A drug chromatogram signal structural analysis method based on an intelligent sensing system

Abstract

The application relates to the technical field of medicine analysis and detection and chromatographic signal processing, and discloses a medicine chromatographic signal structured analysis method based on an intelligent sensing system, which reads original time-response sequences of reference samples and test samples under the same detection method and method version, establishes a medicine chromatographic detection sequence and a peak level reference relation set, forms a candidate corresponding relation according to a peak sequence and a peak interval, determines a stable reference peak from known component peaks and calculates a common drift, deducts the common drift from changes of a candidate target peak relative to a reference target peak, respectively determines states and use permissions of a retention time, a peak interval, a peak width, a peak sequence and a peak boundary relation, re-estimates a baseline, determines a peak boundary and a peak area for a local interval affected by a limited state or a disabled state, and writes a processing process and a result into a structured analysis record. The application can distinguish common drift of a chromatographic signal from local changes of a target peak, and reduces influences of invalid reference relations on peak positioning and integral results.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical analysis and detection and chromatographic signal processing technology, and in particular to a structured analysis method for pharmaceutical chromatographic signals based on an intelligent sensing system. Background Technology

[0002] Drug chromatographic detection involves separating different components in a test sample using a liquid chromatograph or gas chromatograph, and generating a response signal that changes with retention time using a detector. Chromatographic data processing typically uses a reference chromatogram as a reference. Based on the retention time of the target component or a preset retention time window, the corresponding target peak is determined in the test sample chromatogram. Then, baseline processing, peak boundary determination, and peak area integration are used to obtain the data required for component qualitative analysis and content calculation. When continuously detecting multiple test samples, changes in the separation state of the chromatographic column, sample matrix residue, mobile phase, or carrier gas state may cause retention time shifts, peak broadening, or peak boundary changes in chromatographic peaks. Therefore, it is necessary to identify and analyze the target peak based on the current chromatographic signal.

[0003] Existing technologies have proposed various processing methods for target peak identification, calibration, and integration based on reference chromatographic peaks. For example, WO2014149978A1 discloses a chromatographic system calibration method that establishes analyte time intervals based on the retention times of analytes in standard samples. After completing a predetermined number of injections, after a predetermined time has elapsed, or after changes in chromatographic parameters, the standard sample is re-injected, and the corresponding analyte time interval is recalculated. When the retention time of the test sample chromatographic peak falls within the corresponding analyte time interval, the component identity of the chromatographic peak is determined. US20220155269A1 discloses an automatic peak integration method, which determines the initial value of the analyte retention time based on the retention time of chemically related substances and a preset offset or scaling factor, and simultaneously determines the initial value of the analyte peak shape parameters based on the peak shape parameters of chemically related substances. The analyte chromatographic peak is then fitted with the initial values ​​to determine the analyte retention time, peak area, and peak shape. US20240011954A1 discloses a gas chromatography system with a diagnostic and prediction module, which compares the chromatographic parameters such as retention time, peak area, or peak shape obtained from sample chromatographic separation with simulated chromatographic separation or reference chromatographic separation, and judges or predicts chromatographic performance abnormalities based on performance control limits, thereby outputting corresponding maintenance information.

[0004] The aforementioned technologies can respectively complete target peak identification, peak integration, chromatographic system calibration, or performance monitoring through retention time intervals, relevant substance peak parameters, reference chromatograms, or performance control limits. However, in continuous drug chromatographic detection, multiple chromatographic peaks may advance or delay together due to changes in the state of the chromatographic system. At the same time, some target peaks may also be affected by sample matrix, local overlap, or peak shape changes, resulting in additional shifts relative to other chromatographic peaks. Existing processing methods are difficult to separate the common changes of the entire chromatogram from multiple stable chromatographic peaks and then determine the local deviation of the target peak relative to this common change. When only some of the reference relationships such as retention time, peak order, peak interval, peak width, and peak boundary in the reference standard change, it is also difficult to restrict the failure relationships from participating in target peak localization and integration. Therefore, continuing to use the original reference relationships may cause errors in target peak matching or peak boundary determination, while overall recalibration or reintegration may change the results of other chromatographic peaks unrelated to the abnormal target peak. Summary of the Invention

[0005] The technical problem this invention aims to solve is that, during continuous drug chromatographic detection, when the common drift of multiple chromatographic peaks and the local change of the target peak coexist, existing processing methods struggle to determine whether the reference relationships of each peak level in the reference standard are still applicable to the current test sample. Furthermore, they are prone to continuing to fix the target peak position and boundary after the local relationship fails, or re-integrating the entire chromatogram. To address this, we propose a structured analysis method for drug chromatographic signals based on an intelligent sensing system.

