Automatic data calculation method and system for hERG margin

Abstract

The invention belongs to the technical field of drug safety pharmacology, and particularly relates to an hERG margin automatic data calculation method and system. The method comprises the following steps: step 1, calculating the statistical magnitude of hERG IC50 of a drug to be analyzed, wherein the statistical magnitude comprises a sample size, an arithmetic mean value, a lower limit and an upper limit of a 95% confidence interval, and a standard deviation; step 2, taking the statistics obtained in the step 1 as basic data, generating probability distribution of key parameters through Monte Carlo simulation, and calculating distribution of safety windows; and step 3, integrating data of all drugs to be analyzed, and obtaining overall safety window estimation through meta-analysis. The invention further provides a system for implementing the method. The method has the advantages of high efficiency, accuracy and accurate personalized calculation for special drugs, provides more scientific and reliable technical support for hERG safety window calculation and risk assessment, and has a good application prospect.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of drug safety pharmacology, and specifically relates to an hERG margin automatic data calculation method and system. BACKGROUND

[0002] In the field of drug safety pharmacology, cardiac safety assessment has always been a core link of drug discovery and development, especially the prediction of QT interval prolongation and torsade de pointes (TdP) arrhythmia risk. The development of this technical field started more than 20 years ago, with the hERG (human ether-à-go-go related gene) potassium channel being identified as a key target for drug-induced cardiotoxicity, and the hERG assay gradually becoming an industry standard. Early practices mainly relied on simple safety window calculations, such as the ratio of hERG IC 50 to clinical exposure concentration, but lacked standardized methods, resulting in high variability of cross-laboratory data and limited prediction accuracy. With the publication of International Conference on Harmonization (ICH) guidelines (such as ICH S7B and E14), the industry began to shift towards more systematic methods, emphasizing the reliability and clinical relevance of hERG potency assays. In recent years, the ICH E14 / S7B Q&A and training materials further promoted the popularization of best practices, introducing a distribution-based quantitative risk assessment framework, supporting hERG margin as an important part of integrated QTc risk assessment, reflecting the shift from simple torsade de pointes arrhythmia risk prediction to comprehensive QTc prolongation assessment, and highlighting the importance of data integration, statistical rigor, and interdisciplinary collaboration.

[0003] Related key technologies include the quantitative calculation method of hERG margin and the innovation of integrated risk assessment framework. The calculation of hERG margin is no longer dependent on simple arithmetic mean, but through processing the uncertainty of input parameters to improve prediction accuracy. Specifically, this method simulates hERG IC 50 , plasma concentration and protein binding fraction based on log-normal distribution, generates a large random distribution (such as 10,000 values), and derives the margin distribution through the formula Margin = hERG IC 50 / (Concentration × Unbound Fraction). This distribution method allows the inclusion of variability, generates confidence and prediction intervals using mean and standard deviation, and aggregates overall margin through random effects meta-analysis, thereby providing more robust and reliable risk thresholds.

[0004] However, current methods for calculating hERG margins still face numerous challenges in practical application. On one hand, authoritative literature (such as ICH E14 / S7B training materials) only describes theoretical methods, requiring manual or semi-automated data processing in practice. This leads to problems with standardization, automation, reproducibility, and result presentation. On the other hand, existing methods exhibit insufficient adaptability and computational efficiency when dealing with the complex kinetic characteristics of specific drugs (such as ondansetron) or integrating cross-laboratory data. Specifically, drugs like ondansetron may exhibit unique hERG inhibitory properties, and traditional simple calculations may fail to accurately capture their risk profile. Existing technologies lack optimized calculation modules for such specific situations, hindering flexible and accurate personalized analysis. Therefore, the field continues to develop an efficient, flexible, and accurate method for calculating hERG margins. Summary of the Invention

[0005] To address the aforementioned problems in the prior art, this invention provides an automated hERG margin data calculation method and system.

[0006] An automated hERG margin data calculation method includes the following steps: Step 1: Calculate the hERG IC50 of the drug to be analyzed. 50 The statistical measures include sample size, arithmetic mean, lower and upper limits of 95% confidence interval, and standard deviation; Step 2: Using the statistics obtained in Step 1 as the basic data, generate the probability distribution of key parameters through Monte Carlo simulation, and calculate the distribution of the safety window; Step 3: Integrate the data of all drugs to be analyzed and obtain an overall safety window estimate through meta-analysis.

