Anticoagulation individualized medication guidance analysis system based on genetic polymorphism

Abstract

This invention discloses a personalized anticoagulation medication guidance and analysis system based on gene polymorphism, belonging to the field of gene testing and clinical medication guidance technology. It includes: a data acquisition module that acquires raw gene testing data related to the patient's anticoagulation medication and identifies gene polymorphic sites, generating multi-site genotype combination data containing drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites; an annotation mapping module that queries a gene-phenotype mapping knowledge base to determine the metabolic phenotype category and target sensitivity category corresponding to each gene site, generating a site functional annotation set; and a metabolic grading module that uses the site functional annotation set to perform comprehensive grading of anticoagulation drug metabolism rates, obtaining a drug metabolism rate grading index. This system can output initial doses and dynamic dosing plans that are more closely aligned with individual genetic and clinical characteristics.

Description

Technical Field

[0001] This invention relates to the field of gene detection and clinical medication guidance technology, specifically to an anticoagulation personalized medication guidance and analysis system based on gene polymorphism. Background Technology

[0002] The dosage of anticoagulants varies greatly among individuals in clinical application, especially for coumarin-based drugs such as warfarin. Overdose can easily lead to bleeding, while underdose can induce thrombosis, necessitating individualized medication guidance. Current gene-guided anticoagulation therapy typically only genotypes the CYP2C9 metabolic enzyme gene and the VKORC1 target gene, calculating the initial dose based on a simple combination of metabolic rate and target sensitivity. This approach ignores the polymorphic influence of transporter genes on drug absorption, distribution, and efflux, and fails to quantitatively integrate target sensitivity categories with metabolic phenotypes, thus failing to form a multi-gene functional gradation that comprehensively reflects drug exposure and effect intensity. Patients with the same metabolic type may exhibit drastically different blood drug concentration curves due to differences in transporter activity; current protocols cannot capture this information, leaving the genetic basis for dose estimation at the level of isolated single or double genes.

[0003] Regarding dosage adjustment, conventional practices employ static population pharmacokinetic formulas, calculating fixed doses based on genotype, age, and weight, but lack a standardized approach to addressing concomitant medication interference. Patients often take multiple medications concurrently during anticoagulation therapy, some of which induce or inhibit metabolic enzymes such as CYP2C9, continuously altering anticoagulant clearance rates and causing shifts in steady-state dosage requirements. Existing systems do not incorporate the enzyme induction and inhibition coefficients of concomitant drugs into the dosage correction system, nor do they combine dynamic changes in liver and kidney function to output dosage adjustment cycles. This prevents the dosing regimen from matching individual metabolic trends at the beginning of treatment, and subsequent adjustments lack clear timeframes. Furthermore, the generated dosing plan does not perform conflict checks on contraindications, easily leading to overlaps between recommended dosing windows and contraindications, affecting adherence. Therefore, it is necessary to address how to generate a unified metabolic rate grading index based on multi-locus genotype combinations, and how to use this grading index in conjunction with physical status indicators and a list of concomitant medications to calculate the initial dose and adjustment cycle with time-series conflict correction. Summary of the Invention

[0004] This invention provides a personalized anticoagulation medication guidance and analysis system based on gene polymorphism, which integrates the comprehensive influence of drug-metabolizing enzyme genes, target genes, and transporter genes on drug response, and based on this, combines physical condition and combined medication information to achieve accurate calculation of initial dose and adjustment cycle.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides an anticoagulant personalized medication guidance and analysis system based on gene polymorphism. By integrating polymorphic information on gene loci related to drug metabolic enzymes, targets, and transporters, it achieves accurate calculation of anticoagulant drug dosage and formulation of dosing regimens. The system includes a data acquisition module, an annotation and mapping module, a metabolic classification module, a dosage calculation module, and a regimen output module.

[0006] The data acquisition module is used to acquire raw gene detection data related to the patient's anticoagulant medication, and to perform gene polymorphism site identification processing on the raw gene detection data to generate multi-site genotype combination data including drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites. Preferably, the data acquisition module obtains gene sequence fragments by performing base sequence alignment on the raw gene detection data, and extracts target sequence fragments from the gene sequence fragments according to the preset reference sequence positions corresponding to the VKORC1 gene site, CYP2C9 gene site, and CYP4F2 gene site. Then, it performs base variation identification on each target sequence fragment to determine the allele type at each gene site, thereby obtaining the multi-site genotype combination data. Further, when extracting target sequence fragments, the system acquires the start and end coordinates of each gene site on the reference genome, extracts the corresponding original region sequence from the gene sequence fragment, and performs flanking sequence pruning on the original region sequence to remove redundant bases in the non-coding regions at both ends, thereby improving the accuracy of subsequent variation identification.

[0007] The annotation mapping module, based on the multi-locus genotype combination data, queries a pre-defined gene-phenotype mapping knowledge base to determine the metabolic phenotype category and target sensitivity category corresponding to each gene locus, generating a site functional annotation set. As a technical solution of this invention, the annotation mapping module uses each gene locus as a query index to search for phenotypic record entries matching each allele type in the gene-phenotype mapping knowledge base, extracting the metabolic phenotype category and target sensitivity category from them. The metabolic phenotype category may include ultra-rapid metabolizers, extensive metabolizers, intermediate metabolizers, and slow metabolizers; the target sensitivity category may include highly sensitive, normally sensitive, and lowly sensitive. The metabolic phenotype category and target sensitivity category of each gene locus are associated and combined to form the site functional annotation set. Thus, the system can comprehensively integrate dual information on drug metabolism capacity and drug efficacy target sensitivity, providing multi-dimensional functional basis for subsequent dose grading.

[0008] The metabolic grading module utilizes the site functional annotation set to perform comprehensive grading of the anticoagulant metabolism rate in patients, obtaining a corresponding drug metabolism rate grading index. Specifically, the metabolic grading module extracts the metabolic phenotype category and target sensitivity category corresponding to all gene loci from the site functional annotation set. According to preset metabolic phenotype scoring rules, a first functional score value is assigned to the metabolic phenotype category of each gene locus, and according to preset target sensitivity scoring rules, a second functional score value is assigned to the target sensitivity category of each gene locus. The first and second functional score values ​​corresponding to all gene loci are summed to obtain a total functional score. Then, the total functional score is graded according to a preset grading threshold range to obtain the drug metabolism rate grading index. This grading index can quantitatively reflect the overall metabolic clearance trend and target response characteristics of anticoagulants in the patient's body, overcoming the one-sidedness of evaluation based on a single gene locus.

[0009] The dosage estimation module acquires the patient's current physical condition index data and a list of concomitant medications, and inputs the drug metabolism rate classification index, physical condition index data, and concomitant medication list into the dosage estimation rule engine, outputting the initial recommended anticoagulant drug dosage and dosage adjustment cycle. Preferably, the dosage estimation module first performs validity verification processing on the physical condition index data, filtering out invalid index values ​​that exceed the preset physiological range to obtain valid physical condition index data; then, it uses the drug metabolism rate classification index and valid physical condition index data as initial dosage estimation factors, matching the corresponding baseline dosage range in the dosage estimation rule engine; simultaneously, using the name of each concomitant drug in the list of concomitant medications, it queries the drug interaction rule table in the dosage estimation rule engine to obtain the dosage correction coefficient corresponding to each concomitant drug; finally, based on the baseline dosage range and all dosage correction coefficients, it calculates the initial recommended anticoagulant drug dosage and dosage adjustment cycle. When matching the baseline dose range, the system extracts age, body weight, and liver function indicators from effective physical condition data, combines them with the drug metabolism rate grading index to form a multidimensional predictive factor vector, and retrieves the standard dose range corresponding to the standard factor combination closest to this vector from the dose rule table, which serves as the baseline dose range. When obtaining the dose correction coefficient, the system performs string matching between the combined drug name and the interfering drug name field in the drug interaction rule table. If a match is successful, the associated enzyme induction coefficient or enzyme inhibition coefficient is extracted; if a match fails, the dose correction coefficient corresponding to the combined drug is set to a neutral value, thus quantifying the impact of drug interactions into the dose calculation process. These techniques comprehensively consider multiple factors such as genotype, physiological state, and concomitant medications, significantly improving the rationality and safety of the initial dosage.

