline status and gastrectomy characteristics. The first model training unit takes DNA detection information, antibiotic usage time, antibiotic dosage, and antibiotic inhibitory effect as inputs and RNA detection information as outputs to train an antibiotic influence model. The DNA detection information of the target patient is optimized by inputting the target patient's data into the antibiotic influence model. The state analysis module further includes a transformation unit and a fitting unit: the transformation unit is used to unify the antibiotic inhibitory effect at different scales; the fitting unit is used to fit the sample dryness and humidity and the antibiotic inhibitory effect to establish an antibiotic analysis model. The detection module further includes a moisture detection unit, a DNA detection unit, and a timing unit; the moisture detection unit is used to acquire the water content information of the target patient's sample; the DNA detection unit is used to perform DNA detection on the target patient's sample; and the timing unit is used to acquire the time when the target patient used antibiotics.
[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring the recovery of gastrointestinal function after gastrectomy, comprising the following steps:
[0007] Obtain postoperative samples from the target patient after gastrectomy, perform DNA testing on the postoperative samples, and obtain the bacterial community distribution in the samples;
[0008] Obtain preoperative microbial data and gastrectomy characteristics of target patients; obtain the baseline status of target patients based on preoperative microbial data; and obtain population information of target patients based on baseline status and gastrectomy characteristics.
[0009] Based on the population information of the target patients, an antibiotic impact model is trained; the bacterial community distribution in the sample is optimized according to the antibiotic impact model and the bacterial community distribution in the sample, and the optimized bacterial community distribution in the sample is provided to medical staff as a reference.
[0010] Specifically, obtaining population information of the target patients based on their baseline status and gastrectomy surgical characteristics also includes the following steps:
[0011] S1: Obtain the types of microbiota and the proportion of each microbiota from the baseline status of the target patient, obtain the historical baseline status of other patients, and calculate the deviation F between the target patient and the i-th other patient based on the baseline status. i F i =D i +Σ j P j×C ij In the formula, C ij P represents the characteristics of bacterial community differences. j The weights represent the differential characteristics of the microbial community; j is a positive integer in [1, m], where m represents the total number of microbial species in the baseline state of the target patient and the i-th other patient; D i P represents the Euclidean distance between the percentage of the same bacterial species in the baseline states of the target patient and the i-th other patient; if the j-th bacterial species are different in the baseline states of the target patient and the i-th other patient, then P j If not zero, then P j The value is zero; the j-th bacterial species being different means that only one of the two baseline states of the target patient and the i-th other patient contains the j-th bacterial species; based on the bias, the baseline states of the target patient and other patients are classified unsupervised to determine the taxonomic cluster to which the baseline state of the target patient belongs, and the information of other patients in the taxonomic cluster is extracted to form a set U1;
[0012] S2, obtain the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, calculate the Euclidean distance between the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, perform unsupervised classification on the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients based on the Euclidean distance, determine the classification cluster to which the gastrectomy surgical features of the target patient belong, extract the information of other patients in the classification cluster to form a set U2, obtain the intersection of U1 and U2, and obtain the group information U1∩U2 of the target patient.
[0013] Specifically, in step S1, the weights of the differential characteristics of the microbial community are determined through the following steps:
[0014] Obtain the kth distinct bacterial species from the baseline status of the target patient and the i-th other patient, and denote it as Q. k Get Q k The proportion R k According to R k The weight P of the differential feature of the k-th microbial community is obtained. k , In the formula, n represents the number of bacterial species that are the same in the baseline state of the target patient and the i-th other patient, and R base R represents a reference value indicating the percentage of different bacterial species. base Set to Q k The average percentage.
[0015] Specifically, training the antibiotic impact model based on the target patient population information also includes the following steps:
[0016] Historical testing information of samples is obtained from the target patient population information. This testing information includes DNA testing information, RNA testing information, antibiotic usage time, antibiotic dosage, and sample moisture content. An antibiotic analysis model is established based on the sample moisture content. The historical testing information of the samples is then processed using the antibiotic analysis model to obtain the inhibitory effect of the antibiotics.