[0007] Preferably, in step 1, for the drug ondansetron, the geometric mean method is also used to process the logarithmic scale data to better reflect its distribution characteristics.

[0008] Preferably, in step 1, the formula for calculating the standard deviation is: SD = (log10(UL) - log10(LL)) / 3.92 Where SD is the standard deviation, UL is the upper limit of the 95% confidence interval, and LL is the lower limit of the 95% confidence interval.

[0009] Preferably, in step 2, the specific steps for generating the probability distribution of key parameters through Monte Carlo simulation include: setting a random seed to ensure reproducibility of the results, and then simulating 10,000 random samples: assuming that the plasma protein binding rate (PPB) follows a normal distribution, and that the drug concentration follows a normal distribution on a logarithmic scale, then converting to the original scale; the free concentration is calculated by multiplying the total concentration by (1 - PPB), representing the pharmacologically active concentration; IC50... 50 The normal distribution is simulated on a logarithmic scale, and the molecular weight is converted to μM units.

[0010] Preferably, in step 2, the formula for calculating the security window is: Safe Margin = (Molecular weight × IC) 50 ) / Free concentration The distribution of the safety window is calculated as follows: calculate the safety window value for 10,000 samples and output the 2.5% quantile, median and 97.5% quantile.

[0011] Preferably, in step 2, visualization functions are integrated to draw histograms and density curves, displaying PPB, concentration, free concentration, and IC50. 50 Add quantile vertical lines to the distribution of the safety window.

[0012] Preferably, step 3 specifically includes the following steps: Input the safety window statistic of the drug to be analyzed and convert it to a logarithmic scale; The logarithmic mean, standard error, and confidence interval of the overall safety window were calculated using a random effects meta-analysis model. The logarithmic result is inversely transformed back to the original scale to obtain an estimate of the overall safety window.

[0013] Preferred options also include: Step 4: Using visualization techniques, compare the distribution of different drugs and the overall safety window to support decision analysis. Specifically, this includes: calculating the density value of each drug to be analyzed and the overall safety window based on the log-normal distribution model, and using the ggplot2 package to plot the density curve on the logarithmic scale.

[0014] The present invention also provides a system for implementing the above-described automated data calculation method for hERG margin, comprising: The statistics calculation module is configured to calculate the hERG IC of the drug to be analyzed. 50 The statistical measures include sample size, arithmetic mean, lower and upper limits of 95% confidence interval, and standard deviation; The safety window calculation module is configured to use the statistics obtained by the statistics calculation module as the basic data, generate the probability distribution of key parameters through Monte Carlo simulation, and calculate the distribution of the safety window. The overall safety window calculation module is configured to integrate data from all drugs to be analyzed and derive an overall safety window estimate through meta-analysis.

[0015] The present invention also provides a computer-readable storage medium having stored thereon a computer program for implementing the above-described automated data calculation method for hERG margin.

[0016] This invention provides an automated data calculation method and system for hERG margin, which significantly improves the scientificity and reliability of hERG safety window calculation through probability distribution models and meta-analysis methods.

[0017] Specifically, the core theoretical innovation of this invention lies in replacing the traditional point estimation method with a probability distribution model. Traditional methods use only a single numerical value for safety assessment, failing to reflect the true variability of parameters. In contrast, this invention uses a probability distribution model to measure plasma protein binding rate (PPB), drug concentration, and IC50. 50 A normal distribution model is established using key parameters, and a large number of random samples are generated using Monte Carlo simulation to more accurately characterize uncertainties in real-world scenarios. Theoretical analysis shows that this distribution-based method can effectively capture extreme cases and reduce evaluation bias caused by parameter variations. Especially in the calculation of the safety factor, the relationship between free concentration and IC50 is effectively utilized. 50 The ratio distribution can provide a complete risk probability distribution, rather than a single threshold judgment.

[0018] Furthermore, this invention features an optimized computational module specifically designed for special drugs (such as ondansetron), enabling flexible and precise personalized analysis.