[0006] To achieve the above objectives, this application adopts the following technical solution: a method for structured analysis of drug chromatographic signals based on an intelligent sensing system, comprising the following steps:

[0007] Step 1: Read the raw time-response sequences of the reference standard and the test sample under the same detection method and method version, and establish the drug chromatographic detection sequence;

[0008] Step 2: Analyze the reference standard, extract the retention time, peak order, peak interval, peak width and peak boundary of the target peak, and establish a peak-level reference relationship set;

[0009] Step 3: Based on the peak sequence and peak interval, form a candidate correspondence in the current test sample, determine at least two stable reference peaks from the known component peaks, and determine the common drift based on their retention time changes;

[0010] Step 4: Compare the changes of the candidate target peak relative to the control target peak with the common drift to obtain the retention time relationship residual, peak interval relationship residual, and peak width relationship residual, and determine the peak order and peak boundary change results;

[0011] Step 5: Based on the residuals and changes in peak-level relationships, determine the status of each peak-level reference relationship as valid, restricted, disabled, or pending confirmation, and determine its usage rights in candidate peak location, peak boundary determination, and peak area integration.

[0012] Step 6: When all peak-level reference relationships are valid, determine the final peak boundary and peak area based on the peak-level reference relationships of the common drift and valid states. When there are restricted or deactivated states, determine the local re-analysis interval, lock the baseline and peak parameters outside the interval, re-estimate the local baseline within the interval and generate peak boundary candidate results, exclude candidate results that depend on the deactivated state, determine the final peak boundary and peak area based on the peak-level reference relationships of the valid states, and write them into the structured analysis record.

[0013] Preferably, when establishing the drug chromatographic detection sequence, the original time-response sequence of each injection is associated with the detection method identifier, method version, sample name, sample type, injection number, injection time, and data file identifier; each injection retains its own independent chromatographic running time axis, and the original time-response sequences of different injections are not spliced ​​together end to end;

[0014] Only data obtained using the same detection method and the same method version should be included in the same drug chromatographic detection sequence, and the sample types should include at least reference standards and test samples.

[0015] Preferably, the retention time relationship records the initial peak time of the target peak in the reference standard; the peak sequence relationship records the peak emergence order between the target peak and the known peaks before and after it; the peak interval relationship records the relative time distance between the target peak and one or more reference peaks; the peak width relationship records the baseline peak width or half-peak width of the target peak; and the peak boundary relationship records the position of the peak start point and peak end point relative to the left and right valley points, adjacent peak boundaries, and local baselines.

[0016] Preferably, when establishing candidate correspondences, known peaks in the reference standard are used as candidate objects, and candidate corresponding peaks are generated in the current test sample;

[0017] A candidate correspondence is established when the elution order of a candidate corresponding peak is consistent with that of at least one adjacent known peak with which a candidate correspondence has been established, and the difference between the relative time interval between the two peaks in the current test sample and the corresponding relative time interval in the control sample is within a preset initial range.

[0018] The preset initial range is determined based on the system applicability requirements of the corresponding detection method, method validation data, or repeated injection results of qualified reference standards. When one reference standard peak corresponds to multiple candidate corresponding peaks of the test sample, or multiple reference standard peaks correspond to the same candidate corresponding peak of the test sample, the corresponding candidate corresponding peaks are not included in the stable reference peak set.

[0019] Preferably, the stable reference peak has a clear peak in both the reference standard and the current test sample, and the change in the peak position under different local baseline support point conditions is within the allowable range of the stable reference peak position, and there is no change in peak order or serious overlap.

[0020] Let the retention time of the i-th stable reference peak in the reference standard be . The retention time of the corresponding peak in the k-th sample injection is The change in retention time of the stable reference peak. for:

[0021]

[0022] When the number of stable reference peaks (mm) is no less than three, the common drift is... for:

[0023]

[0024] In the formula, m is the number of stable reference peaks involved in the calculation. When there are only two stable reference peaks, the difference in their retention time changes is compared. If the difference does not exceed the allowable value for common drift consistency, the common drift is determined by the average value of their retention time changes. If the difference exceeds the allowable value for common drift consistency, the common drift is not determined, and the corresponding peak-level reference relationship is determined to be in a state of pending confirmation. The allowable values ​​for stable reference peak positions and common drift consistency are determined based on method validation data or repeated injection results of qualified reference standards.

[0025] Preferably, the retention time of the target peak of the reference standard is set as follows: The retention time of the candidate target peak in the current test sample is Then retain the time relation residuals. for:

[0026]

[0027] Let the retention times of the nearest stable reference peak in the reference standard and the current test sample be respectively. and Then the peak interval relationship residual for:

[0028]

[0029] Let the peak widths of the target peak in the reference standard and the current test sample be respectively... , Then the peak width relationship residual for:

[0030] .

[0031] Preferably, for the h-th type of peak-level reference relationship that can be expressed numerically, let its peak-level relationship residual be... :

[0032] The first allowed value is The second allowed value is ,and < .

[0033] when When this happens, the corresponding peak-level reference relationship is determined to be in a valid state;

[0034] when When this occurs, it is determined to be in a restricted state;

[0035] when When the peak reference relationship is insufficient, it is determined to be in a disabled state; when the verification data of the corresponding peak reference relationship is insufficient, it is determined to be in a pending confirmation state.

