A model data standardization processing method and system for drug testing
By performing steps such as synchronous arrangement, gradient sorting, and batch correction of drug test data, the problem of local outliers affecting the overall results in drug testing was solved, thereby improving the stability and reliability of the data.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANDONG AIMENG BIOTECHNOLOGY CO LTD
- Filing Date
- 2026-05-11
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies do not handle local outliers sufficiently well in drug testing, which affects the accuracy of the overall results.
By performing steps such as synchronous arrangement, gradient sorting, multi-well reading screening, baseline correction, and batch correction on drug experimental data, standardized data output with consistency and sequence is formed.
It enhances the correlation between time and dosage dimensions of drug test data, reduces interference from local abnormal readings, improves the continuity and interpretability of data, reduces errors caused by batch environmental differences, and improves the stability and reliability of data.
Smart Images
Figure CN122177510A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of biomedical technology, and in particular to a method and system for standardizing model data processing for drug testing. Background Technology
[0002] The field of biomedical technology encompasses drug discovery, efficacy evaluation, toxicology research, in vitro model construction, animal model experiments, and preclinical data management. The core of this field lies in the quantitative recording and comparability analysis of the response processes of candidate drugs in different experimental models. Its overall system consists of cell experiment data acquisition, animal experiment indicator recording, integration of physiological and biochemical test results, organization of time-series observation information, and cross-experimental platform data archiving.
[0003] Chinese patent document CN117556034B discloses a data processing system for standardizing the output results of an electronic medical record question-and-answer model. The system includes a processor and a memory storing a computer program. When the computer program is executed by the processor, the following steps are performed: obtaining a key entity set based on a sample database; inputting the key entity set and a target entity set into a first intermediate model; obtaining a key entity vector set and a target entity vector set; inputting the key entity vector set and the target entity vector set into a second intermediate model; obtaining a final entity set; obtaining a target model; inputting a first candidate entity into the target model; and obtaining a second candidate entity set to achieve standardization processing of the target text. This invention does not limit itself to a single method when obtaining target priorities. By combining multiple methods, the accuracy of obtaining the priority corresponding to the entity is improved, thereby making the standardization results corresponding to the output results of the electronic medical record question-and-answer model more accurate.
[0004] The existing technology has the following problems: Existing technologies focus on data field uniformity and format constraints during actual operation, but lack fine-grained processing mechanisms for internal differences in the original detection signals. The reading fluctuations between different well positions are easily masked by overall averaging, resulting in local outliers being retained and entering the subsequent analysis stage, thus affecting the accuracy of the overall results. Summary of the Invention
[0005] The main objective of this invention is to provide a method and system for standardizing model data processing for drug testing, which can effectively solve the problem that local outliers are retained and enter the subsequent analysis stage, thereby affecting the accuracy of the overall results.
[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for standardizing model data for drug testing includes the following steps: S1: Acquisition of raw records for drug experiments: Obtain the cell culture plate number, drug concentration value, culture time scale, and absorbance reading of the enzyme-linked immunosorbent assay (ELISA) reader for the humanized mouse model used in drug testing; arrange the cell culture plate number with the corresponding culture time scale; sort the drug concentration values in a gradient; match the sorted drug concentration values with the absorbance reading of the ELISA reader in the same row to form the experimental record sequence value; S2: Detection signal consistency screening: Based on the experimental record sequence value, obtain the absorbance reading of each well under the same cell culture plate number, calculate the difference between each well reading and the average absorbance reading of the plate, and if the difference exceeds the set absorbance offset threshold, it is removed to obtain the effective absorbance set value. S3: Concentration response normalization processing: Based on the effective absorbance set value, obtain the corresponding drug concentration value and perform logarithmic transformation. Pair the transformed concentration value with the corresponding absorbance reading and calculate the ratio of the absorbance reading at each concentration point to the lowest concentration absorbance reference value. Then, scale each ratio with the highest concentration absorbance reference value to form a normalized sequence in the range of 0 to 1. Perform adjacent difference calculation on the normalized sequence. If the difference result meets the set response change threshold, it is retained to generate the concentration response normalization coefficient. S4: Inter-batch difference correction: Call the concentration response normalization coefficient, obtain the set of normalization coefficients under different experimental batch identifiers, calculate the mean of normalization coefficients for each batch, extract the mean difference between batches, and if this difference exceeds the set batch offset threshold, perform a uniform shift value adjustment on the corresponding batch to obtain the batch correction coefficient value. S5: Standardized data output construction: Based on the batch correction coefficient value, obtain the corresponding concentration response normalization coefficient and perform superposition operation. Combine and map the superposition result with the drug concentration value, rearrange the output according to the cell culture plate number order, and generate standardized pharmacodynamic data values.