[0017] The historical detection information of the processed samples is divided into a training set and a test set. The DNA detection information, antibiotic usage time, antibiotic dosage and antibiotic inhibitory effect in the training set are used as inputs, and the RNA detection information is used as outputs to train the antibiotic impact model. The antibiotic impact model is then validated using data from the test set. The antibiotic impact model is obtained after validation.
[0018] Specifically, the antibiotic analysis model is trained through the following steps:
[0019] S10, from the historical detection information of the i-th other patient sample in the target patient population information, obtain the inhibitory effect Z of antibiotics on the bacterial flora under the same antibiotic usage time and dosage in the u1-th sample under different humidity conditions. u1 To obtain the inhibitory effect of antibiotics on bacterial flora under different sample humidity conditions and Z-strain. u1 The proportional relationship between them, and the inhibitory effect of antibiotics on bacterial flora under different sample dryness and humidity conditions, were all expressed by Z-axis. u1 To represent; give Z u1 Set a baseline value, establish a multinomial regression model, with the sample dryness and humidity as the input and the antibiotics inhibiting the bacterial community under different sample dryness and humidity conditions as the output.
[0020] S20, Obtain historical testing information of the vth other patient sample from the target patient population information, and represent the inhibitory effect of antibiotics on the bacterial flora under different humidity conditions of the vth other patient sample through a single variable Z. u2 Based on the trained multinomial regression model, the corresponding sample humidity is used as input to obtain the output, and Z is obtained from the output. u2 and Z u1 The proportional relationship between them will be Z u2 By performing Z u1 express;
[0021] S30, change the value of v, repeat step S20 until all other patient samples in the target patient's population information pass the Z-test. u1 The antibiotic analysis model is established by fitting: y=f(x), where y represents the inhibitory effect of the antibiotic, x represents the dryness or wetness of the sample, and f represents the fitting function.
[0022] Specifically, optimizing the microbial community distribution in the sample based on the antibiotic effect model and the microbial community distribution in the sample also includes the following steps:
[0023] The DNA testing information and moisture content of the target patient sample are obtained. The moisture content is then input into the antibiotic analysis model to obtain the inhibitory effect of the antibiotic. The inhibitory effect of the antibiotic, the DNA testing information, the time of antibiotic use, and the dosage of antibiotic use are then input into the antibiotic effect model to obtain the optimized bacterial community distribution in the sample, which is provided to medical staff as a reference.
[0024] Compared with the prior art, the beneficial effects of the present invention are: analyzing the influence of intestinal transit time on antibiotic absorption and the influence of antibiotics on the gut microbiota, establishing an antibiotic influence model, optimizing DNA detection information, and reducing the interference of using antimicrobial agents on DNA detection after gastrectomy without RNA detection, thereby improving the reference value and usability of the data. Attached Figure Description
[0025] Figure 1 This is a schematic diagram of the structure of the gastrointestinal function recovery monitoring system after gastrectomy according to the present invention;
[0026] Figure 2 This is a flowchart of the method for monitoring the recovery of gastrointestinal function after gastrectomy according to the present invention. Detailed Implementation
[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0028] Example: Figure 1 As shown, this invention provides a monitoring system for gastrointestinal function recovery after gastrectomy, including a detection module, a data storage module, a sample analysis module, and a status analysis module. The detection module is used to perform DNA testing on samples from target patients to obtain the microbial community distribution. The data storage module is used to store sample testing data from historical patients. The sample analysis module is used to optimize the DNA testing data of the target patient's samples and provide the optimized microbial community distribution in the samples to medical staff for reference. The status analysis module is used to analyze the moisture content of the samples to determine the inhibitory effect of antibiotics.