[0036] Preferably, when the peak order change results in the peak order being maintained, the peak order relationship is determined to be a valid state;

[0037] When the peak sequence cannot be reliably confirmed, the peak sequence relationship is determined to be in a restricted state; when the peak sequence changes and conflicts with the known peak relationship of the target component, the peak sequence relationship is determined to be in a deactivated state; the retention time relationship in an effective state is used for candidate position ranking, the peak sequence relationship and peak interval relationship in an effective state are used to determine candidate intervals, and the peak width relationship and peak boundary relationship in an effective state are used to evaluate peak boundary candidate results;

[0038] Peak-level reference relationships in a restricted state are only used to expand the search range of candidate peaks or to assist in the evaluation of candidate results. Peak-level reference relationships in a deactivated state do not participate in peak location, peak boundary determination, or peak area integration.

[0039] The peak width relationship residual is not used as the sole basis for determining the peak width relationship to be in a disabled state. When the peak width relationship residual exceeds the corresponding second allowable value, and there is a change in the peak boundary, an abnormal peak spacing relationship, or an increase in the degree of overlap with adjacent peaks, the peak width relationship is determined to be in a disabled state.

[0040] Preferably, the local reanalysis interval is determined based on the relationship between adjacent peaks that are still in a valid state, stable valley points, peak boundaries of adjacent confirmed peaks, and the continuous range of local baselines;

[0041] When the target peak is located between two stable reference peaks, the time range between the right boundary of the previous stable reference peak and the left boundary of the next stable reference peak is defined as the local reanalysis interval.

[0042] Within the local reanalysis interval, local baseline support points are selected from the sampling points that have never fallen into the candidate peak interval, and the local baseline support points located on both sides of the candidate peak are linearly connected to form local baseline candidate results.

[0043] Peak boundary candidate results are generated based on the local maximum response location, adjacent valley points, and the intersection location of the chromatographic signal and the local baseline of the response signal after removing the local baseline.

[0044] Let the original detection response of the q-th sampling point be... The corresponding local baseline response is The candidate peak start point and peak end point correspond to the sampling points respectively. and The time interval between adjacent sampling points is Then the candidate peak area for:

[0045] .

[0046] Preferably, when there are two or more sets of peak boundary candidate results that simultaneously meet the peak sequence relationship, peak interval relationship and local baseline continuity in a valid state, the peak area change rate of the same peak boundary candidate result under different local baseline candidate results is calculated respectively.

[0047] Let the maximum peak area corresponding to the candidate results of the same peak boundary be . The minimum peak area is Then the peak area change rate ;

[0048] Select the peak boundary candidate with the smallest peak area change rate; when the peak area change rates are the same or their difference is lower than the preset resolution, select the peak boundary candidate with the smaller sum of the time distances between the peak start point and peak end point and the corresponding local valley point. The preset resolution is determined based on the repeated injection results of qualified reference standards or method validation data.

[0049] If a unique peak boundary candidate result cannot be obtained after comparison, the peak area and drug content results of the target peak are not automatically updated; if a unique peak boundary candidate result is obtained, the drug content result directly associated with the target peak is updated according to the corresponding final peak area, and the drug chromatographic detection sequence, stable reference peak, common drift, peak level relationship residual, current status of each peak level reference relationship, local reanalysis interval, final peak parameters and drug content results are written into the structured analysis record.

[0050] The technical effects and advantages of this invention are as follows:

[0051] This invention first uses multiple stable reference peaks to determine the common drift of the current chromatographic signal, and then judges whether the reference relationships of each peak level are still applicable based on the remaining changes of the candidate target peak relative to the common drift. This can reduce the situation where multiple peaks are mistakenly advanced or delayed as local anomalies of the target peak. By setting the current state of retention time, peak order, peak interval, peak width, and peak boundary relationship and restricting their usage, the still valid reference information can be retained, and the invalid relationship can be prevented from continuing to fix the target peak position and boundary. When the local relationship changes, only the affected interval is re-analyzed and the peak parameters outside the interval are locked, which can narrow the reprocessing range. The common drift, relationship residuals, state changes, and the basis for result formation are saved through structured analysis records. Attached Figure Description

[0052] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts:

[0053] Figure 1 This is a schematic diagram of the drug chromatographic signal structure analysis method of the present invention;

[0054] Figure 2 This is a schematic diagram of the peak-level reference relationship state judgment and local re-analysis logic of the present invention. Detailed Implementation

[0055] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0056] Example 1

[0057] This embodiment is used to perform structured analysis of the time-response signal formed during the detection of pharmaceuticals by liquid chromatography or gas chromatography. When performing liquid chromatography detection, a high-performance liquid chromatograph including a pump, autosampler, column oven and ultraviolet detector can be used. The mobile phase composition, flow rate, injection volume, column temperature and detection wavelength are set according to the detection method corresponding to the drug to be tested. After the system is balanced, blank sample, reference standard and multiple test samples are run in sequence.