[0007] Preferably, the experimental record sequence values include cell culture plate number sequence, drug concentration gradient value, and absorbance matching data; the effective absorbance set values include well absorbance data, mean deviation screening value, and effective reading subset; the concentration response normalization coefficient includes logarithmic concentration value, normalized ratio sequence, and response difference characteristic value; the batch correction coefficient values include batch mean parameter, batch difference offset, and translation adjustment coefficient; and the standardized pharmacodynamic data values include corrected normalized data, concentration mapping data, and sorting output results.
[0008] Preferably, in the S1 step of collecting original records of drug experiments, multi-dimensional physiological and pathological raw data of in vivo drug efficacy test of humanized mouse model corresponding to drug test are collected simultaneously. The data includes blood biochemical indicators, urine metabolic indicators, weight change value, blood glucose dynamic value, behavioral test score, imaging parameters, semi-quantitative results of pathological staining, proportion of cell subsets by flow cytometry, and gray value of protein expression by Western blot. The cell culture plate number corresponds to the mouse individual number, the drug concentration value corresponds to the drug dosage, the culture time scale corresponds to the drug intervention duration, and the multidimensional physiological and pathological raw data corresponds to the absorbance reading of the enzyme-linked immunosorbent assay (ELISA) reader. The individual mouse numbers were arranged with their corresponding drug administration durations, and the drug doses were sorted in a gradient. The sorted drug doses were matched with the original in vivo detection data from peers, and together with the data collected at the cellular level, a complete experimental record sequence value was formed.
[0009] Preferably, in the S2 step of signal consistency screening, for the original in vivo multidimensional physiological and pathological data contained in the experimental record sequence values, for the parallel repeated detection data of the same mouse under the same drug intervention duration, the absorbance readings of the multi-well positions under the same cell culture plate number are used to calculate the relative deviation between the repeated detection values and the mean of the data set. If the relative deviation exceeds the set parallel sample deviation threshold, the data set is removed. For data from parallel experiments of the same batch of models, the mean deviation between groups is calculated. If the deviation exceeds the set operational deviation threshold, invalid data is removed. For semi-quantitative pathological staining data, the mean is read by two people in a double-blind manner, and data with a reading difference exceeding the set subjective deviation threshold is removed. Finally, the selected in vivo effective data and cellular effective data together constitute the effective absorbance aggregate value.
[0010] Preferably, before the concentration response normalization process in step S3, a baseline correction step for individual humanized mouse models is added: before drug administration, baseline values of basic physiological indicators for each mouse are collected, including baseline blood glucose, baseline body weight, baseline target protein expression level, and baseline behavioral score. The baseline values correspond to the in vivo scenario baseline of the lowest concentration absorbance benchmark value. After obtaining the effective absorbance aggregate value, the in vivo detection value corresponding to each drug administration intervention duration is subtracted from the corresponding individual baseline value of the mouse to obtain the baseline-corrected detection value. Then, the concentration response normalization process is performed together with the effective data at the cellular level to finally generate the concentration response normalization coefficient.