[0029] The sample analysis module also includes a population analysis unit and a first model training unit. The population analysis unit acquires the preoperative microbial data and gastrectomy characteristics of the target patient, obtains the baseline status of the target patient based on the preoperative microbial data, and obtains the population information of the target patient based on the baseline status and gastrectomy characteristics. The first model training unit takes DNA detection information, antibiotic usage time, antibiotic dosage, and antibiotic inhibitory effect as inputs and RNA detection information as outputs to train an antibiotic influence model. The target patient's data is input into the antibiotic influence model to optimize the target patient's DNA detection information. The state analysis module also includes a transformation unit and a fitting unit: the transformation unit is used to unify the antibiotic inhibitory effect at different scales; the fitting unit is used to fit the sample dryness and humidity and the antibiotic inhibitory effect to establish an antibiotic analysis model. The detection module also includes a moisture detection unit, a DNA detection unit, and a timing unit; the moisture detection unit is used to acquire the water content information of the target patient's sample; the DNA detection unit is used to perform DNA detection on the target patient's sample; the timing unit is used to acquire the time when the target patient used antibiotics.
[0030] In another embodiment of the present invention, the present invention provides a method for monitoring the recovery of gastrointestinal function after gastrectomy, comprising the following steps:
[0031] Obtain postoperative samples from the target patient after gastrectomy, perform DNA testing on the postoperative samples, and obtain the bacterial community distribution in the samples;
[0032] Obtain preoperative microbial data and gastrectomy characteristics of target patients; obtain the baseline status of target patients based on preoperative microbial data; and obtain population information of target patients based on baseline status and gastrectomy characteristics.
[0033] Based on the population information of the target patients, an antibiotic impact model is trained; the bacterial community distribution in the sample is optimized according to the antibiotic impact model and the bacterial community distribution in the sample, and the optimized bacterial community distribution in the sample is provided to medical staff as a reference.
[0034] Specifically, obtaining population information of the target patients based on their baseline status and gastrectomy surgical characteristics also includes the following steps:
[0035] S1: Obtain the types of microbiota and the proportion of each microbiota from the baseline status of the target patient, obtain the historical baseline status of other patients, and calculate the deviation F between the target patient and the i-th other patient based on the baseline status. i F i =D i +Σ j P j ×C ijIn the formula, C ij P represents the characteristics of bacterial community differences. j The weights represent the differential characteristics of the microbial community; j is a positive integer in [1, m], where m represents the total number of microbial species in the baseline state of the target patient and the i-th other patient; D i P represents the Euclidean distance between the percentage of the same bacterial species in the baseline states of the target patient and the i-th other patient; if the j-th bacterial species are different in the baseline states of the target patient and the i-th other patient, then P j If not zero, then P j The value is zero; the j-th bacterial species being different means that only one of the two baseline states of the target patient and the i-th other patient contains the j-th bacterial species; based on the bias, the baseline states of the target patient and other patients are classified unsupervised to determine the taxonomic cluster to which the baseline state of the target patient belongs, and the information of other patients in the taxonomic cluster is extracted to form a set U1;
[0036] S2, obtain the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, calculate the Euclidean distance between the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, perform unsupervised classification on the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients based on the Euclidean distance, determine the classification cluster to which the gastrectomy surgical features of the target patient belong, extract the information of other patients in the classification cluster to form a set U2, obtain the intersection of U1 and U2, and obtain the group information U1∩U2 of the target patient.
[0037] For example, the baseline status of the target patient before surgery includes bacterial flora types Q1, Q2, Q3, ..., Q... 200 Another patient's baseline preoperative bacterial flora includes species Q1, Q2, Q3, ..., Q. 150 W1, ..., W 100 Then Q1, Q2, Q3, ..., Q 150 For the same bacterial species, W1, ..., W 100 and Q 151 ..., W 200 For different bacterial species, based on the same bacterial species Q1, Q2, Q3, ..., Q... 150 The proportion is used to calculate the Euclidean distance between the baseline status of the target patient before surgery and the baseline status of another patient before surgery.