[0019] In terms of visualization, this invention uses density curves to intuitively display the distribution characteristics of safety coefficients for different drugs. This visualization not only facilitates comparison of the safety characteristics of different drugs but also visually displays overlapping distribution areas, providing important evidence for risk assessment.

[0020] Compared with existing technologies, the theoretical advantages of this invention are reflected in several aspects: First, the modeling method based on log-normal distribution is more consistent with the biological characteristics of drug metabolism parameters and can better handle skewed distribution data. Second, the introduction of meta-analysis methods enables the integration of data from multiple studies, improving the robustness of the evaluation results. The combined distribution curve in the figure shows the overall evaluation results, which have a smoother distribution shape and stronger representativeness. This integrated analysis method provides an effective solution to the problem of insufficient sample size in a single study.

[0021] This invention also demonstrates the statistical superiority of the method through theoretical derivation. The probability distribution calculation of the safety factor is based on a rigorous mathematical formula: Safe Margin = (Molecular Weight × IC) 50 The method uses unbounded concentration, where all parameters are input in distributed form. This approach not only provides point estimates but also complete confidence interval estimates, significantly enhancing the reliability of the evaluation results. Theoretical analysis shows that this method is significantly less sensitive to outliers than traditional methods, exhibiting better robustness.

[0022] In summary, this invention achieves a significant breakthrough in drug safety assessment methods at the theoretical level through innovative probability distribution models and advanced visualization technology. Its distributed assessment framework, intuitive graphical presentation, and rigorous statistical foundation constitute the core advantages of this invention, providing more scientific and reliable technical support for hERG safety window calculation and risk assessment.

[0023] Obviously, based on the above description of the present invention, and according to common technical knowledge and conventional methods in the field, various other modifications, substitutions or alterations can be made without departing from the basic technical concept of the present invention.

[0024] The following detailed embodiments further illustrate the above-described content of the present invention. However, this should not be construed as limiting the scope of the present invention to the following examples. All technologies implemented based on the above-described content of the present invention fall within the scope of the present invention. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating Embodiment 1 of the present invention.

[0026] Figure 2 This is an example diagram showing the data distribution analysis results of the multi-filit safety window in Embodiment 1 of the present invention; Figure 3 This is an example diagram showing the data distribution analysis results of the safety window of moxifloxacin in Embodiment 1 of the present invention; Figure 4 This is an example diagram showing the data distribution analysis results of Ondansetron's security window in Embodiment 1 of the present invention; Figure 5 This is an example diagram showing the calculation results of the overall security window in Embodiment 1 of the present invention. Detailed Implementation

[0027] It should be noted that the algorithms for data acquisition, transmission, storage and processing steps not specifically described in the embodiments, as well as the hardware structures and circuit connections not specifically described, can all be implemented using content already disclosed in the prior art.

[0028] Example 1: hERG margin automated data calculation method and system The system provided in this embodiment is used for automated hERG margin data calculation, including: The statistics calculation module is configured to calculate the hERG IC of the drug to be analyzed. 50 The statistical measures include sample size, arithmetic mean, lower and upper limits of 95% confidence interval, and standard deviation; The safety window calculation module is configured to use the statistics obtained by the statistics calculation module as the basic data, generate the probability distribution of key parameters through Monte Carlo simulation, and calculate the distribution of the safety window. The overall safety window calculation module is configured to integrate data from all drugs to be analyzed and derive an overall safety window estimate through meta-analysis.

[0029] The calculation method using the system in this embodiment is as follows: Figure 1 As shown, this method can be implemented using the R language. Specifically, it includes the following steps: Step 1: Accurately calculate the drug's hERG IC 50 The statistics provide basic data for subsequent safety window calculations.

[0030] IC 50 This is the half-maximal effective concentration (IC50) of a drug that inhibits the hERG channel, and its variability directly affects the reliability of risk assessment. This example illustrates this. The user inputs the IC50 values ​​of three drugs (domfelithe, moxifloxacin, and ondansetron). 50 For experimental data, the system automatically calculates the sample size (n), arithmetic mean (Mean), and the lower bound (LL) and upper bound (UL) of the 95% confidence interval (CI). For Ondansetron, the geometric mean method is also used to process the log-scaled data to better reflect its distribution characteristics. The standard deviation (SD) is calculated using the confidence intervals. The formula is: SD = (log10(UL) - log10(LL)) / 3.92, where 3.92 is the standard multiple of the 95% confidence interval (based on the t-distribution).