[0010] The protocol output module generates an individualized medication guidance plan for the patient based on the initial recommended anticoagulant dosage and dosage adjustment cycle. In one embodiment of the invention, the protocol output module decomposes the initial recommended anticoagulant dosage into single-dose dose values ​​and dosing interval values ​​according to the dosage adjustment cycle, and generates a time-series dosing schedule including the start and end times of dosing based on these values. Subsequently, the time-series dosing schedule is checked against the patient's contraindication labels. When the dosing time points in the schedule overlap with the contraindication time periods in the contraindication labels, the dosing time points are corrected for offset, resulting in a corrected time-series dosing schedule, which is then output as the individualized medication guidance plan. When breaking down single-dose dosage and dosing interval values, the system retrieves the total daily dose value included in the initial recommended anticoagulant dosage, the total number of adjustment days in the dose adjustment cycle, and the patient's medication adherence time window parameter, which includes the acceptable range of daily dosing frequencies. The single-dose dosage value is calculated based on the total daily dose value and the acceptable daily dosing frequency range, and the dosing interval value is determined based on the total adjustment days and the acceptable daily dosing frequency range. Through this method, the system can improve patient medication adherence and safety while ensuring efficacy, avoiding risks caused by inappropriate dosing timing.

[0011] This invention organically integrates gene polymorphism detection data, physical condition indicators, and combined medication information. Through knowledge base mapping, hierarchical scoring, and rule engine reasoning, it realizes individualized dosage recommendations and dynamic dosing regimens for anticoagulant drugs. It features high accuracy and strong clinical applicability, and helps to promote the transformation of anticoagulation therapy from empirical dosing to precision medicine.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0013] To address the challenge of quantifying the polymorphisms of drug-metabolizing enzyme genes, target genes, and transporter genes at the functional level, an annotation mapping module, based on a gene-phenotype mapping knowledge base, assigns a metabolic phenotype category and a target sensitivity category to each gene locus, generating a site functional annotation set. The target sensitivity category includes classifications of high sensitivity, normal sensitivity, and low sensitivity, while transporter gene loci are included in the overall assessment based on their impact on drug transmembrane transport efficiency, expressed as a functional score. The metabolic grading module extracts the functional annotations from all loci and converts the metabolic phenotype category and target sensitivity category into a first functional score and a second functional score, respectively, according to preset scoring rules. These scores are then summed, and a drug metabolic rate grading index is determined based on a grading threshold range. This grading index no longer reflects the activity level of a single metabolic enzyme but rather integrates three dimensions: drug metabolism rate in vivo, target tissue response sensitivity, and cellular transport flux. Compared to the traditional method of classifying metabolic types solely based on CYP2C9 and VKORC1 genotypes, the introduction of multi-gene functional scoring allows for a finer granular characterization of an individual's anticoagulant handling capacity. Different combinations of transporter activity and target sensitivity can generate sufficiently distinct gradient values ​​on the grading index, directly translating into differences in the baseline dose range. This avoids the problem of giving the same dose recommendation due to differences in transporter and target genes under the same metabolic phenotype.

[0014] To effectively integrate genetic grading indices with dynamic clinical factors during treatment and output executable dosing regimens, the dosage estimation module, after obtaining the drug metabolism rate grading index, first verifies the validity of body status indicator data, filtering out outliers outside the physiological range. It then extracts age, body weight, and liver function indicators, combining them with the grading index to form a multidimensional estimation factor vector, which is then matched with the corresponding baseline dose range in the rule engine. Simultaneously, it uses the drug names in the concomitant medication list to query the drug interaction rule table, obtaining the enzyme induction coefficient or enzyme inhibition coefficient of each concomitant drug as a dose correction coefficient. When the concomitant drug is an enzyme inducer, the correction coefficient increases the recommended dose; when it is an enzyme inhibitor, it decreases it; unmatched drugs use a neutral value to avoid interference from irrelevant drugs in the calculation. This comprehensive rule matching and coefficient-based correction replaces the practice of clinicians estimating drug interactions based on experience, systematically incorporating the interference effects of concomitant drugs into the initial dose generation process. In addition to outputting the initial recommended dose, the dosage estimation rule engine also provides a dose adjustment cycle. This cycle is determined based on indicators such as the metabolic rate grading index and liver function, providing a clear time reference for subsequent dose evaluation and adjustment, rather than blindly waiting for coagulation indicator monitoring results or patient feedback. The protocol output module decomposes the initial dose into single-dose dose values ​​and dosing interval values, introduces patient medication adherence time window constraints, generates a time-series dosing plan, and uses medication contraindication tags to perform conflict verification on dosing time points. It performs offset correction on time points that overlap with contraindication time periods, controlling the offset direction to avoid contraindication intervals. The resulting individualized medication guidance plan simultaneously matches genetic background, real-time physical condition, concomitant medication interference, and individual lifestyle contraindications in terms of dose values ​​and dosing sequence, reducing dosage inappropriateness and medication time conflicts caused by single-factor decision-making, and achieving a higher degree of safe medication prescription generation without manual intervention. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a schematic diagram of the structure of an anticoagulation personalized medication guidance and analysis system based on gene polymorphism;

[0017] Figure 2 This is a flowchart illustrating the method for generating individualized medication guidance protocols for anticoagulants.

[0018] Figure 3 This is a sequencing depth distribution map of the target sequences at the VKORC1, CYP2C9, and CYP4F2 gene loci;

[0019] Figure 4 This is a schematic diagram showing the total distribution and grading thresholds of the functional score for the metabolic rate of anticoagulants;

[0020] Figure 5 This is a distribution chart of recommended initial anticoagulant doses corresponding to different drug metabolism rate classifications;

[0021] Figure 6 This is a distribution chart of single-dose dosage and dosing interval in individualized anticoagulant regimens. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0023] See Figure 1 This invention provides a personalized anticoagulation medication guidance and analysis system based on gene polymorphism, comprising a data acquisition module, an annotation and mapping module, a metabolic grading module, a dose estimation module, and a protocol output module. The data acquisition module acquires raw gene detection data related to the patient's anticoagulation medication and performs gene polymorphism site identification processing on the raw gene detection data to generate multi-site genotype combination data including drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites. The annotation and mapping module, based on the multi-site genotype combination data, queries a preset gene-phenotype mapping knowledge base to determine the metabolic phenotype category and target sensitivity category corresponding to each gene site, generating a site functional annotation set. The metabolic grading module uses the site functional annotation set to perform comprehensive grading of the patient's anticoagulation drug metabolism rate, obtaining the patient's corresponding drug metabolism rate grading index. The dose estimation module acquires the patient's current physical condition indicators and a list of concomitant medications, and inputs the drug metabolism rate grading index, physical condition indicators, and concomitant medication list into the dose estimation rule engine, outputting the patient's initial recommended anticoagulation drug dose and dose adjustment cycle. The protocol output module generates a personalized medication guidance plan for the patient based on the initial recommended dose of anticoagulant and the dose adjustment cycle.

[0024] Example 1:

[0025] In practice, the data acquisition module obtains raw gene testing data related to the patient's anticoagulant medication. This raw gene testing data originates from high-throughput sequencing of the patient's blood samples, and the data format is FASTQ sequencing read files. The data acquisition module performs gene polymorphism site identification processing on the raw gene testing data, generating multi-site genotype combination data containing drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites. The specific implementation process is as follows.