[0038] Then, based on W1, ..., W 100 and Q 151 ..., W 200The weights are determined by the proportion of the microbiota. Since this part of the microbiota only appears in the target patient or other patients, the smaller the proportion of this part of the microbiota, the higher the similarity, and the smaller the deviation between the target patient and other patients. Therefore, the weight of the microbiota difference feature is positively correlated with the proportion. In order to eliminate the influence of scale, the obtained Euclidean distance is averaged with 150. Based on the average, W1, ..., W2 are calculated according to the proportion. 100 and Q 151 ..., W 200 The weights; Q1, Q2, Q3, ..., Q 150 The weight is 0; the deviation is obtained based on the weight and the Euclidean distance.
[0039] Features of gastrectomy surgery include, but are not limited to, the cutting method and the degree of cutting. Cutting methods include distal gastrectomy, proximal gastrectomy, and total gastrectomy. The cutting method feature is quantified by assignment. The degree of cutting refers to the portion of the stomach that is removed. For example, if half of the stomach is removed, the degree of cutting feature is 1 / 2. Optionally, the DBSCAN algorithm is used for unsupervised classification.
[0040] Specifically, in step S1, the weights of the differential characteristics of the microbial community are determined through the following steps:
[0041] Obtain the kth distinct bacterial species from the baseline status of the target patient and the i-th other patient, and denote it as Q. k Get Q k The proportion R k According to R k The weight P of the differential feature of the k-th microbial community is obtained. k , In the formula, n represents the number of bacterial species that are the same in the baseline state of the target patient and the i-th other patient, and R base R represents a reference value indicating the percentage of different bacterial species. base Set to Q k The average percentage.
[0042] Optionally, for bacterial species Q 150 In other words, it is possible to obtain the pre-gastrectomy gut microbiota types Q of other patients. 150 The proportion of bacterial species is used to determine the types of bacteria, Q. 150 The average proportion of the bacterial community type Q 150 R base .
[0043] Specifically, training the antibiotic impact model based on the target patient population information also includes the following steps:
[0044] Historical testing information of samples is obtained from the target patient population information. This testing information includes DNA testing information, RNA testing information, antibiotic usage time, antibiotic dosage, and sample moisture content. An antibiotic analysis model is established based on the sample moisture content. The historical testing information of the samples is then processed using the antibiotic analysis model to obtain the inhibitory effect of the antibiotics.
[0045] The historical detection information of the processed samples is divided into a training set and a test set. The DNA detection information, antibiotic usage time, antibiotic dosage and antibiotic inhibitory effect in the training set are used as inputs, and the RNA detection information is used as outputs to train the antibiotic impact model. The antibiotic impact model is then validated using data from the test set. The antibiotic impact model is obtained after validation.
[0046] Optionally, the sample is selected as the patient's stool. Because post-gastrectomy patients often experience irregular bowel movements (such as diarrhea or constipation), this can alter the duration of antibiotic action in the intestines, thus changing the antibiotic's effect on bacteria. Diarrhea shortens intestinal transit time, resulting in insufficient antibiotic absorption; constipation prolongs intestinal transit time, causing the intestinal mucosa to secrete more mucus, diluting the antibiotic and affecting its effectiveness. Whether constipation, diarrhea, or other conditions, the dryness or wetness of the stool affects intestinal transit time and antibiotic concentration, thereby influencing antibiotic efficacy. Therefore, an antibiotic analysis model is established based on stool dryness / wetness to eliminate the influence of antibiotic absorption status on the test results. The sample's detection information is processed, and the processed information is used to train the antibiotic effect model. The sample's dryness / wetness can be characterized by its water content.