[0031] All results are rounded to two decimal places, and tables and summaries are output, including the original data, mean, confidence intervals, log10 mean, and standard deviation. Automated calculations avoid errors from manual calculations, ensuring the accuracy and repeatability of the statistics. Logarithmic transformation addresses the skewed distribution of the data, improving the reliability of subsequent simulations.

[0032] Step 2: Safety window calculation and distribution image drawing. The probability distribution of key parameters is generated through Monte Carlo simulation, the distribution of the safety window is calculated, and the results are visualized to quantify the uncertainty of risk assessment.

[0033] First, a random seed is set to ensure reproducibility of results. Then, 10,000 random samples are simulated: plasma protein binding rate (PPB) is assumed to follow a normal distribution, and drug concentration follows a normal distribution on a logarithmic scale, before being converted back to the original scale; free concentration is calculated by multiplying the total concentration by (1 - PPB), representing the pharmacologically active concentration; IC50... 50 Similarly, a normal distribution is simulated on a logarithmic scale, and molecular weight is converted to μM units.

[0034] The core calculation formula for the security window is: Safe Margin = (Molecular weight × IC) 50 Free concentration Calculate the safety window values ​​for 10,000 samples and output quantiles (2.5%, median, 97.5%) for risk assessment (a lower limit >1 indicates safety). Simultaneously, the technical solution integrates visualization functions, plotting histograms and density curves to display PPB, concentration, free concentration, and IC50. 50 The distribution of the safety window is simulated, and quantile vertical lines are added. All graphs are integrated into a single page with a 2x3 layout, including magnified views of the safety window (1%-99% range). Professional color schemes (such as light blue and red) and labels are used to ensure scientific accuracy and aesthetic appeal. Results are saved as a CSV file containing parameter summaries. This step, through distribution simulation and visualization, overcomes the limitations of point estimation, making risk assessment more comprehensive and intuitive.

[0035] Step 3: Calculate the overall safety window value.

[0036] This step aims to integrate data from multiple drugs and derive an overall safety window estimate through meta-analysis, thereby improving the robustness of the assessment. The technical approach first inputs safety window statistics (such as mean and confidence interval) for three drugs and converts them to a logarithmic scale (such as log10 mean). Then, a random-effects meta-analysis model (using the metamean function in R) is used to calculate the logarithmic mean, standard error, and confidence interval of the overall safety window. The meta-analysis, based on logarithmic transformation and standard deviation weighting of the drug data, handles the variability between different drugs, ensuring more representative results. Finally, the logarithmic results are inversely transformed back to the original scale to obtain an estimate of the overall safety window (such as the median and 95% confidence interval). This step, through data integration and statistical modeling, addresses the problem of small sample sizes for single drugs and provides a more reliable framework for overall laboratory risk assessment.

[0037] Step 4: Overall Safety Window Image Plotting. This step uses advanced visualization techniques to compare the distribution of different drugs and the overall safety window, supporting decision analysis. The technical solution is based on a log-normal distribution model, calculating the density values ​​of each drug (dofetilide, ondansetron, moxifloxacin) and the overall safety window, and using the ggplot2 package to plot density curves on a logarithmic scale. In the image, each drug is distinguished by different colors and line styles, the coordinate axes are set to a logarithmic scale (e.g., 0.5 to 300), and the legend, labels, and themes are optimized (using Times New Roman font) to ensure the image meets publication standards. Finally, the graph is saved in PNG format for reporting or publication. The visualization output of this step clearly shows the distribution differences and overall trends among drugs; for example, the Combined curve represents the summary of laboratory data, enhancing the interpretability and dissemination of the results.

[0038] Figures 2-5 An example of the calculation results output in this embodiment is provided.