[0011] Preferably, the batch-to-batch difference correction in step S4 further includes additional processing for humanized mouse model strains and construction batch difference correction: a baseline response database of mouse models with different construction routes and genetic backgrounds conforming to industry standards is established in advance, and the baseline threshold of the baseline response database corresponds to the extended limitation of the batch offset threshold; after obtaining the set of normalized coefficients under different experimental batch identifiers, the normalized coefficients of each batch are first compared with the baseline response coefficients of the corresponding models to calculate the model background deviation value. If the deviation value exceeds the set model background threshold, the model background correction coefficient is superimposed on the translation value adjustment process, and the batch correction coefficient value is finally output.
[0012] Preferably, the standardized data output construction in step S5 also includes the integrated and standardized output of multi-omics data: synchronously collecting raw proteomic and metabolomic data of humanized mice after drug intervention; performing outlier removal, abundance normalization, and batch correction on the proteomic data; and performing peak alignment, baseline correction, and relative quantification normalization on the metabolomic data; mapping and matching the standardized proteomic and metabolomic data with the standardized pharmacodynamic data values one-to-one based on mouse individual number / cell culture plate number, drug concentration / dose, culture time scale / drug intervention duration; and rearranging them according to the cell culture plate number order to generate standardized pharmacodynamic data values containing pharmacodynamic, proteomic, and metabolomic dimensions.
[0013] Preferably, after the standardized data output is constructed in step S5, an additional step of standardized data verification and full-process traceability is added: the generated standardized drug efficacy data values are cross-validated with the corresponding drug testing model verification data, wherein the model verification data includes electrophysiological recording results, lesion pathological feature verification results, and disease-related signaling pathway activation level verification results. Simultaneously, based on the cell culture plate number, mouse individual number, and experimental batch identifier, a full-process data traceability chain is established to fully record model construction parameters, drug administration parameters, detection equipment parameters, and experimental operator information. Finally, based on standardized efficacy data values, a traceable standardized data report that meets the requirements for drug research and development registration is generated.
[0014] A model data standardization processing system for drug testing includes a data acquisition module, a signal filtering module, a normalization processing module, a batch calibration module, a data integration and output module, and a validation and traceability module. Data acquisition module: Used to collect raw data from in vitro cell culture plate experiments and in vivo multidimensional physiological and pathological raw data of humanized mouse models used for drug testing. The raw data includes absorbance readings from an ELISA reader, blood biochemical indicators, urine metabolic indicators, weight changes, blood glucose dynamics, behavioral test scores, imaging parameters, semi-quantitative results of pathological staining, cell subset ratios from flow cytometry, and gray values from Western blotting. The module associates cell culture plate number, mouse individual number, drug concentration / dosage, and culture time scale / intervention duration to form an experimental record sequence. Signal filtering module: used to perform consistency verification on the collected experimental record sequences, and remove invalid data that exceeds the set deviation threshold, including data with abnormal well absorbance, excessive parallel sample deviation, abnormal batch operation deviation, and abnormal subjective interpretation difference; generate valid in vitro and in vivo data sets; Normalization module: Used to perform baseline correction, logarithmic concentration conversion, concentration response ratio calculation and adjacent difference screening on valid datasets, and generate concentration response normalization coefficients; at the same time, it performs outlier removal, abundance / baseline normalization and batch correction on proteomics and metabolomics data. Batch correction module: used to calculate the difference in mean normalization coefficients between different experimental batches, and combined with the humanized mouse model construction route and genetic background baseline response database, to perform translation adjustment and model background correction, and generate batch correction coefficients; Data integration and output module: This module maps and matches the batch-corrected concentration response normalization coefficients with standardized proteomics and metabolomics data, and outputs the data in sorted order by cell culture plate number, mouse individual number, drug concentration / dosage, and intervention duration, forming standardized pharmacodynamic data values that include pharmacodynamic, proteomics, and metabolomics dimensions. The verification and traceability module is used to cross-validate standardized pharmacodynamic data with model validation data, establish a full-process data traceability chain, record model construction, dosing parameters, testing equipment and experimental operation information, and generate traceable standardized data reports that meet the requirements for drug research and development registration.