[0026] The data acquisition module performs base sequence alignment on the raw gene detection data to obtain the patient's gene sequence fragments. In some implementations, the base sequence alignment process uses the BWA-MEM algorithm to map FASTQ format sequencing reads onto the human reference genome GRCh38, generating a SAM format alignment file. The data acquisition module then uses samtools software to convert the SAM format alignment file to a BAM format file, and performs sorting and indexing to obtain gene sequence fragments containing the patient's whole genome alignment information.

[0027] Optionally, after obtaining the gene sequence fragment, the data acquisition module extracts target sequence fragments from the gene sequence fragment according to the preset reference sequence positions corresponding to the VKORC1, CYP2C9, and CYP4F2 gene loci. Specifically, the data acquisition module acquires the preset reference sequence position information corresponding to the VKORC1, CYP2C9, and CYP4F2 gene loci. This reference sequence position information includes the start and end coordinates of each gene locus on the reference genome. The reference sequence position information for the VKORC1 gene locus is the interval from 31100000 to 31101000 on chromosome 16; the reference sequence position information for the CYP2C9 gene locus is the interval from 96700000 to 96702000 on chromosome 10; and the reference sequence position information for the CYP4F2 gene locus is the interval from 15900000 to 15901000 on chromosome 19. The reference sequence position information is stored in a BED format file, and the data acquisition module obtains the start and end coordinates by reading the BED format file.

[0028] The data acquisition module extracts the original region sequence corresponding to each gene locus from the gene sequence fragment based on the start and end coordinates. Specifically, the data acquisition module uses the view command of the samtools software to specify the chromosome number and coordinate range in the BAM format file, outputs the read alignment information of the corresponding region, and forms the original region sequence corresponding to each gene locus. The original region sequence is a base sequence containing the core gene locus and its upstream and downstream flanking sequences of a certain length.

[0029] After obtaining the original region sequences, the data acquisition module performs flanking sequence trimming on each original region sequence, removing redundant bases from the non-coding regions at both ends of the original region sequence to obtain the target sequence fragment. The flanking sequence trimming is based on the exon boundaries and coding region start and end positions marked in the gene annotation file (GTF). The data acquisition module uses the `bedtoolsintersect` tool to find the intersection of the original region sequence and the coding region coordinates in the gene annotation file, trimming the non-coding bases outside the coding region, retaining the core coding sequence and alternative splicing regions, forming the target sequence fragment for each gene locus. The length of the target sequence fragment is determined based on the actual coding region length of the corresponding gene locus. For example, the target sequence fragment length for the VKORC1 gene locus is approximately 512 base pairs, for the CYP2C9 gene locus it is approximately 1400 base pairs, and for the CYP4F2 gene locus it is approximately 1200 base pairs.

[0030] After obtaining the target sequence fragment corresponding to each gene locus, the data acquisition module performs base variation identification processing on each target sequence fragment to determine the allele type at each gene locus. The base variation identification processing uses the GATKHaplotypeCaller tool to perform local re-alignment and detect single nucleotide variants and small insertions / deletions on the target sequence fragments. The GATKHaplotypeCaller tool outputs a variant detection result file in VCF format, which includes the chromosomal location of the variant site, reference base, substituted base, genotype information, and quality value. The data acquisition module reads the specific allele type based on the genotype information recorded in the VCF format variant detection result file. For example, for the CYP2C9 gene locus, it identifies the genotype of the single nucleotide polymorphism (SNP) site rs1799853 corresponding to the CYP2C9*2 allele as TT, TC, or CC, determining the CYP2C9*2 carrier status; it identifies the genotype of the single nucleotide polymorphism site rs1057910 corresponding to the CYP2C9*3 allele as AA, AC, or CC, determining the CYP2C9*3 carrier status. The module then comprehensively determines the allele type of the CYP2C9 gene locus as a combination of CYP2C9*1 / *2, CYP2C9*1 / *3, etc. For the VKORC1 gene locus, it identifies the genotype of the single nucleotide polymorphism rs9923231 at the promoter region -1639 as GG, GA, or AA, determining the allele type of the VKORC1 gene locus. For the CYP4F2 gene locus, the genotype of the single nucleotide polymorphism site rs2108622 is identified as CC, CT, or TT to determine the allele type of the CYP4F2 gene locus. The data acquisition module combines the allele types of the VKORC1 gene locus, the CYP2C9 gene locus, and the CYP4F2 gene locus to form multi-site genotype combination data containing drug-metabolizing enzyme gene loci, target gene loci, and transporter gene loci. Among them, the VKORC1 gene locus is a target gene locus, the CYP2C9 gene locus is a drug-metabolizing enzyme gene locus, and the CYP4F2 gene locus is a transporter gene locus.

[0031] See Figure 3 The figure shows the sequencing depth distribution of target sequence fragments for the VKORC1, CYP2C9, and CYP4F2 gene loci in the data acquisition module. The horizontal axis represents base position (bp), covering the length of each gene fragment. The VKORC1 target sequence fragment is approximately 512 bp long, the CYP2C9 target sequence fragment is approximately 1400 bp long, and the CYP4F2 target sequence fragment is approximately 1200 bp long. The vertical axis represents sequencing depth (×), with values ​​ranging from 20 to 65.

[0032] The blue solid line represents the sequencing depth of the target sequence fragment of the VKORC1 gene. The curve fluctuates significantly overall, with the depth mostly concentrated in the range of 25× to 60×. The sequencing depth of some base sites is significantly higher than 50×, indicating that the sequencing coverage of this gene fragment is good. The changes in depth reflect the differences in sequencing efficiency in different regions.

[0033] The orange dashed line represents the sequencing depth of the target sequence fragment of the CYP2C9 gene. The sequencing depth distribution is relatively uniform, with the depth of most sites concentrated in the range of 35× to 55×. The overall sequencing depth is higher than that of VKORC1, and the sequencing depth remains stable at base positions after 1000bp, indicating that the high-throughput sequencing quality of this fragment is stable and the coverage is uniform.

[0034] The green dotted line represents the sequencing depth of the target sequence fragment of the CYP4F2 gene. The curve shows a similar fluctuation trend to CYP2C9, with the sequencing depth concentrated in the range of 30× to 50×. The overall curve is relatively smooth, indicating that the sequencing data of this gene fragment is of high quality and sufficient to support subsequent variant identification.

[0035] Example 2:

[0036] In practice, the annotation mapping module receives multi-site genotype combination data output by the data acquisition module. This data includes allele types for the VKORC1, CYP2C9, and CYP4F2 gene loci. Based on this data, the annotation mapping module queries a pre-defined gene-phenotype mapping knowledge base to determine the corresponding metabolic phenotype and target sensitivity categories for the VKORC1, CYP2C9, and CYP4F2 gene loci, generating a site functional annotation set.

[0037] The pre-defined gene-phenotype mapping knowledge base is stored in the system memory as a relational database table. The knowledge base includes fields for gene locus name, allele type, metabolic phenotype category, and target sensitivity category. The gene locus name field has a value set of {VKORC1, CYP2C9, CYP4F2}, and the allele type field stores textual descriptions of various allele types for the corresponding gene locus. The metabolic phenotype category field has a value set of {ultra-rapid metabolizer, extensive metabolizer, intermediate metabolizer, slow metabolizer, not applicable}, and the target sensitivity category field has a value set of {highly sensitive, normally sensitive, lowly sensitive}. Each record in the gene-phenotype mapping knowledge base corresponds to a mapping relationship between a combination of allele types at a gene locus and the relevant phenotype. This mapping relationship is pre-constructed based on the guidelines published by the Clinical Pharmacogenomics Implementation Consortium and published population pharmacokinetic study data.