[0058] Under certain implementation conditions, the baseline response is continuously observed for ten minutes during instrument equilibration. When the baseline fluctuation is no higher than 0.5 mAU, the detection sequence is started. For gas chromatography detection, an Agilent 7820A gas chromatograph can be used. Select either an FID detector or an ECD detector according to the analyte, and select either direct injection or headspace injection mode according to the sample status. Set the injection port conditions, column conditions, column oven temperature program, carrier gas flow rate, and detector conditions according to the corresponding methods. After the instrument is ready, the sample sequence is run.

[0059] The sample enters the chromatographic column through the injection port and is separated. Each component enters the detector in sequence, and the detector converts the component response into an electronic signal that changes over time. The data system saves this electronic signal as the original time-response sequence.

[0060] In this embodiment, the peak-level reference relationship refers to the correspondence between a target peak extracted from a reference standard and used for subsequent chromatographic analysis of the test sample. This relationship includes, but is not limited to, the retention time relationship of the target peak, the peak order relationship between the target peak and its adjacent peaks, the time interval relationship between the target peak and one or more reference peaks, the peak width relationship, and the positional relationship between the peak start point and peak end point relative to the local valley point.

[0061] The above relationships corresponding to the same target peak are saved and judged separately. For example, when the absolute retention time relationship of the target peak has deviated, the peak elution order between it and the adjacent peaks may still remain unchanged. At this time, all reference information cannot be discarded, nor can the deviated absolute retention time be used to directly determine the boundary of the target peak.

[0062] To facilitate the explanation of subsequent processing procedures, several specific terms involved in this embodiment are defined as follows:

[0063] A stable reference peak is a chromatographic peak that can be reliably identified in both the reference standard and subsequent test samples, and which does not exhibit significant changes in peak order, severe overlap, or unstable peak position in the current detection sequence. Stable reference peaks are mainly used to indicate the common changes in the chromatogram of the current test sample relative to the chromatogram of the reference standard, and are not required to be the quantitative target component in the final report.

[0064] Common drift refers to the change in retention time of multiple stable reference peaks relative to the reference standard in the current test sample. Common drift is used to explain the synchronous advancement or delay of multiple peaks caused by changes in the overall state of the chromatographic system, and does not directly indicate that the reference relationship of a certain target peak has become invalid.

[0065] Peak-level relation residuals refer to the remaining deviations after deducting the common drift from the actual changes of the target peak. These residuals reflect whether the target peak still maintains the original relative relationship with the stable reference peak in the control, and are the direct basis for judging whether a certain peak-level reference relationship can continue to be used.

[0066] To complete the reading, comparison, reanalysis, and result output of the above-mentioned chromatographic signals, the intelligent sensing system includes a chromatographic data reading unit, a detection sequence construction unit, a reference standard analysis unit, a stable reference peak determination unit, a common drift calculation unit, a peak-level reference relationship status determination unit, a local reanalysis unit, a quantitative result update unit, and a structured output unit. Each unit can be deployed in the computer where the chromatography workstation is located, a server connected to the chromatography workstation, or an analysis terminal with chromatographic data reading permissions. The intelligent sensing system takes the original time-response sequence already formed by the chromatograph as the main processing object, without changing the hardware structure of the pump, autosampler, column oven, injection port, chromatographic column, and detector, and does not take the instrument operation control as the processing center.

[0067] Based on the above-described chromatographic detection process and data processing object, this embodiment provides a structured analysis method for drug chromatographic signals based on an intelligent sensing system, comprising the following steps:

[0068] S1: Read the original time-response sequences of the reference standard, blank sample and multiple test samples obtained using the same detection method and the same method version, and establish the drug chromatographic detection sequence according to the actual injection order of each sample;

[0069] S2: Analyze the original time-response sequence of the reference standard, determine the peak start, peak apex and peak end of each target peak, extract the retention time, peak order, peak interval, peak width and peak boundary of each target peak, and establish the corresponding peak level reference relationship set;

[0070] S3: Establish candidate correspondences in the current test sample based on peak order and peak spacing relationships, identify at least two stable reference peaks from the known component peaks that can be identified in both the reference sample and the current test sample, and determine the common drift of the current injection based on the retention time changes of multiple stable reference peaks relative to the reference sample.

[0071] S4: Compare the actual changes of the target peak with the common drift to obtain the residuals of retention time relationship, peak interval relationship, and peak width relationship, and determine the peak sequence change results and peak boundary change results of the target peak;

[0072] S5: Based on the residuals of each peak level relationship, the results of peak sequence changes, the results of peak boundary changes, and the allowable range of the corresponding detection methods, each peak level reference relationship is determined as an effective state, a restricted state, a discontinued state, or a pending confirmation state.