[0047] DNA test results contain interference from dead bacteria, while RNA test results are less affected by dead bacteria. Therefore, by training an antibiotic impact model using historical RNA and DNA test results, and combining this model with the target patient's DNA test results, DNA test results can be optimized without the need for RNA testing, providing a reference for healthcare professionals. Optionally, the DNA and RNA test information can be the percentage of bacterial species, and the model can be a machine learning model, including but not limited to perceptrons, neural networks, and support vector machines.
[0048] Specifically, the antibiotic analysis model is trained through the following steps:
[0049] S10, from the historical detection information of the i-th other patient sample in the target patient population information, obtain the inhibitory effect Z of antibiotics on the bacterial flora under the same antibiotic usage time and dosage in the u1-th sample under different humidity conditions. u1To obtain the inhibitory effect of antibiotics on bacterial flora under different sample humidity conditions and Z-strain. u1 The proportional relationship between them, and the inhibitory effect of antibiotics on bacterial flora under different sample dryness and humidity conditions, were all expressed by Z-axis. u1 To represent; give Z u1 Set a baseline value, establish a multinomial regression model, with the sample dryness and humidity as the input and the antibiotics inhibiting the bacterial community under different sample dryness and humidity conditions as the output.
[0050] S20, Obtain historical testing information of the vth other patient sample from the target patient population information, and represent the inhibitory effect of antibiotics on the bacterial flora under different humidity conditions of the vth other patient sample through a single variable Z. u2 Based on the trained multinomial regression model, the corresponding sample humidity is used as input to obtain the output, and Z is obtained from the output. u2 and Z u1 The proportional relationship between them will be Z u2 By performing Z u1 express;
[0051] S30, change the value of v, repeat step S20 until all other patient samples in the target patient's population information pass the Z-test. u1 The antibiotic analysis model is established by fitting: y=f(x), where y represents the inhibitory effect of the antibiotic, x represents the dryness or wetness of the sample, and f represents the fitting function.
[0052] For patients, antibiotics are taken at fixed times for 3-7 days post-surgery to reduce the probability of infection. Since the antibiotics taken by patients are generally consistent within this short period, meeting the requirement of identical dosage, patient samples with similar antibiotic administration times are selected for analysis. For the i-th other patient, the inhibition effect is z1 when the sample dryness is x1, and z2 when the sample dryness is x2. Since DNA detection is affected by dead bacteria, while RNA is minimally affected, the inhibition effect can be calculated by determining the proportion of the first bacterial group based on the DNA detection results, the proportion of the second bacterial group based on the RNA detection results, and then using the proportions of the second and first bacterial groups to obtain the inhibition effect. Based on the ratio of z1 to z2, z1 is expressed using z2. This step is repeated to express the inhibition effect under different sample dryness conditions using z1.
[0053] For the vth other patient, the inhibitory effect can be represented by a single variable in the same way; at this point, it is necessary to establish the variable relationship between the vth other patient and the ith other patient, since Z u2 Corresponding to the dryness / humidity of a sample, inputting this sample dryness / humidity into a multinomial regression model yields an output that can correlate with Z.u1 Correspondingly, realize Z u2 and Z u1 The unification; the fitting function can take the form f(x) = a0 + a1·x + a2·x 2 +a3·x 3 +…, a1, a2, a3… are coefficients, a0 is the bias. The coefficients and bias can be obtained by least squares method based on the input x and the output y.
[0054] Specifically, optimizing the microbial community distribution in the sample based on the antibiotic effect model and the microbial community distribution in the sample also includes the following steps:
[0055] The DNA testing information and moisture content of the target patient sample are obtained. The moisture content is then input into the antibiotic analysis model to obtain the inhibitory effect of the antibiotic. The inhibitory effect of the antibiotic, the DNA testing information, the time of antibiotic use, and the dosage of antibiotic use are then input into the antibiotic effect model to obtain the optimized bacterial community distribution in the sample, which is provided to medical staff as a reference.