[0038] In some implementations, the metabolic phenotype category field of the VKORC1 gene locus is recorded as "not applicable". The correspondence between the target sensitivity category field of the VKORC1 gene locus and the allele type of the VKORC1 gene locus is as follows: when the allele type of the VKORC1 gene locus is -1639GG, the target sensitivity category field is recorded as low sensitivity; when the allele type of the VKORC1 gene locus is -1639GA, the target sensitivity category field is recorded as normal sensitivity; when the allele type of the VKORC1 gene locus is -1639AA, the target sensitivity category field is recorded as high sensitivity.

[0039] The correspondence between the metabolic phenotype category field and the allele type of the CYP2C9 gene locus is as follows: when the allele type of the CYP2C9 gene locus is *1 / *1, the metabolic phenotype category field is recorded as extensive metabolizer; when the allele type of the CYP2C9 gene locus is *1 / *2 or *1 / *3, the metabolic phenotype category field is recorded as intermediate metabolizer; when the allele type of the CYP2C9 gene locus is *2 / *2, *2 / *3, or *3 / *3, the metabolic phenotype category field is recorded as slow metabolizer. The target sensitivity category field of the CYP2C9 gene locus is recorded as normal sensitivity in all cases.

[0040] The correspondence between the metabolic phenotype category field and the allele type of the CYP4F2 gene locus is as follows: when the allele type of the CYP4F2 gene locus is CC, the metabolic phenotype category field is recorded as extensive metabolizer; when the allele type of the CYP4F2 gene locus is CT, the metabolic phenotype category field is recorded as intermediate metabolizer; and when the allele type of the CYP4F2 gene locus is TT, the metabolic phenotype category field is recorded as slow metabolizer. The target sensitivity category field of the CYP4F2 gene locus is recorded as normal sensitivity in all cases.

[0041] Optionally, the construction process of the gene-phenotype mapping knowledge base includes data collection, literature mining, and expert review. Evidence of the association between gene loci and phenotypes is extracted from the PharmGKB database, CPIC guidelines, and relevant clinical studies. After manual verification, a mapping relationship table is formed and stored in the system's relational database.

[0042] In its implementation, the annotation mapping module uses each gene locus in the multi-locus genotype combination data as a query index to search for phenotypic records in the gene-phenotype mapping knowledge base that match the allele type of each gene locus. The query process is implemented using a structured query language. The annotation mapping module concatenates the gene locus name and the corresponding allele type into the query conditions. For example, for the VKORC1 gene locus, the query condition is constructed as "SELECT metabolic phenotype category, target sensitivity category FROM gene-phenotype mapping knowledge base WHERE gene locus name='VKORC1' AND allele type='VKORC1 gene locus allele type'", where the allele type of the VKORC1 gene locus is provided by the multi-locus genotype combination data. The annotation mapping module sequentially performs queries on the VKORC1 gene locus, the CYP2C9 gene locus, and the CYP4F2 gene locus, obtaining phenotypic records that uniquely match the allele type of each gene locus.

[0043] It is understandable that when the allele type of a certain gene locus in the multi-locus genotype combination data cannot be matched with a completely consistent phenotypic record in the gene-phenotype mapping knowledge base, the annotation mapping module adopts the nearest allele type matching strategy, selects the phenotypic record corresponding to the known allele type that is closest to the functional effect of the current allele type as a substitute, and generates a matching confidence label.

[0044] The annotation mapping module extracts the metabolic phenotype category corresponding to each gene locus from the queried phenotypic record entries. The extraction process is achieved by reading the metabolic phenotype category field from the phenotypic record entries. The extracted metabolic phenotype category for the VKORC1 gene locus is "Not Applicable," the extracted metabolic phenotype category for the CYP2C9 gene locus is one of extensive metabolizer, intermediate metabolizer, or slow metabolizer, and the extracted metabolic phenotype category for the CYP4F2 gene locus is one of extensive metabolizer, intermediate metabolizer, or slow metabolizer.

[0045] The annotation mapping module extracts the target sensitivity category corresponding to each gene locus from the queried phenotypic record entries. The extraction process is achieved by reading the target sensitivity category field from the phenotypic record entries. The target sensitivity category extracted from the VKORC1 gene locus is one of high sensitivity, normal sensitivity, or low sensitivity; the target sensitivity category extracted from the CYP2C9 gene locus is normal sensitivity; and the target sensitivity category extracted from the CYP4F2 gene locus is normal sensitivity.

[0046] In some implementations, the annotation mapping module associates and combines the metabolic phenotype category and target sensitivity category for each gene locus to generate a locus functional annotation set. The locus functional annotation set is organized using a key-value pair structure, with the gene locus name as the key and a numerical pair containing the metabolic phenotype category and target sensitivity category as the value. A specific record format for the locus functional annotation set is: {"VKORC1": (metabolic phenotype category = not applicable, target sensitivity category = high sensitivity), "CYP2C9": (metabolic phenotype category = intermediate metabolite, target sensitivity category = normal sensitivity), "CYP4F2": (metabolic phenotype category = extensive metabolite, target sensitivity category = normal sensitivity)}. After generation, the locus functional annotation set is passed to the metabolic grading module for subsequent processing.

[0047] Example 3:

[0048] In its implementation, the metabolic grading module receives a site function annotation set output by the annotation mapping module. This set includes the metabolic phenotype and target sensitivity categories for the VKORC1, CYP2C9, and CYP4F2 gene loci. Using this site function annotation set, the metabolic grading module performs a comprehensive grading of the patient's anticoagulant drug metabolism rate, obtaining the corresponding drug metabolism rate grading index.

[0049] The metabolic classification module extracts the metabolic phenotype and target sensitivity categories corresponding to all gene loci from the site functional annotation set. The site functional annotation set is stored in a key-value pair structure. The metabolic classification module iterates through each key-value pair element in the site functional annotation set, reading the metabolic phenotype and target sensitivity categories from the values ​​with the key name "VKORC1", "CYP2C9", and "CYP4F2". After extraction, the metabolic classification module obtains six independent phenotype category values: the metabolic phenotype category of the VKORC1 gene locus, the target sensitivity category of the VKORC1 gene locus, the metabolic phenotype category of the CYP2C9 gene locus, the target sensitivity category of the CYP2C9 gene locus, the metabolic phenotype category of the CYP4F2 gene locus, and the target sensitivity category of the CYP4F2 gene locus.

[0050] The metabolic grading module assigns a first functional score value to the metabolic phenotype category of each gene locus according to preset metabolic phenotype scoring rules. These preset rules are stored in the module's configuration parameters as a mapping table, defining the correspondence between metabolic phenotype categories and first functional score values. Specifically, the mapping table is as follows: for ultra-rapid metabolizers, the first functional score value is recorded as 4; for extensive metabolizers, as 3; for intermediate metabolizers, as 2; for slow metabolizers, as 1; and for inapplicable metabolizers, as 0. The first functional score value is determined based on the gradient of the influence of drug-metabolizing enzyme activity on drug metabolism rate; a higher value indicates stronger metabolic capacity. The metabolic grading module iterates through the metabolic phenotype categories of the VKORC1, CYP2C9, and CYP4F2 gene loci, searching the mapping table for the first functional score corresponding to each category. The VKORC1 gene locus's metabolic phenotype category is always "inapplicable," resulting in a first functional score of 0. The CYP2C9 gene locus's metabolic phenotype category is one of extensive metabolizer, intermediate metabolizer, or slow metabolizer, resulting in a first functional score of 3, 2, or 1, respectively. Similarly, the CYP4F2 gene locus's metabolic phenotype category is also one of extensive metabolizer, intermediate metabolizer, or slow metabolizer, resulting in a first functional score of 3, 2, or 1, respectively.