[0073] S6: Based on the current status of each peak-level reference relationship, determine the usage rights of the corresponding relationship in candidate peak positioning, peak boundary determination and peak area integration, and determine the local re-analysis interval affected by the restricted or disabled status;

[0074] S7: Lock the confirmed baseline and peak parameters outside the local reanalysis interval, re-estimate the local baseline within the local reanalysis interval, generate peak boundary candidate results, and determine the final peak boundary and peak area of ​​the target peak based on the peak-level reference relationship that is still in a valid state.

[0075] S8: Update the drug content results directly related to the target peak based on the final peak area and the quantitative relationship corresponding to the current drug testing items;

[0076] S9: Write the detection sequence information, common drift, peak level relationship residuals, peak level reference relationship status, local re-analysis process, final peak parameters and content results into the structured analysis record.

[0077] The above steps are performed sequentially according to the formation, comparison and reanalysis of chromatographic detection data. Step S1 establishes the sequential relationship between each injection. Step S2 forms the reference standard required for the analysis of subsequent test samples in the sequence. Step S3 extracts the common drift that can represent the common changes in chromatographic signals from the current test sample. Step S4 determines the local changes of the target peak after excluding common changes. Steps S5 and S6 convert the local changes into the actual usage rights of the corresponding reference relationship. Step S7 only reanalyzes the local chromatographic signals affected by the restricted or discontinued state relationship. Steps S8 and S9 respectively complete the content result update and structured output.

[0078] In step S1, the chromatographic data reading unit reads the raw time-response sequence formed by each chromatographic run from the data file, data export interface or detection task storage location of the chromatographic workstation. Each raw time-response sequence consists of the sampling time and detection response value arranged in the sampling order, and is associated with the corresponding instrument type, detector type, detection method identifier, method version, sample name, sample type, injection sequence number, injection time and data file identifier.

[0079] To avoid the difference in peak position and shape caused by different detection conditions being mistaken for changes caused by continuous injection, the detection sequence construction unit only includes data obtained using the same detection method and the same method version into the same drug chromatographic detection sequence. The sample types include at least reference standards, blank samples, and test samples, and may also include parallel samples, repeated injection samples, and quality control samples. When the detection method or method version changes, a new set of peak level reference relationships is established, and the peak level reference relationships before the change are not directly used.

[0080] When forming the detection sequence, each injection retains its own independent chromatographic run time axis, without splicing multiple chromatograms end to end. The detection sequence only saves the sequential relationship of each injection, enabling the system to determine the specific injection position of the current test sample after the most recent reference sample, and to update the reference used for subsequent test samples when rerunning the reference sample or quality control sample in the middle of the detection sequence.

[0081] When the chromatography workstation can output method call records, operation logs, pump flow rates, mobile phase gradients, column temperatures, column oven temperature programs, carrier gas flow rates, or detector status events, the data reading unit will associate the corresponding records with the corresponding injections to confirm whether the chromatographic signal is formed according to the predetermined detection method. When the workstation cannot output continuous operating parameters, the system uses the parameter nodes, sample sequences, and readable operation logs already saved in the method file, and does not construct real-time status data that the instrument has not actually collected.

[0082] The drug chromatographic detection sequence formed in step S1 enables subsequent processing to clarify the temporal and sequential relationship between the reference standard and each test sample. The detection sequence serves as the data basis for establishing peak-level reference relationships in step S2 and determining stable reference peaks in step S3.

[0083] After completing the sequential association of each injection data, the system executes step S2 to establish a peak-level reference relationship based on the chromatographic signal of the reference standard in the detection sequence.

[0084] The reference standard analysis unit first performs local baseline estimation and candidate peak identification on the original time-response sequence of the reference standard. Signal smoothing is used to reduce the influence of random noise on the peak, local extrema and slope judgment. The smoothed signal is only used to help determine the peak interval. The peak area is still calculated based on the original detection response value and the corresponding local baseline to reduce the change of smoothing processing on the quantitative results.

[0085] For each target component, the reference standard analysis unit determines the peak start point, peak apex, and peak end point based on the target component information and reference standard chromatographic signal recorded in the detection task. The peak apex is the sampling position with the largest detection response after subtracting the local baseline within the corresponding peak interval. The sampling time corresponding to the peak apex is used as the target peak retention time. The peak start point and peak end point are determined jointly based on the intersection position of the chromatographic signal and the local baseline, the change in left and right slopes, and adjacent valley points. For target peaks with tailing, the peak end point is also determined by combining the continuity of the tail response regression to the local baseline.

[0086] After obtaining the basic parameters of the target peak, the reference analysis unit establishes retention time relationship, peak order relationship, peak interval relationship, peak width relationship and peak boundary relationship respectively. The retention time relationship records the initial peak time of the target peak in the reference. The peak order relationship records which known peaks the target peak is located before or after. The peak interval relationship records the relative time distance between the target peak and one or more reference peaks. The peak width relationship records the baseline peak width, half peak width or other peak width parameters used by the detection method for the target peak. The peak boundary relationship records the position of the peak start point and peak end point relative to the left and right valley points, adjacent peak boundaries and local baselines.