[0039] Figure 2 The figures are as follows: Figure 1: Plasma Protein Binding (PPB), showing the distribution of dofetilide protein binding in human plasma. Figure 2: Concentration (logscale), showing the logarithmic distribution of total plasma critical concentrations associated with clinical QTc assessment. Figure 3: Unbound Concentration, calculated from total clinical plasma concentrations and plasma protein binding, representing the concentration of pharmacologically active free drug. Figure 4: IC. 50 hERG channel half-maximal inhibitory concentration / hERG IC50 50 This demonstrates the IC50 of dofetilide as measured in in vitro hERG experiments. 50 The distribution of values. Figure Safety Margin: hERG Safety Window / hERG Safety Margin, showing the distribution of values ​​by hERG IC. 50The overall distribution of the safety window value calculated by dividing by the clinical free plasma concentration. The green dashed line represents a specific risk assessment threshold. Figure Safety Margin (Zoomed 1%-99%): Zoomed View of Safety Margin Distribution (1st - 99th Percentile) is a zoomed-in view of Figure Safety Margin, removing the influence of the extreme values ​​at both ends of the distribution (0.5% each) to more clearly show the shape of the main range of the safety window.

[0040] Figure 3 The figures are as follows: Figure 1: Plasma Protein Binding (PPB), showing the distribution of moxifloxacin protein binding in human plasma. Figure 2: Concentration (logscale), showing the logarithmic distribution of total plasma critical concentrations associated with clinical QTc assessment. Figure 3: Unbound Concentration, calculated from total clinical plasma concentrations and plasma protein binding, representing the concentration of pharmacologically active free drug. Figure 4: IC. 50 hERG channel half-maximal inhibitory concentration / hERG IC50 50 This demonstrates the IC50 of moxifloxacin as measured in an in vitro hERG assay. 50 The distribution of values. Figure Safety Margin: hERG Safety Window / hERG Safety Margin, showing the distribution of values ​​by hERG IC. 50 The overall distribution of the safety window value calculated by dividing by the clinical free plasma concentration. The green dashed line represents a specific risk assessment threshold. Figure Safety Margin (Zoomed 1%-99%): Zoomed View of Safety Margin Distribution (1st - 99th Percentile) is a zoomed-in view of Figure Safety Margin, removing the influence of the extreme values ​​at both ends of the distribution (0.5% each) to more clearly show the shape of the main range of the safety window.

[0041] Figure 4The figures are as follows: Figure 1: Plasma Protein Binding (PPB), showing the distribution of ondansetron protein binding in human plasma. Figure 2: Concentration (logscale), showing the logarithmic distribution of total plasma critical concentrations relevant to clinical QTc assessment. Figure 3: Unbound Concentration, calculated from total clinical plasma concentrations and plasma protein binding, representing the concentration of pharmacologically active free drug. Figure 4: IC. 50 hERG channel half-maximal inhibitory concentration / hERG IC50 50 This demonstrates the IC50 of ondansetron as measured in in vitro hERG assays. 50 The distribution of values. Figure Safety Margin: hERG Safety Window / hERG Safety Margin, showing the distribution of values ​​by hERG IC. 50 The overall distribution of the safety window value calculated by dividing by the clinical free plasma concentration. The green dashed line represents a specific risk assessment threshold. Figure Safety Margin (Zoomed 1%-99%): Zoomed View of Safety Margin Distribution (1st - 99th Percentile) is a zoomed-in view of Figure Safety Margin, removing the influence of the extreme values ​​at both ends of the distribution (0.5% each) to more clearly show the shape of the main range of the safety window.

[0042] Figure 5 The figure illustrates the overall data distribution of the hERG safety window assessment system in this embodiment, which is based on the comprehensive risk assessment framework proposed by Leishman et al. (2024). The graphs compare the hERG safety window distributions of three well-defined positive control drugs—dofetilide (a potent hERG blocker), moxifloxacin (a moderate hERG blocker, a clinical positive control), and ondansetron (a weak hERG blocker, known for its low safety window)—using kernel density estimation curves. All safety windows were calculated using free plasma concentrations associated with high clinical exposure and 10 msQTc prolongation, according to the ICH E14 / S7B guidelines.