[0051] The metabolic grading module assigns a second functional score value to the target sensitivity category of each gene locus according to preset target sensitivity scoring rules. These preset rules are stored in the module's configuration parameters as a mapping table, defining the correspondence between target sensitivity categories and second functional score values. Specifically, the mapping table records the second functional score value as follows: 3 for a high-sensitivity target category; 2 for a normal-sensitivity target category; and 1 for a low-sensitivity target category. The second functional score value is determined based on the intensity of the target's drug response; a higher value indicates stronger sensitivity, and the required drug dosage is correspondingly reduced. The metabolic grading module iterates through the target sensitivity categories of the VKORC1, CYP2C9, and CYP4F2 gene loci, searching the mapping table for the corresponding second functional score value for each target sensitivity category. The target sensitivity category of the VKORC1 gene locus is either high-sensitivity, normal-sensitivity, or low-sensitivity, with corresponding secondary function scores of 3, 2, or 1, respectively. The target sensitivity category of the CYP2C9 gene locus is consistently normal-sensitivity, with a corresponding secondary function score of 2. The target sensitivity category of the CYP4F2 gene locus is consistently normal-sensitivity, with a corresponding secondary function score of 2.

[0052] The metabolic grading module sums the first and second functional scores corresponding to all gene loci to obtain a total functional score. This summation covers the first and second functional scores corresponding to the VKORC1, CYP2C9, and CYP4F2 gene loci, totaling six scores. The formula for calculating the total functional score is as follows:

[0053]

[0054] in, This represents the sum of the functional scores; This represents the first functional score corresponding to the VKORC1 gene locus, with a value range of {0}; This represents the second functional score corresponding to the VKORC1 gene locus, with a value range of {1,2,3}. This represents the first functional score corresponding to the CYP2C9 gene locus, with a value range of {1,2,3}. This represents the second functional score corresponding to the CYP2C9 gene locus, with a value range of {2}; This represents the first functional score corresponding to the CYP4F2 gene locus, with a value range of {1,2,3}. This represents the second functional score corresponding to the CYP4F2 gene locus, with a value range of {2}. The theoretical value range is an integer between 7 and 15.

[0055] The metabolic grading module processes the sum of functional scores according to a preset grading threshold range to determine the drug metabolism rate grading index. The preset threshold range defines the correspondence between the drug metabolism rate grading index and the range of values ​​for the sum of functional scores, and is stored in the parameter configuration file of the metabolic grading module. The grading threshold range is divided as follows: when the sum of functional scores is greater than or equal to 13 and less than or equal to 15, the drug metabolism rate grading index is determined to be in the rapid metabolism category; when the sum of functional scores is greater than or equal to 10 and less than or equal to 12, the drug metabolism rate grading index is determined to be in the normal metabolism category; when the sum of functional scores is greater than or equal to 7 and less than or equal to 9, the drug metabolism rate grading index is determined to be in the slow metabolism category. The drug metabolism rate grading index is represented by an enumerated type value, with 3 coded for rapid metabolism, 2 for normal metabolism, and 1 for slow metabolism. The metabolic grading module then calculates the... The value is compared with the upper and lower bounds of the grading threshold interval to determine... The corresponding drug metabolism rate classification index is obtained from the range of values ​​that the value falls into. After the drug metabolism rate classification index is generated, the metabolism classification module transfers the drug metabolism rate classification index to the dose estimation module.

[0056] See Figure 4 In the graph, the vertical axis represents the number of patients, and the horizontal axis represents the range of the total functional score F, which ranges from 7 to 15. The graph uses a bar chart to show the distribution of patients across different ranges of total functional score. The legend uses red and blue dashed lines to indicate the threshold ranges for slow metabolism (F=7-9), normal metabolism (F=10-12), and rapid metabolism (F≥13).

[0057] As shown in the bar chart, the total functional score (F) of patients is mainly concentrated in the range of 10 to 12, with the number of patients peaking in this range, reflecting a high proportion of patients with normal metabolic rate. The next highest number of patients have a total functional score (F) in the range of 7 to 9, corresponding to slow metabolic rate patients. The number of patients with a total functional score (F) of 13 or higher is relatively small, reflecting a low proportion of fast metabolic rate patients in the overall sample.

[0058] Two threshold lines clearly delineate the functional score ranges for different drug metabolism rate grading indices, supporting the metabolic grading module's comprehensive assessment of patients' metabolic rates. Specifically, the slow metabolism level corresponds to a total functional score of 7 to 9, the normal metabolism level to 10 to 12, and the rapid metabolism level to 13 to 15, accurately reflecting the preset rules of the metabolic grading module for determining the grade based on the total functional score.

[0059] Example 4:

[0060] In practice, the dosage estimation module acquires the patient's current physical condition indicators and a list of concomitant medications. The physical condition indicators are derived from the patient's latest test records stored in the hospital or laboratory information system, including fields such as age, weight, serum creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and albumin. The list of concomitant medications is derived from the patient's medication records or a doctor's prescription list, including fields such as the name of the concomitant medications, daily dosage, dosing frequency, and start and end dates of use. The dosage estimation module receives the drug metabolism rate classification index output by the metabolic classification module and inputs the drug metabolism rate classification index, physical condition indicators, and the list of concomitant medications into the dosage estimation rule engine, outputting the patient's initial recommended anticoagulant dosage and dosage adjustment cycle.

[0061] The dosage calculation module performs validity verification on the body status indicator data, filtering out invalid indicator values ​​that exceed the preset physiological range to obtain valid body status indicator data. The validity verification process is as follows: the dosage calculation module reads each indicator item from the body status indicator data and compares it one by one with the upper and lower threshold values ​​in the preset physiological range configuration file. The preset physiological range configuration file is stored in the parameter directory of the dosage calculation module, defining the valid value range for each body status indicator. The preset physiological range for age is 0 to 120 years, for body weight is 10 kg to 300 kg, for serum creatinine is 10 μmol / L to 1500 μmol / L, for alanine aminotransferase (ALT) is 1 unit / L to 1000 units / L, for aspartate aminotransferase (AST) is 1 unit / L to 1000 units / L, and for albumin is 10 g / L to 60 g / L. The dosage calculation module evaluates each indicator. When the value of an indicator is within a preset physiological range, its original value is retained. When the value exceeds the preset physiological range, the indicator is marked as invalid and filtered out from the body condition indicator data. The remaining indicators after filtering out invalid values ​​constitute the valid body condition indicator data. During the filtering process, when an indicator is marked as invalid, the dosage calculation module sets the corresponding value to null, excluding it from subsequent calculations, and records the name and original value of the invalid indicator in the system log.

[0062] The dosage estimation module uses the drug metabolism rate grading index and effective physical condition indicators as initial dosage estimation factors, matching them with the corresponding basic dosage intervals in the dosage estimation rule engine. The dosage estimation rule engine is a rule-based forward matching inference module, containing a dosage rule table, a drug interaction rule table, and an inference control program. The dosage rule table is organized in a two-dimensional matrix, with row indices representing the drug metabolism rate grading index and column indices representing composite conditions including age, body weight, and liver function indicators. Table entries store the lower and upper bounds of the standard dosage interval.

[0063] During the matching of the baseline dose range, the dose estimation module extracts the patient's age, body weight, and liver function indicators from the effective physical condition data. The liver function indicators are calculated by combining the alanine aminotransferase (ALT), aspartate aminotransferase (AST), and albumin levels, multiplied by the albumin level (the ratio of ALT to AST). The dose estimation module combines the drug metabolism rate grading index, age, body weight, and liver function indicators into a multidimensional estimation factor vector. This multidimensional estimation factor vector is in the form of a quadruple, where the first element is the drug metabolism rate grading index, the second element is the age, the third element is the body weight, and the fourth element is the liver function indicator.