[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention provides a model data standardization processing method and system for drug testing. By synchronizing the culture plate number and time scale in the original experimental records and performing gradient sorting on the drug concentration, the data under different time and dosage conditions form a sequence structure with order constraints, thereby enhancing the correspondence between experimental data in the time and dosage dimensions.
[0016] 2. This invention provides a model data standardization processing method and system for drug testing. By combining the absorbance readings of multiple wells with the deviation of the mean value within the plate, data that deviates from the threshold are screened out, so that the signal source has a consistent basis and the interference of local abnormal readings on the overall results is reduced.
[0017] 3. This invention provides a model data standardization processing method and system for drug testing. In the processing of concentration-response relationship, logarithmic transformation and interval scaling are introduced to compress the response changes corresponding to different orders of magnitude concentrations into a uniform scale interval. At the same time, the response points with stable change trends are retained through differential constraints, thereby improving the continuity and interpretability of the concentration response curve.
[0018] 4. This invention provides a model data standardization processing method and system for drug testing. By extracting the mean difference and uniformly shifting the data across batches, the systematic bias between different experimental batches is corrected, and the structural error caused by batch environmental differences is reduced.
[0019] 5. This invention provides a model data standardization processing method and system for drug testing. By mapping and rearranging the normalized response coefficient with the original concentration, a data output structure with a unified expression and consistent order is formed, which enables the data to have higher comparative stability and trend consistency in subsequent analysis, while enhancing the data fusion ability and statistical reliability under different experimental conditions. Attached Figure Description
[0020] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the S1 process structure of the present invention; Figure 3 This is a schematic diagram of the S2 process structure of the present invention; Figure 4 This is a schematic diagram of the S3 process structure of the present invention; Figure 5 This is a schematic diagram of the S4 process structure of the present invention; Figure 6 This is a schematic diagram of the S5 process structure of the present invention; Figure 7 This is a schematic diagram of the system flow of the present invention. Detailed Implementation
[0021] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.
[0022] Example 1: Standardization of efficacy data of Aβ scavenger X in a humanized mouse model of Alzheimer's disease.
[0023] We constructed APPswe / PS1ΔE9 double mutant humanized Alzheimer's mice (C57BL / 6 background) using CRISPR / Cas9 technology to evaluate the in vitro and in vivo efficacy of the candidate Aβ scavenger X. The experiment consisted of three independent batches, each including cellular (96-well plate SH-SY5Y cell Aβ aggregation inhibition assay) and animal (20 mice, randomly divided into control, low / medium / high dose groups, n=5 per group) data. The assays included cellular absorbance, blood biochemistry, brain pathological staining, protein expression, and proteomics data.
[0024] Full-dimensional collection of raw records Simultaneously collect data at the cellular and animal levels, establish unique association identifiers, and bind model construction parameters (CRISPR editing sites, embryo transfer batches), drug administration parameters, and detection data one by one to generate the original drug experiment record association data shown in Table 1.
[0025] Table 1: Relationship Table of Original Records for Drug Experiments in Humanized Mouse Models of Alzheimer's Disease Table 1 shows the correlation records between cell culture data and animal model data under different detection dimensions.
[0026] Multimodal data hierarchical consistency screening Differential deviation control was performed for different data types to remove invalid data and generate the multimodal data stratified consistency screening results shown in Table 2.
[0027] Table 2: Statistical Table of Hierarchical Consistency Screening Results for Multimodal Detection Data Table 2 shows the consistency screening results for different types of experimental data. Invalid data were removed by stratified bias control to ensure the data quality of subsequent standardization processing.
[0028] Baseline correction and concentration response normalization First, the baseline values of individual mice, such as the basic Aβ expression level and basic behavioral score before drug administration, were subtracted. Then, the drug concentration was logarithmically transformed and normalized to 0-1 with the lowest absorbance of the concentration being 0 and the highest absorbance of the concentration being 1. The effective response points with a difference of >0.1 were retained, and the concentration response normalized data shown in Table 3 were generated.