[0087] Each peak-level reference relationship for the same target peak forms an independent field. If the absolute retention time in subsequent test samples shifts, but the peak order relationship and peak interval relationship remain stable, the system can continue to retain the unchanged relationship. If a certain relationship has deviated significantly, it will not continue to participate in the peak boundary determination unconditionally because other relationships are normal.

[0088] The peak-level reference relationship set output in step S2 provides a control basis for selecting stable reference peaks and calculating the common drift of the current test sample in step S3.

[0089] Since multiple chromatographic peaks of subsequent test samples may advance or delay together with continuous injection, it is difficult to distinguish between the overall shift of the chromatographic signal and the local change of a single target peak by only comparing the absolute retention time of the target peak. Therefore, after establishing the peak-level reference relationship, the system executes step S3 to select a stable reference peak from the current test sample and determine the common drift.

[0090] The stable reference peak determination unit uses known peaks in the reference standard as candidates and searches for corresponding peaks in the chromatographic signal of the current test sample. The candidate peaks have clear peak apexes in both the reference standard and the current test sample. The changes in the peak apex position under different local baseline support point conditions are within the allowable range of the stable reference peak position. They can be repeatedly identified under different signal smoothing scales and do not show obvious changes in peak order or serious overlap. The allowable value of the stable reference peak position is determined based on method validation data or repeated injection results of qualified reference standards.

[0091] To avoid establishing erroneous peak correspondences based solely on a single absolute retention time before obtaining common drift, the system uses known peaks in the reference standard as candidate objects and generates candidate corresponding peaks in the current test sample. When a candidate corresponding peak has the same elution order as at least one adjacent known peak with which a candidate correspondence has been established, and the difference between their relative time interval in the current test sample and the corresponding relative time interval in the reference standard is within a preset initial range, a candidate correspondence is established. The preset initial range is determined based on the system applicability requirements of the corresponding detection method, method validation data, or repeated injection results of qualified reference standards. If one reference standard peak corresponds to multiple candidate corresponding peaks in the test sample, or multiple reference standard peaks correspond to the same candidate corresponding peak in the test sample, the corresponding candidate corresponding peak is not included in the stable reference peak set. Candidate peaks that cannot form a unique correspondence are treated as objects to be confirmed and do not participate in common drift calculation.

[0092] To reduce the impact of local changes in a single reference peak on the common drift, the stable reference peak determination unit selects at least two stable reference peaks from the candidate peaks. The stable reference peaks are preferentially distributed at different retention time positions before and after the target peak. When a candidate reference peak in the current test sample has an unclear peak top, a changed peak order, or serious overlap with adjacent peaks, the peak is removed from the current set of stable reference peaks.

[0093] Let the retention time of the i-th stable reference peak in the reference standard be . The retention time of the peak corresponding to the k-th sample injection in the current detection sequence is... The change in retention time of the stable reference peak. Calculate according to formula (1):

[0094] (1)

[0095] In the formula, A positive value indicates that the stable reference peak in the current test sample appears later than that of the control sample. A negative value indicates an earlier occurrence. This change in time only reflects the movement of a stable reference peak and cannot be directly used as a basis for judging local mismatch of the target peak.

[0096] After obtaining the time variations of each stable reference peak, the common drift calculation unit determines the common drift of the current injection based on multiple variations. When using the median value for calculation, the common drift... Calculate according to formula (2):

[0097] (2)

[0098] In the formula, m is the number of stable reference peaks involved in the calculation. Using the median value can reduce the impact of slight anomalies in a single stable reference peak on the common drift result, and is suitable for situations where the reliability of each stable reference peak is similar.

[0099] When there are differences in the sharpness of the peaks, the separation between peaks, or the stability of the signals of different stable reference peaks, the weights can be determined based on the reliability of each stable reference peak, and the common drift can be addressed. Calculate according to formula (3):

[0100] (3)

[0101] In the formula, This represents the weight of the i-th stable reference peak in the k-th injection. The more stable the peak position and the clearer the separation from adjacent peaks, the higher the corresponding weight. The weight is only used to indicate the contribution of each stable reference peak to the common drift calculation result and does not change the original chromatographic detection response.

[0102] Equations (2) and (3) are two alternative methods for determining common drift. One of them can be used in the same injection process. When the reliability of each stable reference peak is close and the number of stable reference peaks is not less than three, Equation (2) can be used. When there are differences in the reliability of each stable reference peak, Equation (3) can be used. When there are only two stable reference peaks, the difference in the change of retention time between the two is compared. If the difference does not exceed the allowable value for common drift consistency, the average value of the change of retention time between the two is used to determine the common drift. If the difference exceeds the allowable value for common drift consistency, the common drift is not determined, and the automatic peak area and content update of the current test sample is stopped. The allowable value for common drift consistency is determined based on the method validation data or the repeated injection results of qualified reference standards.