[0043] Overall image display: Dofetilide has the most concentrated distribution and the narrowest safety window. Moxifloxacin has a wider distribution range, with a safety window between that of dofetilide and ondansetron. Ondansetron exhibits a unique and relatively high safety window, but its absolute IC50 value is low. 50 The value is low.

[0044] The "Combined" curve represents the aggregated distribution of all data, and its shape indicates that the method and system of this embodiment can clearly distinguish and characterize compound categories with different hERG inhibition risks. This visualizes the overall hERG safety window.

Claims

1. A method for hERG margin automated data calculation, characterized in that, The method comprises the following steps: Step 1, calculate the hERG IC of the drug to be analyzed 50 the statistics including sample size, arithmetic mean, lower and upper limits of the 95% confidence interval, standard deviation; Step 2, based on the statistical quantity obtained in step 1, the probability distribution of the key parameters is generated by Monte Carlo simulation, and the distribution of the safety window is calculated; Step 3, the data of all drugs to be analyzed are integrated, and the overall safety window estimate is obtained by meta-analysis.

2. The hERG margin automated data calculation method of claim 1, wherein, In step 1, for the drug ondansetron, the geometric mean method is also used to process the log scale data to better conform to its distribution characteristics.

3. The hERG margin automated data calculation method of claim 1, wherein, In step 1, the calculation formula of the standard deviation is: SD = (log10(UL) - log10(LL)) / 3.92 Wherein, SD is the standard deviation, UL is the upper limit of the 95% confidence interval, and LL is the lower limit of the 95% confidence interval.

4. The hERG margin automated data calculation method of claim 1, wherein, In Step 2, the specific steps to generate the probability distribution of the key parameters by Monte Carlo simulation include: setting a random seed to ensure the results are reproducible, and then simulating 10,000 random samples: the plasma protein binding rate PPB is assumed to follow a normal distribution, and the drug concentration follows a normal distribution on a logarithmic scale, which is then converted to the original scale; the free concentration, representing the pharmacologically active concentration, is calculated from the total concentration multiplied by (1 - PPB); IC 50 Simulate a normal distribution on a logarithmic scale and convert to μΜ units considering the molecular weight.

5. The hERG margin automated data calculation method of claim 1, wherein, In step 2, the calculation formula of the safety window is: Safe Margin = (molecular weight x IC 50 ) / free concentration The distribution calculation method of the safety window is: the safety window values of 10000 samples are calculated, and the 2.5% quantile, the median and the 97.5% quantile are output.

6. The hERG margin automated data calculation method of claim 5, wherein, In Step 2, integrate visualization functions to plot histograms and density curves for PPB, concentration, free concentration, IC 50 and safety windows, add quantile vertical lines.

7. The hERG margin automated data calculation method of claim 1, wherein, In step 3, it specifically comprises the following steps: Input the safety window statistics of the drug to be analyzed and convert it to log scale; The random effect meta-analysis model is used to calculate the log mean, standard error and confidence interval of the overall safety window; The log result is converted back to the original scale to obtain the estimate value of the overall safety window.

8. The hERG margin automated data calculation method of claim 1, wherein, It also includes: Step 4, by using visualization technology, the safety window distribution of different drugs and the overall safety window is compared to support decision analysis, which specifically includes: based on the lognormal distribution model, the density value of each drug to be analyzed and the overall safety window is calculated, and the density curve on the log scale is drawn by using the ggplot2 package.

9. A system for implementing the hERG margin automated data calculation method of any of claims 1-8, characterized by, It includes: The statistics calculation module is configured to calculate the hERG IC of the drug to be analyzed. 50 The statistical measures include sample size, arithmetic mean, lower and upper limits of 95% confidence interval, and standard deviation; The safety window calculation module is configured to take the statistical quantity obtained by the statistical quantity calculation module as the basis data, generate the probability distribution of the key parameters by Monte Carlo simulation, and calculate the distribution of the safety window; The overall safety window calculation module is configured to integrate the data of all drugs to be analyzed, and obtain the overall safety window estimate by meta-analysis.

10. A computer-readable storage medium, characterized in that, Stored thereon are: a computer program for implementing the hERG margin automatic data calculation method of any one of claims 1-8.

Images