[0029] Table 3: Results of Normalized Treatment of Dosage Concentration-Absorbance Response Table 3 shows the concentration data after logarithmic transformation, the normalized absorbance response values, and the results of screening effective response points after adjacent difference calculation.
[0030] Batch and Model Background Comprehensive Correction The pre-established APP / PS1 humanized mouse baseline response database was called to correct the systematic bias between the three batches, generating the batch and model background comprehensive correction data shown in Table 4.
[0031] Table 4: Comprehensive Correction Coefficients for Experimental Batches and Model Background Table 4 lists the mean normalization coefficients, deviations from the model baseline, and shift adjustment coefficients for different experimental batches, ultimately yielding a comprehensive correction coefficient that can be used for data correction.
[0032] Multi-omics integration and data validation and source tracing Standardized pharmacodynamic data and TMT-labeled quantitative proteomic data were correlated based on mouse IDs to generate a multidimensional dataset. Validation results showed that the area of brain Aβ plaques in the high-dose drug X group decreased by 71.2%, with 96.2% consistency with the synaptic function recovery rate (68.5%) recorded by electrophysiological recordings. Simultaneously, a complete traceability chain was established to record parameters at every stage, including gene editing, embryo transfer, drug administration, and detection.
[0033] Example 2: Standardization of efficacy data of hypoglycemic drug Y in an insulin-resistant humanized mouse model.
[0034] A humanized mouse model of insulin resistance was constructed using CRISPR / Cas9 knockout mice with the IRS-1 gene, and the hypoglycemic effect of a novel GLP-1 receptor agonist, Y, was tested. Two independent batches were set up, with 16 mice in each batch. Fasting blood glucose, serum insulin, glycated hemoglobin, and urinary metabolic indicators were measured.
[0035] Raw data collection: The mouse individual number, drug dosage (0, 0.1, 0.3, 1 mg / kg) and daily dynamic blood glucose value were linked, and serum insulin concentration and model construction batch information were collected simultaneously.
[0036] Consistency screening: Data with a coefficient of variation >8% in parallel blood glucose tests were removed, totaling 5 groups, with an overall effective rate of 96.9%; Serum insulin data were analyzed by taking the mean of two replicates and removing samples with a coefficient of variation >10%.
[0037] Baseline correction: Based on fasting blood glucose and fasting insulin levels before drug administration, the measured values at each time point were corrected to eliminate individual differences in basal metabolism among mice.
[0038] Batch calibration: The mean difference of the normalization coefficients of the two batches is 0.028, which is lower than the preset threshold of ±0.04, so no additional translation adjustment is required; compared with the insulin resistance model benchmark database, the model background correction coefficient of 0.022 is superimposed.
[0039] Standardized output: A dose-glucose reduction rate standardized curve was generated. The results showed that after 28 days of drug Y 1mg / kg group, fasting blood glucose decreased by 42.3% and insulin sensitivity increased by 57.8%. The data can be directly used for cross-laboratory drug efficacy comparison.
[0040] Example 3: Standardization of cytotoxicity test data of antitumor drug Z in humanized immune reconstituted mice.
[0041] Human peripheral blood mononuclear cells (PBMCs) were transplanted into NOD / SCID mice to construct a humanized immune reconstitution model, and the antitumor cytotoxicity of the PD-1 inhibitor Z was tested. The experiments included cellular level (CFSE T cell proliferation assay) and animal level (tumor growth curve, flow cytometry detection of T cell subset proportions).
[0042] Raw data acquisition: Correlate cell culture plate number, drug concentration and T cell proliferation rate, and link mouse tumor volume, CD3+ / CD8+ T cell ratio with drug administration time and immune reconstitution cycle.