[0103] When stable reference peaks are distributed at different positions along the chromatographic run time axis, and the shifts of the initial and later stable peaks differ slowly, a common drift curve that continuously changes with retention time can be generated based on the initial retention times and their changes of each stable reference peak. The common drift value corresponding to the time position of the target peak can then be obtained from this curve. The common drift curve is still determined by the actual changes of multiple stable reference peaks, and is not backfitted based on the candidate position of a single target peak.

[0104] When using a single common drift, the common drift at the target peak location is still recorded as... When using a common drift curve, the common drift corresponding to the retention time position of the target peak j in the k-th injection is denoted as... That is, to obtain the curve value corresponding to the initial retention time position of the target peak of the reference standard, and then calculate the residual of the retention time relationship in equation (4). When using a common drift curve, replace with This is to avoid treating common drift values ​​at different retention times as the same.

[0105] When there are fewer than two available stable reference peaks, the system does not forcibly determine the common drift. Instead, it sets the relevant peak-level reference relationship as pending confirmation, does not automatically update the peak area and content results of the current target peak, and saves the current original time-response sequence and the reason for pending confirmation. Subsequent quality control samples, repeated injections, or re-runs of reference standards are used to re-establish the reference basis and determine whether the current test sample needs to be re-tested. When the number of stable reference peaks meets the requirements, the common drift obtained in step S3 is used to explain the changes in the common advance or delay of multiple chromatographic peaks and serves as the common benchmark for calculating the local changes of the target peak in step S4.

[0106] After obtaining the common drift, the system does not directly translate the target peak boundary based on it. Instead, it executes step S4 to subtract the common drift from the actual change of the target peak in order to obtain the peak-level relationship residual that can reflect the relative change of the target peak.

[0107] Let the retention time of the j-th target peak in the reference standard be . The retention time of the candidate target peak in the current test sample is The retention time relationship residual of the target peak Calculate according to formula (4):

[0108] (4)

[0109] If the target peak moves in near-synchronous fashion with multiple stable reference peaks, the retention time residual is close to zero, indicating that the change in the target peak can be explained by the common movement of the current chromatographic signal. If multiple stable reference peaks maintain a common change, while the target peak still shows additional advancement or retardation, the retention time residual increases, indicating that there may be a local mismatch in the original retention time relationship between the target peak and the reference standard.

[0110] To avoid judging the relationship status solely based on absolute retention time, the system further calculates the peak interval relationship residual between the target peak and the nearest stable reference peak. Let the retention times of the p-th stable reference peak in the reference standard and the current test sample be respectively... and Then the peak interval relationship residual Calculate according to formula (5):

[0111] (5)

[0112] Peak interval relationship residuals are used to determine whether the relative distance between the target peak and the nearby stable reference peak has changed additionally. Even if the absolute retention time of the target peak changes, as long as the relative interval between the target peak and the nearby stable reference peak remains stable, the peak interval relationship can still be used as the basis for subsequent peak analysis.

[0113] To determine whether the target peak exhibits abnormal broadening or narrowing relative to the reference standard, the system uses the peak width of the target peak in the reference standard. and the peak width of the target peak in the current test sample. Calculate the peak width relationship residual Specifically, calculate according to formula (6):

[0114] (6)

[0115] Peak width residuals are used to reflect the broadening state of the target peak in the current sample. When the target peak is tailed due to the influence of the sample matrix, even if the peak position changes slightly, its peak width and peak boundary relationships may still deviate significantly.

[0116] For peak sequence relationships that are not suitable for representation by continuous numerical values, the system records them as peak sequence maintained, peak sequence uncertain, or peak sequence changed, depending on whether the target peak is still located between the corresponding preceding and following reference peaks in the control sample. For peak boundary relationships, the system compares the relative positions of the target peak's starting point, peak ending point, left and right valley points, and adjacent peak boundaries, and records them as peak boundary maintained, peak boundary uncertain, or peak boundary changed. Each result is saved independently and is not converted into a comprehensive score to avoid a significant abnormality being offset by other normal indicators.

[0117] The residuals of peak-level relationships, peak sequence changes, and peak boundary changes generated in step S4 provide a basis for determining the usage status of each peak-level reference relationship in step S5.

[0118] After generating residuals for each peak-level relationship, the system executes step S5 to convert the residuals and changes into the state of the corresponding peak-level reference relationship. This addresses the issue that it is unreasonable to retain or discard all relationships when some relationships in the reference standard have changed but others can still be used.

[0119] The peak-level reference relationship determination unit sets a first allowable value and a second allowable value for each peak-level reference relationship based on the system applicability requirements of the corresponding drug detection method, method validation data, repeated injection results of qualified reference standards, or historical qualified detection data. The first allowable value is used to distinguish between normal changes and changes that require restricted use, and the second allowable value is used to distinguish between restricted use and discontinuation of use. Different allowable values ​​can be used for different drugs, different detection methods, and different target components. In one embodiment, the first allowable value is determined based on the allowable distribution range of the corresponding peak-level relationship residuals obtained from repeated injections of qualified reference standards, and the second allowable value is determined based on the first allowable value and an amplification factor greater than one.