[0064] The dosage estimation module uses the multidimensional estimation factor vector as the matching key to retrieve the standard dose interval corresponding to the standard factor combination that is closest to the multidimensional estimation factor vector from the dosage rule table of the dosage estimation rule engine. The dosage rule table stores several standard factor combinations, each corresponding to a standard dose interval. Each standard factor combination is also organized in the form of a quadruple. The first element of the quadruple is the standard drug metabolism rate grading index, the second element is the standard median age, the third element is the standard median body weight, and the fourth element is the standard median liver function index. The dosage estimation module uses a weighted Euclidean distance algorithm to calculate the similarity distance between the multidimensional estimation factor vector and each standard factor combination. The formula for calculating the similarity distance is as follows:

[0065]

[0066] in, This represents the similarity distance between the multidimensional inferred factor vector and the standard factor combination; This represents the drug metabolism rate grading index in the multidimensional extrapolation factor vector. The value range is {1,2,3}, which corresponds to the slow metabolic level, normal metabolic level and fast metabolic level, respectively. This represents the standard drug metabolism rate grading index in the standard factor combination. The value range is {1,2,3}, which is preset by the dosage rule table; This represents the age value in the multidimensional extrapolation factor vector, in years; This represents the median standard age in the standard factor combination, in years. The median standard age is pre-set by the dose rule table according to age segments, typically in intervals of 10 years, with the midpoint of each interval serving as the median standard age. The normalization factor representing the age value. Setting it to 10 years old is used to eliminate the difference in units between age values ​​and other dimensions. This represents the volume mass value in the multidimensional extrapolation factor vector, in kilograms. This represents the median body mass of the standard body in the standard factor combination, in kilograms. The median body mass is pre-set by the dose rule table according to body mass segments, with typical segments being intervals of 10 kilograms each. The median of the interval is used as the median body mass. The normalization factor representing the body mass value. Set to 10 kg to eliminate the dimensional differences between body mass values ​​and other dimensions; This represents the numerical value of liver function indicators in the multidimensional extrapolation factor vector; This represents the median of the standard liver function index in the standard factor combination. The median of the standard liver function index is pre-set by the dose rule table according to the segmentation of the liver function index values. The segmentation is based on the decimal division of the liver function index values, and the median of each segment is used as the median of the standard liver function index. The normalization factor representing the values ​​of liver function indicators. Set as the average difference between the medians of liver function index values ​​across segments; The dimension weighting coefficients of the drug metabolism rate grading index are represented. The value was set to 0.4, based on the clinical validation results of the independent contribution of the metabolic rate grading of traditional Chinese medicine to dose adjustment. The dimension weight coefficient representing the age value. The value was set to 0.2, based on the assessment results of the contribution of age to drug clearance. The dimension weighting coefficients representing the body mass value. Set to 0.2, based on the evaluation results of the contribution of body mass to the distribution volume; The dimensional weighting coefficients representing the values ​​of liver function indicators. The value was set to 0.2, based on the assessment results of the contribution of liver function to drug metabolism.

[0067] The dose estimation module calculates the similarity distance between the multidimensional estimation factor vector and each standard factor combination in the dose rule table. Select the distance that makes the similarity similar. The standard factor combination with the smallest value is selected as the closest standard factor combination, and the standard dose interval corresponding to this closest standard factor combination is extracted. This standard dose interval is then determined as the basal dose interval. The basal dose interval includes a lower limit and an upper limit of the basal dose, in milligrams per day.

[0068] Optionally, when there are multiple standard factor combinations corresponding to similarity distances When the values ​​are equal and all are minimum, the dose estimation module obtains multiple candidate standard factor combinations, extracts the candidate dose intervals corresponding to the candidate standard factor combinations, takes the average of the lower bound values ​​of the candidate dose intervals as the lower bound value of the basic dose interval, and takes the average of the upper bound values ​​of the candidate dose intervals as the upper bound value of the basic dose interval.

[0069] After obtaining the baseline dose range, the dosage estimation module uses the name of each combined drug in the combined drug list to query the drug interaction rule table in the dosage estimation rule engine to obtain the corresponding dose correction coefficient for each combined drug. The drug interaction rule table is stored in the knowledge base of the dosage estimation rule engine and includes fields for interfering drug name, enzyme induction coefficient, and enzyme inhibition coefficient. A value greater than 1 in the enzyme induction coefficient field indicates that the interfering drug accelerates the metabolism of anticoagulants, requiring an upward dose adjustment; a value between 0 and 1 in the enzyme inhibition coefficient field indicates that the interfering drug inhibits the metabolism of anticoagulants, requiring a downward dose adjustment. The drug interaction rule table is constructed based on published drug interaction research literature and pharmacokinetic interaction parameters recorded in drug instructions.

[0070] Specifically, the dosage estimation module performs string matching on each combined drug name in the combined drug list against the interfering drug name field in the drug interaction rule table. The string matching uses a case-insensitive full-field exact match algorithm. When a combined drug name successfully matches the interfering drug name field of a record in the drug interaction rule table, the dosage estimation module extracts the enzyme induction coefficient or enzyme inhibition coefficient associated with that interfering drug name from the drug interaction rule table. When a record contains both enzyme induction coefficient and enzyme inhibition coefficient fields, the enzyme induction coefficient is extracted first if it has a non-default value, and the enzyme inhibition coefficient is extracted first if it has a non-default value. Only when both fields have values ​​is the product of the two field values ​​taken as the mixed effect coefficient. When a combined drug name fails to match the interfering drug name field of all records in the drug interaction rule table, the dosage estimation module sets the dosage correction coefficient corresponding to that combined drug to a neutral value of 1.0. The dosage estimation module then summarizes all extracted enzyme induction coefficients, enzyme inhibition coefficients, and neutral values ​​into a dosage correction coefficient corresponding to each combined drug.

[0071] In some implementations, the combined medication list contains p combined drugs, where p is an integer greater than or equal to zero. When p equals 0, the combined medication list is empty, the dosage estimation module does not perform the drug interaction rule table lookup operation, and the set of all dosage correction coefficients is empty. The dosage estimation module calculates the initial recommended anticoagulant dosage and dosage adjustment period based on the baseline dosage range and all dosage correction coefficients. The initial recommended anticoagulant dosage is obtained by multiplying all dosage correction coefficients together with the median of the baseline dosage range, using the median of the baseline dosage range as a reference. The dosage adjustment period is determined based on the variability of the drug metabolism rate grading index and the dosage correction coefficients: when the drug metabolism rate grading index is in the slow metabolism category and the standard deviation of the dosage correction coefficient is greater than 0.3, the dosage adjustment period is set to 3 days; when the drug metabolism rate grading index is in the slow metabolism category and the standard deviation of the dosage correction coefficient is less than or equal to 0.3, the dosage adjustment period is set to 5 days; when the drug metabolism rate grading index is in the normal metabolism category, the dosage adjustment period is set to 7 days; and when the drug metabolism rate grading index is in the rapid metabolism category, the dosage adjustment period is set to 5 days. The dose adjustment cycle is set based on population statistics from clinical pharmacokinetic studies of patients with different metabolic phenotypes who require time to reach steady-state plasma drug concentrations. The dose estimation module outputs the calculated initial recommended anticoagulant dose and dose adjustment cycle to the protocol output module.

[0072] See Figure 5In the graph, the horizontal axis represents the product of dose correction factors, ranging from approximately 0.4 to 2.3; the vertical axis represents the recommended initial anticoagulant dose in milligrams per day, ranging from approximately 1 to 16 mg / day. The data points in the graph are divided into three categories: slow metabolizer (red dots), normal metabolizer (green squares), and fast metabolizer (blue triangles). The classification of these three metabolic levels is based on the drug metabolism rate grading index output by the metabolic grading module.