[0043] Consistency screening: Flow cytometry data were corrected for isotype control and parallel samples with a coefficient of variation >12% were removed; tumor volume measurement was performed by two independent measurements and the mean was taken, and data with a difference >±15% were removed. The overall effective rate was 94.2%.
[0044] Normalization: The T cell proliferation rate in the untreated group was taken as the baseline 0, and the positive control group was taken as the baseline 1. The normalization coefficients for each concentration point were calculated, and the concentration points with response changes >0.15 were retained.
[0045] Model background correction: By comparing with the NOD / SCID immune reconstitution mouse baseline response database, the differences in immune activity at different reconstitution cycles after transplantation were corrected, with a superimposed correction coefficient of 0.041.
[0046] Standardized output: Generate standardized data on drug concentration-tumor inhibition rate. The results show that the tumor growth inhibition rate of drug Z 10mg / kg group was 62.7%, and the proportion of CD8+ effector T cells increased by 2.3 times. The data can be directly used for pharmacodynamic evaluation in drug application.
[0047] The working principle of this invention is as follows: First, by synchronizing the culture plate numbers and time scales in the original experimental records and performing gradient sorting based on drug concentration, data under different time and dosage conditions form a sequence structure with sequential constraints, enhancing the correspondence between experimental data in the time and dosage dimensions. Then, by combining the absorbance readings at multiple wells with the mean deviation within the plate, data deviating from the threshold are screened out, ensuring a consistent signal source and reducing the interference of local abnormal readings on the overall results. Second, logarithmic transformation and interval scaling are introduced in the concentration-response relationship processing to ensure that the responses corresponding to different orders of magnitude of concentrations are consistent. The changes are compressed to a uniform scale range, and response points with stable trends are preserved through differential constraints, improving the continuity and interpretability of the concentration response curve. Then, cross-batch data are extracted by mean difference and uniformly shifted to correct systematic biases between different experimental batches and reduce structural errors caused by batch environmental differences. Finally, the normalized response coefficients are mapped and rearranged with the original concentrations to form a data output structure with a uniform expression and consistent order, which makes the data have higher comparative stability and trend consistency in subsequent analysis, while enhancing the data fusion ability and statistical reliability under different experimental conditions.
[0048] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended technical solutions and their equivalents.
Claims
1. A method for standardizing model data for drug testing, comprising the following steps, characterized in that: S1: Acquisition of raw records for drug experiments: Obtain the cell culture plate number, drug concentration value, culture time scale, and absorbance reading of the enzyme-linked immunosorbent assay (ELISA) reader for the humanized mouse model used in drug testing; arrange the cell culture plate number with the corresponding culture time scale; sort the drug concentration values in a gradient; match the sorted drug concentration values with the absorbance reading of the ELISA reader in the same row to form the experimental record sequence value; S2: Detection signal consistency screening: Based on the experimental record sequence value, obtain the absorbance reading of each well under the same cell culture plate number, calculate the difference between each well reading and the average absorbance reading of the plate, and if the difference exceeds the set absorbance offset threshold, it is removed to obtain the effective absorbance set value. S3: Concentration response normalization processing: Based on the effective absorbance set value, obtain the corresponding drug concentration value and perform logarithmic transformation. Pair the transformed concentration value with the corresponding absorbance reading and calculate the ratio of the absorbance reading at each concentration point to the lowest concentration absorbance reference value. Then, scale each ratio with the highest concentration absorbance reference value to form a normalized sequence in the range of 0 to 1. Perform adjacent difference calculation on the normalized sequence. If the difference result meets the set response change threshold, it is retained to generate the concentration response normalization coefficient. S4: Inter-batch difference correction: Call the concentration response normalization coefficient, obtain the set of normalization coefficients under different experimental batch identifiers, calculate the mean of normalization coefficients for each batch, extract the mean difference between batches, and if this difference exceeds the set batch offset threshold, perform a uniform shift value adjustment on the corresponding batch to obtain the batch correction coefficient value. S5: Standardized data output construction: Based on the batch correction coefficient value, obtain the corresponding concentration response normalization coefficient and perform superposition operation. Combine and map the superposition result with the drug concentration value, rearrange the output according to the cell culture plate number order, and generate standardized pharmacodynamic data values.