[0056] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. A method for monitoring the recovery of gastrointestinal function after gastrectomy, characterized in that, Includes the following steps: Obtain postoperative samples from the target patient after gastrectomy, perform DNA testing on the postoperative samples, and obtain the bacterial community distribution in the samples; Obtain preoperative microbial data and gastrectomy characteristics of target patients; obtain the baseline status of target patients based on preoperative microbial data; and obtain population information of target patients based on baseline status and gastrectomy characteristics. Based on the population information of the target patients, an antibiotic impact model is trained; the bacterial community distribution in the sample is optimized according to the antibiotic impact model and the bacterial community distribution in the sample, and the optimized bacterial community distribution in the sample is provided to medical staff as a reference.
2. The method for monitoring gastrointestinal function recovery after gastrectomy according to claim 1, characterized in that, The process of obtaining population information of target patients based on their baseline status and gastrectomy surgical characteristics also includes the following steps: S1: Obtain the types of microbiota and the proportion of each microbiota from the baseline status of the target patient, obtain the historical baseline status of other patients, and calculate the deviation F between the target patient and the i-th other patient based on the baseline status. i F i =D i +Σ j P j ×C ij In the formula, C ij P represents the characteristics of bacterial community differences. j The weights represent the differential characteristics of the microbial community; j is a positive integer in [1, m], where m represents the total number of microbial species in the baseline state of the target patient and the i-th other patient; D i P represents the Euclidean distance between the percentage of the same bacterial species in the baseline states of the target patient and the i-th other patient; if the j-th bacterial species are different in the baseline states of the target patient and the i-th other patient, then P j If not zero, then P j The value is zero; the j-th bacterial species being different means that only one of the two baseline states of the target patient and the i-th other patient contains the j-th bacterial species; based on the bias, the baseline states of the target patient and other patients are classified unsupervised to determine the taxonomic cluster to which the baseline state of the target patient belongs, and the information of other patients in the taxonomic cluster is extracted to form a set U1; S2, obtain the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, calculate the Euclidean distance between the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients, perform unsupervised classification on the gastrectomy surgical features of the target patient and the gastrectomy surgical features of other patients based on the Euclidean distance, determine the classification cluster to which the gastrectomy surgical features of the target patient belong, extract the information of other patients in the classification cluster to form a set U2, obtain the intersection of U1 and U2, and obtain the group information U1∩U2 of the target patient.
3. The method for monitoring gastrointestinal function recovery after gastrectomy according to claim 2, characterized in that, In step S1, the weights of the differential characteristics of the microbial community are determined through the following steps: Obtain the kth distinct bacterial species from the baseline status of the target patient and the i-th other patient, and denote it as Q. k Get Q k The proportion R k According to R k The weight P of the differential feature of the k-th microbial community is obtained. k , In the formula, n represents the number of bacterial species that are the same in the baseline state of the target patient and the i-th other patient, and R base R represents a reference value indicating the percentage of different bacterial species. base Set to Q k The average percentage.
4. The method for monitoring gastrointestinal function recovery after gastrectomy according to claim 3, characterized in that, The training of the antibiotic impact model based on the target patient population information also includes the following steps: Historical testing information of samples is obtained from the target patient population information. This testing information includes DNA testing information, RNA testing information, antibiotic usage time, antibiotic dosage, and sample moisture content. An antibiotic analysis model is established based on the sample moisture content. The historical testing information of the samples is then processed using the antibiotic analysis model to obtain the inhibitory effect of the antibiotics. The historical detection information of the processed samples is divided into a training set and a test set. The DNA detection information, antibiotic usage time, antibiotic dosage and antibiotic inhibitory effect in the training set are used as inputs, and the RNA detection information is used as outputs to train the antibiotic impact model. The antibiotic impact model is then validated using data from the test set. The antibiotic impact model is obtained after validation.