[0073] As shown in the figure, the recommended initial anticoagulant doses for all three metabolic grades monotonically increase with the product of dose correction factors, but there is a clear stratification among the recommended doses for different metabolic grades. The recommended dose for the rapid metabolizer grade is generally the highest, ranging from approximately 3 mg / day to 16 mg / day, and the dose varies significantly with the correction factor. The recommended dose for the normal metabolizer grade is in the middle, ranging from approximately 2 mg / day to 11 mg / day, and the data points are relatively dense and evenly distributed. The recommended dose for the slow metabolizer grade is the lowest, ranging from approximately 1 mg / day to 5.7 mg / day, and the curve trend is relatively flat.

[0074] Example 5:

[0075] In specific implementation, please refer to Figure 2 The protocol output module receives the initial recommended anticoagulant dosage and dosage adjustment period from the dosage calculation module. The initial recommended anticoagulant dosage includes the total daily dosage in milligrams. The dosage adjustment period includes the total number of adjustment days in days. Based on the initial recommended anticoagulant dosage and dosage adjustment period, the protocol output module generates an individualized medication guidance plan for the patient.

[0076] The protocol output module breaks down the initial recommended anticoagulant dosage into single-dose doses and dosing intervals according to the dosage adjustment cycle. Specifically, the breakdown process involves the protocol output module obtaining the total daily dose value included in the initial recommended anticoagulant dosage and the total number of adjustment days within the dosage adjustment cycle. The total daily dose value is the total number of milligrams of anticoagulant medication taken daily, calculated by the dosage estimation module. The total number of adjustment days represents the length of the dose stability observation period, calculated from the start date of medication.

[0077] The protocol output module obtains the patient's medication adherence time window parameters, which include the range of acceptable daily dosing frequencies. These parameters are stored in the patient's personalized configuration file and are pre-set by the physician based on the patient's lifestyle and medication habits. The range of acceptable daily dosing frequencies is defined as a closed interval, with the lower bound being the minimum acceptable number of doses and the upper bound being the maximum acceptable number of doses. Both the lower and upper bounds are positive integers, and the lower bound is less than or equal to the upper bound. Typically, the acceptable daily dosing frequency range is set to 1 to 3 times.

[0078] The protocol output module calculates the single-dose dose based on the total daily dose and the acceptable daily dosing frequency range. The module determines the actual average daily dosing frequency from the acceptable daily dosing frequency range. The rules for determining the average daily dosing frequency are as follows: when the lower bound of the acceptable daily dosing frequency range equals the upper bound, the average daily dosing frequency is taken as this unique value; when the lower bound of the acceptable daily dosing frequency range is less than the upper bound, the average daily dosing frequency is taken as the midpoint of the acceptable daily dosing frequency range, which is the sum of the lower and upper bound values ​​divided by 2, rounded down. The single-dose dose is obtained by dividing the total daily dose by the average daily dosing frequency, as expressed in the formula:

[0079]

[0080] in, This indicates the single-dose dose, expressed in milligrams. This indicates the total daily dose, expressed in milligrams. Extracted from the recommended initial anticoagulant dosage; Indicates the average number of times medication is administered per day. Determined by the acceptable daily dosing frequency range in the medication adherence time window parameter. It is a positive integer, and its value ranges from 1 to 3. The method for determining this is as follows: Let the lower bound of the acceptable daily dosing frequency range be... The upper boundary is ,when equal hour, Assignment ;when Less than hour, Assigned value ,in This indicates the floor function.

[0081] The protocol output module determines the dosing interval value based on the total number of adjustment days and the acceptable daily dosing frequency range. The dosing interval value represents the time interval between two consecutive dosings, in hours. It is determined by dividing a 24-hour day by the average daily dosing frequency. The dosing interval value is obtained. The dosing interval value is retained to one decimal place.

[0082] In some implementations, when the calculated single-dose dose is less than the single-dose content of the smallest-sized anticoagulant tablet, the protocol output module adjusts the single-dose dose to the single-dose content of the smallest-sized tablet and extends the dosing interval accordingly. When the single-dose dose is greater than the maximum safe dose threshold, the protocol output module truncates the single-dose dose to the maximum safe dose threshold and distributes the excess dose to additional dosing sessions, recalculating the dosing interval.

[0083] The protocol output module generates a time-series dosing plan table, including the dosing start time and dosing end time, based on the single-dose dosage value and dosing interval value. The dosing start time is set to the first dosing time when the patient begins medication, which defaults to 8:00 AM on the day the prescription is issued. The dosing end time is calculated by adding the total number of adjustment days to the dosing start time, and the end time is the same moment after the total number of adjustment days following the start time. The time-series dosing plan table is organized in list form, with each item in the list recording a specific dosing time point and its corresponding dosage value. Starting from the dosing start time, the protocol output module generates a dosing time point every time a dosing interval value is elapsed, until the dosing end time is exceeded. The dosage value at each dosing time point is set to the single-dose dosage value.

[0084] In practice, the protocol output module performs conflict verification between the time-series dosing schedule and the patient's medication contraindication labels. Medication contraindication labels are stored in the patient's electronic health record and entered by healthcare professionals, containing a list of contraindicated time periods. Each contraindicated time period has a start and end time, represented in HH:MM format, and is independent of a specific date. Medication contraindication labels can include fasting periods, sleep-supporting periods, and fixed dosing periods for other medications.

[0085] The conflict resolution process is as follows: The protocol output module iterates through each dosing time point in the time-series dosing schedule, extracting the hour and minute portions of each dosing time point. The extracted hour and minute portions are compared with the start and end times of each contraindication period in the contraindication label. If the hour and minute portions of a dosing time point fall within a closed interval of a contraindication period, the dosing time point is determined to overlap with that contraindication period. When overlap is determined, the protocol output module performs offset correction processing on the dosing time point. The offset correction process involves shifting the overlapping dosing time point in the future direction by a fixed step size, set to 30 minutes. After the shift, it checks whether the new dosing time point still overlaps with a contraindication period. If it still overlaps, the shift continues with a step size of 30 minutes until it falls within a non-contraindication period. If the shifted dosing time point results in an interval less than the minimum safe interval threshold (set to 4 hours), the protocol output module no longer shifts the dosing time point. Instead, it merges the dose of that dosing time point into the previous dosing time point and deletes the overlapping dosing time point record from the time-series dosing schedule. After the above offset correction and dose merging processes, the protocol output module obtains the corrected time-series dosing schedule.

[0086] The protocol output module outputs the corrected dosing schedule as a personalized medication guidance plan. This plan is output in structured text format, including patient identification, drug name, single-dose dosage, dosing interval, the corrected dosing schedule, and medication reminders. These reminders are automatically generated based on the patient's contraindications and offset corrections, highlighting precautions to take at each time point. The personalized medication guidance plan is then output to the doctor's workstation interface in the hospital information system and to the patient's mobile application for review by doctors and implementation by patients.

[0087] See Figure 6 The horizontal axis of the graph represents the patient's total daily dose in milligrams, ranging from 1 to 15 milligrams; the vertical axis represents the dosing interval in hours, ranging from 0 to 25 hours. The legend distinguishes three different daily dosing frequencies: red dots represent 1 daily dosing frequency, blue squares represent 2 daily dosing frequencies, and green diamonds represent 3 daily dosing frequencies.

[0088] As can be seen from the figure, when the average number of doses per day is 1, the dosing interval values ​​of all data points are concentrated around 24 hours, and the total daily dose values ​​are distributed in the range of 1 to 15 mg. This indicates that when the protocol output module sets the medication adherence time window parameter in the patient's individualized medication guidance protocol to once a day, the calculated dosing interval is stable at 24 hours, and the single dose is equal to the total daily dose value.

[0089] When the average number of doses per day is 2, the dosing interval values ​​of the data points are concentrated around 12 hours, and the distribution of the total daily dose values ​​also covers the range of 1 to 15 mg. This indicates that when the average number of doses per day is 2, the dosing interval value of the protocol output module is clearly equal to 24 hours divided by 2, that is, 12 hours, and the single dose is the total daily dose value divided by 2.