2. The method for standardizing model data for drug testing according to claim 1, characterized in that: The experimental record sequence values include cell culture plate number sequence, drug concentration gradient value, and absorbance matching data; the effective absorbance set values include well absorbance data, mean deviation screening value, and effective reading subset; the concentration response normalization coefficient includes logarithmic concentration value, normalized ratio sequence, and response difference characteristic value; the batch correction coefficient values include batch mean parameter, batch difference offset, and translation adjustment coefficient; the standardized pharmacodynamic data values include corrected normalized data, concentration mapping data, and sorted output results.
3. The method for standardizing model data for drug testing according to claim 1, characterized in that: In the S1 step of the drug experiment raw record collection, multi-dimensional physiological and pathological raw data of the in vivo drug efficacy test of the humanized mouse model corresponding to the drug test are collected simultaneously. The data includes blood biochemical indicators, urine metabolic indicators, weight change value, blood glucose dynamic value, behavioral test score, imaging parameters, semi-quantitative results of pathological staining, proportion of cell subsets by flow cytometry, and gray value of protein expression by Western blot. The cell culture plate number corresponds to the mouse individual number, the drug concentration value corresponds to the drug dosage, the culture time scale corresponds to the drug intervention duration, and the multidimensional physiological and pathological raw data corresponds to the absorbance reading of the enzyme-linked immunosorbent assay (ELISA) reader. The individual mouse numbers were arranged with their corresponding drug administration durations, and the drug doses were sorted in a gradient. The sorted drug doses were matched with the original in vivo detection data from peers, and together with the data collected at the cellular level, a complete experimental record sequence value was formed.
4. The method for standardizing model data for drug testing according to claim 3, characterized in that: In the S2 step of signal consistency screening, for the original physiological and pathological data of the body contained in the experimental record sequence value, for the parallel repeated detection data of the same mouse under the same drug intervention duration, the absorbance reading of the multi-well position under the same cell culture plate number is used to calculate the relative deviation between the repeated detection value and the mean of the data group. If the relative deviation exceeds the set parallel sample deviation threshold, the data group is removed. For data from parallel experiments of the same batch of models, the mean deviation between groups is calculated. If the deviation exceeds the set operational deviation threshold, invalid data is removed. For semi-quantitative pathological staining data, the mean is read by two people in a double-blind manner, and data with a reading difference exceeding the set subjective deviation threshold is removed. Finally, the selected in vivo effective data and cellular effective data together constitute the effective absorbance aggregate value.
5. The method for standardizing model data for drug testing according to claim 4, characterized in that: Before the concentration response normalization process in step S3, a baseline correction step for individual humanized mouse models is added: Before drug administration, baseline values of basic physiological indicators of each mouse are collected, including baseline blood glucose, baseline body weight, baseline target protein expression level, and baseline behavioral score. The baseline values correspond to the in vivo scene baseline of the lowest concentration absorbance benchmark value. After obtaining the effective absorbance aggregate value, the in vivo detection value corresponding to each drug administration intervention duration is subtracted from the corresponding mouse's individual baseline value to obtain the baseline-corrected detection value. Then, the concentration response normalization process is performed together with the effective data at the cellular level to finally generate the concentration response normalization coefficient.
6. The method for standardizing model data for drug testing according to claim 5, characterized in that: The S4 step of batch-to-batch difference correction also includes additional processing for humanized mouse model strains and construction batch difference correction: a baseline response database of mouse models with different construction routes and genetic backgrounds that conform to industry standards is established in advance, and the baseline threshold of the baseline response database corresponds to the extended limitation of the batch offset threshold; after obtaining the set of normalized coefficients under different experimental batch identifiers, the normalized coefficients of each batch are first compared with the baseline response coefficients of the corresponding models to calculate the model background deviation value. If the deviation value exceeds the set model background threshold, the model background correction coefficient is superimposed on the translation value adjustment process, and the batch correction coefficient value is finally output.