5. The method for monitoring the recovery of gastrointestinal function after gastrectomy according to claim 4, characterized in that, The antibiotic analysis model was trained through the following steps: S10, from the historical detection information of the i-th other patient sample in the target patient population information, obtain the inhibitory effect Z of antibiotics on the bacterial flora under the same antibiotic usage time and dosage in the u1-th sample under different humidity conditions. u1 To obtain the inhibitory effect of antibiotics on bacterial flora under different sample humidity conditions and Z-strain. u1 The proportional relationship between them, and the inhibitory effect of antibiotics on bacterial flora under different sample dryness and humidity conditions, were all expressed by Z-axis. u1 To represent; give Z u1 Set a baseline value, establish a multinomial regression model, with the sample dryness and humidity as the input and the antibiotics inhibiting the bacterial community under different sample dryness and humidity conditions as the output. S20, Obtain historical testing information of the vth other patient sample from the target patient population information, and represent the inhibitory effect of antibiotics on the bacterial flora under different humidity conditions of the vth other patient sample through a single variable Z. u2 Based on the trained multinomial regression model, the corresponding sample humidity is used as input to obtain the output, and Z is obtained from the output. u2 and Z u1 The proportional relationship between them will be Z u2 By performing Z u1 express; S30, change the value of v, repeat step S20 until all other patient samples in the target patient's population information pass the Z-test. u1 The antibiotic analysis model is established by fitting: y=f(x), where y represents the inhibitory effect of the antibiotic, x represents the dryness or wetness of the sample, and f represents the fitting function.
6. The method for monitoring gastrointestinal function recovery after gastrectomy according to claim 5, characterized in that, The optimization of the microbial community distribution in the sample based on the antibiotic effect model and the microbial community distribution in the sample also includes the following steps: The DNA testing information and moisture content of the target patient sample are obtained. The moisture content is then input into the antibiotic analysis model to obtain the inhibitory effect of the antibiotic. The inhibitory effect of the antibiotic, the DNA testing information, the time of antibiotic use, and the dosage of antibiotic use are then input into the antibiotic effect model to obtain the optimized bacterial community distribution in the sample, which is provided to medical staff as a reference.
7. A system for monitoring the recovery of gastrointestinal function after gastrectomy, characterized in that, It includes a detection module, a data storage module, a sample analysis module, and a status analysis module. The detection module is used to perform DNA testing on samples from target patients to obtain the bacterial community distribution. The data storage module is used to store sample testing data from historical patients. The sample analysis module is used to optimize the DNA testing data of the target patient samples and provide the optimized bacterial community distribution in the samples to medical staff for reference. The status analysis module is used to analyze the moisture content of the samples to determine the inhibitory effect of antibiotics.
8. The monitoring system for gastrointestinal function recovery after gastrectomy according to claim 7, characterized in that, The sample analysis module further includes a population analysis unit and a first model training unit; the population analysis unit acquires the preoperative microbial data and gastrectomy characteristics of the target patient, obtains the baseline status of the target patient based on the preoperative microbial data, and obtains the population information of the target patient based on the baseline status and gastrectomy characteristics. The first model training unit takes DNA detection information, antibiotic usage time, antibiotic dosage, and antibiotic inhibitory effect as inputs and RNA detection information as outputs to train the antibiotic impact model; the target patient's data is input into the antibiotic impact model to optimize the target patient's DNA detection information.
9. The monitoring system for gastrointestinal function recovery after gastrectomy according to claim 8, characterized in that, The state analysis module also includes a transformation unit and a fitting unit: the transformation unit is used to unify the antibiotic inhibition effect at different scales; the fitting unit is used to fit the sample dryness and humidity and the antibiotic inhibition effect to establish an antibiotic analysis model.
10. The monitoring system for gastrointestinal function recovery after gastrectomy according to claim 9, characterized in that, The detection module further includes a moisture detection unit, a DNA detection unit, and a timing unit; the moisture detection unit is used to obtain the water content information of the target patient sample; the DNA detection unit is used to perform DNA detection on the target patient sample; and the timing unit is used to obtain the time when the target patient used antibiotics.