[0090] When the average number of doses per day is 3, the dosing interval values ​​of the data points are approximately 8 hours, and the distribution is relatively concentrated. Furthermore, the total daily dose values ​​cover the same range, indicating that when the average number of doses per day is 3, the protocol output module sets the dosing interval value to 24 hours divided by 3, i.e., 8 hours, and the single dosing dose is the total daily dose value divided by 3.

[0091] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A personalized anticoagulation medication guidance and analysis system based on gene polymorphism, characterized in that, include: The data acquisition module acquires the original gene detection data related to the patient's anticoagulant medication, and performs gene polymorphism site identification processing on the original gene detection data to generate multi-site genotype combination data containing drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites. The annotation mapping module queries a preset gene-phenotype mapping knowledge base based on the multi-site genotype combination data to determine the metabolic phenotype category and target sensitivity category corresponding to each gene site, and generates a site function annotation set. The metabolic grading module uses the site function annotation set to perform comprehensive grading of the anticoagulant drug metabolism rate of the patient, and obtains the drug metabolism rate grading index corresponding to the patient. The dosage estimation module acquires the patient's current physical condition index data and a list of concomitant medications, and inputs the drug metabolism rate classification index, the physical condition index data, and the list of concomitant medications into the dosage estimation rule engine. It then outputs the patient's initial recommended anticoagulant dosage and dosage adjustment cycle, including: The physical condition index data is subjected to index validity verification processing to filter out invalid index values ​​that exceed the preset physiological range, thereby obtaining valid physical condition index data; The drug metabolism rate classification index and the effective physical state index data are used as initial dose estimation factors, and the corresponding basic dose range is matched in the dose estimation rule engine. Using the name of each combined drug in the combined drug list, query the drug interaction rule table in the dose estimation rule engine to obtain the dose correction coefficient corresponding to each combined drug; Calculate the recommended initial anticoagulant dose and the dose adjustment cycle based on the baseline dose range and all the dose correction factors. The protocol output module generates an individualized medication guidance protocol for the patient based on the initial recommended dose of anticoagulant and the dose adjustment cycle.

2. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, The original gene detection data is processed to identify gene polymorphic sites, generating multi-site genotype combination data containing drug-metabolizing enzyme gene sites, target gene sites, and transporter gene sites, including: The original gene detection data were subjected to base sequence alignment processing to obtain the gene sequence fragment of the patient; Based on the preset reference sequence positions corresponding to the VKORC1 gene locus, CYP2C9 gene locus, and CYP4F2 gene locus, target sequence fragments are extracted from the regions where each gene locus is located from the gene sequence fragment. Each target sequence fragment is subjected to base variation identification processing to determine the allele type at each gene locus, thereby obtaining the multi-locus genotype combination data.

3. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 2, characterized in that, The step of extracting target sequence fragments from the gene sequence fragments based on the preset reference sequence positions corresponding to the VKORC1, CYP2C9, and CYP4F2 gene loci includes: Obtain the reference sequence location information corresponding to the preset VKORC1 gene locus, CYP2C9 gene locus, and CYP4F2 gene locus. The reference sequence location information includes the start coordinates and end coordinates of each gene locus on the reference genome. Based on the start coordinates and the end coordinates, extract the original region sequence corresponding to each gene locus from the gene sequence fragment; Each of the original region sequences is pruned to remove redundant bases from the non-coding regions at both ends of the original region sequence, thereby obtaining the target sequence fragment.

4. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, Based on the multi-locus genotype combination data, a pre-defined gene-phenotype mapping knowledge base is queried to determine the metabolic phenotype category and target sensitivity category corresponding to each gene locus, generating a locus function annotation set, including: Using each gene locus in the multi-locus genotype combination data as a query index, find phenotypic record entries in the gene-phenotype mapping knowledge base that match the allele type of each gene locus. The metabolic phenotype category corresponding to each gene locus is extracted from the phenotype record entries. The metabolic phenotype categories include ultra-rapid metabolizer, extensive metabolizer, intermediate metabolizer, and slow metabolizer. Extract the target sensitivity category corresponding to each gene locus from the phenotypic record entries. The target sensitivity category includes high sensitivity, normal sensitivity, and low sensitivity. The metabolic phenotype category and the target sensitivity category of each gene locus are associated and combined to generate the site functional annotation set.

5. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, The process of using the site functional annotation set to perform comprehensive classification of the patient's anticoagulant drug metabolism rate yields a corresponding drug metabolism rate classification index for the patient, including: Extract the metabolic phenotype category and target sensitivity category corresponding to all gene loci from the site function annotation set; According to the preset metabolic phenotype scoring rules, a first functional score value is assigned to the metabolic phenotype category of each gene locus; According to the preset target sensitivity scoring rules, a second functional score value is assigned to the target sensitivity category of each gene locus. The first and second functional scores corresponding to all the gene loci are summed to obtain the total functional score. The sum of the functional scores is graded according to a preset grading threshold range to obtain the drug metabolism rate grading index.

6. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, The step of using the drug metabolism rate classification index and the effective physical status index data as initial dose estimation factors, and matching the corresponding basic dose range in the dose estimation rule engine, includes: Extract the patient's age, body weight, and liver function indicators from the effective physical condition index data; The drug metabolism rate grading index, the age value, the body weight value, and the liver function index value are combined into a multidimensional extrapolation factor vector. Using the multidimensional extrapolation factor vector as the matching key, the standard dose interval corresponding to the standard factor combination that is closest to the multidimensional extrapolation factor vector is retrieved from the dose rule table of the dose extrapolation rule engine. The standard dose range is defined as the base dose range.

7. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, The step involves using the name of each combined drug in the combined drug list to query the drug interaction rule table in the dosage estimation rule engine to obtain the dosage correction coefficient corresponding to each combined drug, including: Each combined drug name in the combined drug list is matched with the interfering drug name field in the drug interaction rule table; When the merged drug name successfully matches the interfering drug name field, the enzyme induction coefficient or enzyme inhibition coefficient associated with the interfering drug name is extracted from the drug interaction rule table; When the combined drug name fails to match the interfering drug name field, the dose correction coefficient corresponding to the combined drug is set to a neutral value. All extracted enzyme induction coefficients and enzyme inhibition coefficients, along with the neutral values, are summarized into dose correction coefficients for each combined drug.

8. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 1, characterized in that, The process of generating an individualized medication guidance plan for the patient based on the initial recommended anticoagulant dose and the dose adjustment cycle includes: According to the aforementioned dose adjustment cycle, the recommended initial anticoagulant dose is broken down into single-dose dose values ​​and dosing interval values; Based on the single-dose dose value and the dosing interval value, a time-series dosing plan table including the dosing start time and dosing end time is generated; The timing dosing schedule is checked for conflict with the patient's medication contraindication label. When the dosing time point in the timing dosing schedule overlaps with the contraindication time period in the medication contraindication label, the dosing time point is offset and corrected to obtain the corrected timing dosing schedule. The corrected timing dosing schedule will be used as the individualized medication guidance plan.

9. The anticoagulation personalized medication guidance and analysis system based on gene polymorphism according to claim 8, characterized in that, The step of breaking down the initial recommended anticoagulant dose into single-dose dose values ​​and dosing interval values ​​according to the dose adjustment cycle includes: Obtain the total daily dose value included in the initial recommended anticoagulant drug dose and the total number of adjustment days included in the dose adjustment cycle; Obtain the patient's medication adherence time window parameters, which include the range of acceptable daily dosing frequencies; The single-dose dose value is calculated based on the total daily dose value and the acceptable daily dosing frequency range; The dosing interval value is determined based on the total number of adjustment days and the acceptable range of daily dosing frequency.

Images