7. The method for standardizing model data for drug testing according to claim 1, characterized in that: The S5 step of standardized data output construction also includes the integrated and standardized output of multi-omics data: synchronously collecting raw proteomic and metabolomic data of humanized mice after drug intervention, performing outlier removal, abundance normalization, and batch correction on the proteomic data, and performing peak alignment, baseline correction, and relative quantification normalization on the metabolomic data; mapping and matching the standardized proteomic and metabolomic data with the standardized pharmacodynamic data values one-to-one based on mouse individual number / cell culture plate number, drug concentration / dose, culture time scale / drug intervention duration, and rearranging them according to the cell culture plate number order to generate standardized pharmacodynamic data values containing pharmacodynamic, proteomic, and metabolomic dimensions.
8. The method for standardizing model data for drug testing according to claim 1, characterized in that: After the standardized data output is constructed in step S5, an additional step of standardized data verification and full-process traceability is added: the generated standardized drug efficacy data values are cross-validated with the corresponding drug testing model verification data. The model verification data includes electrophysiological recording results, lesion pathological feature verification results, and disease-related signaling pathway activation level verification results. Simultaneously, based on the cell culture plate number, mouse individual number, and experimental batch identifier, a full-process data traceability chain is established to fully record model construction parameters, drug administration parameters, detection equipment parameters, and experimental operator information. Finally, based on standardized efficacy data values, a traceable standardized data report that meets the requirements for drug research and development registration is generated.
9. A model data standardization processing system for drug testing, utilizing the model data standardization processing method for drug testing as described in claims 1-8, comprising a data acquisition module, a signal filtering module, a normalization processing module, a batch calibration module, a data integration and output module, and a verification and traceability module, characterized in that: Data acquisition module: Used to collect raw data from in vitro cell culture plate experiments and in vivo multidimensional physiological and pathological raw data of humanized mouse models used for drug testing. The raw data includes absorbance readings from an ELISA reader, blood biochemical indicators, urine metabolic indicators, weight changes, blood glucose dynamics, behavioral test scores, imaging parameters, semi-quantitative results of pathological staining, cell subset ratios from flow cytometry, and gray values from Western blotting. The module associates cell culture plate number, mouse individual number, drug concentration / dosage, and culture time scale / intervention duration to form an experimental record sequence. Signal filtering module: used to perform consistency verification on the collected experimental record sequences and remove invalid data that exceeds the set deviation threshold, including data with abnormal well absorbance, excessive parallel sample deviation, abnormal batch operation deviation, and abnormal subjective interpretation difference. Generate a valid dataset of in vitro and in vivo data; Normalization module: Used to perform baseline correction, logarithmic concentration conversion, concentration response ratio calculation and adjacent difference screening on valid datasets, and generate concentration response normalization coefficients; at the same time, it performs outlier removal, abundance / baseline normalization and batch correction on proteomics and metabolomics data. Batch correction module: used to calculate the difference in mean normalization coefficients between different experimental batches, and combined with the humanized mouse model construction route and genetic background baseline response database, to perform translation adjustment and model background correction, and generate batch correction coefficients; Data integration and output module: This module maps and matches the batch-corrected concentration response normalization coefficients with standardized proteomics and metabolomics data, and outputs the data in sorted order by cell culture plate number, mouse individual number, drug concentration / dosage, and intervention duration, forming standardized pharmacodynamic data values that include pharmacodynamic, proteomics, and metabolomics dimensions. The verification and traceability module is used to cross-validate standardized pharmacodynamic data with model validation data, establish a full-process data traceability chain, record model construction, dosing parameters, testing equipment and experimental operation information, and generate traceable standardized data reports that meet the requirements for drug research and development registration.