[0120] For a peak-level reference relationship of type h that can be expressed numerically, let its residual be... The first allowed value is The second allowed value is ,and Less than The corresponding peak-level reference relationship status Determine according to formula (7):

[0121] (7)

[0122] The "valid" status indicates that the correspondence can continue to participate in candidate peak location, candidate interval determination, or peak boundary candidate result evaluation according to its relationship type. The "restricted" status indicates that the correspondence has deviated and can only be used to expand the candidate peak search range or assist in the evaluation of candidate results. It cannot be used to fix the peak apex, peak start point, peak end point, and peak area of ​​the target peak. The "discontinued" status indicates that the deviation of the correspondence has exceeded the second allowable value, or it conflicts with the peak sequence relationship or peak interval relationship that is still in the valid status. This relationship cannot continue to participate in the target peak location, boundary division, and area integration of the current test sample. The "pending confirmation" status indicates that the verification data of the corresponding peak-level reference relationship is insufficient and needs to be supplemented and confirmed by subsequent reference standards, quality control samples, parallel samples, or repeated injections.

[0123] Each peak-level reference relationship determines its state. For example, if the absolute retention time relationship of the target peak is in a restricted state, but the target peak is still located between two stable reference peaks, and the peak interval relationship with the adjacent stable reference peak is still in a valid state, the system no longer directly fixes the target peak according to the original absolute retention time, but continues to use the valid peak sequence relationship and peak interval relationship to limit the range of candidate peaks.

[0124] For peak sequence and peak boundary relationships, the state determination unit determines the state by combining the discrete change results and the peak sequence and boundary change conditions allowed by the corresponding detection method. When the order of the target peak and the adjacent peaks remains unchanged, the peak sequence relationship can remain valid; when the peak sequence cannot be reliably confirmed, the peak sequence relationship enters a restricted state; when the peak sequence changes and conflicts with the known peak relationship of the target component, the peak sequence relationship enters a deactivated state.

[0125] When candidate results given by different peak-level reference relationships conflict with each other, the candidate interval to which the target peak belongs is first determined by using the peak order relationship that is still in a valid state and at least one peak interval relationship. The retention time relationship is only used for sorting the candidate positions within the candidate interval. When the peak width relationship conflicts with the local valley point in the current chromatographic signal or the peak boundary relationship that is in a valid state, the peak start point and peak end point are not directly truncated according to the original peak width. Instead, the peak width relationship is only used as the evaluation basis for the candidate boundary. The peak width relationship residual is not used alone as the basis for determining the peak width relationship as a discontinued state. The peak width relationship residual is determined as a discontinued state only when the peak width relationship residual exceeds the second allowable value and there is a change in the peak boundary, an abnormal peak interval relationship, or an increase in the degree of overlap with adjacent peaks.

[0126] The relationship status obtained in step S5 is not only used to show whether the target peak is abnormal, but also serves as the direct basis for controlling the usage permissions of the corresponding reference relationship and determining the local re-parsing range in step S6.

[0127] After the peak-level reference relationship status is determined, the system executes step S6, which determines the subsequent parsing actions that the corresponding reference relationship can participate in based on different relationship statuses, and determines whether the current target peak needs to enter local re-parsing.

[0128] When all peak-level reference relationships of the target peak are in a valid state, the system completes the target peak location, peak boundary determination, and peak area integration based on the candidate region after common drift correction and the corresponding peak-level reference relationships, without initiating local re-analysis. At this time, the change of the target peak can be explained by the common movement of the current chromatographic signal or the allowable range of the detection method.

[0129] When the target peak has at least one peak-level reference relationship in a restricted or discontinued state, the system determines the local reanalysis interval based on the adjacent peak relationships that are still in a valid state, the stable valley point in the current chromatographic signal, the peak boundaries of adjacent confirmed peaks, and the continuous range of the local baseline.

[0130] When the target peak is located between two stable reference peaks, the time range between the right boundary of the previous stable reference peak and the left boundary of the next stable reference peak can be used as a local candidate interval. When there is no stable reference peak on one side of the target peak, the nearest stable valley point or the confirmed peak boundary on that side is used as the corresponding interval boundary. When the target peak is in the overlapping peak region, the local re-analysis interval can be extended to cover the entire overlapping region, but not to other stable peaks that are unrelated to the current mismatch relationship.

[0131] The local reanalysis unit first selects local baseline support points from the sampling points that do not fall into the candidate peak interval within the local reanalysis interval, and then linearly connects the local baseline support points located on both sides of the candidate peak to form local baseline candidate results. The locally redefined baseline is continuous with the locked baseline outside the interval at the left and right ends of the interval to avoid baseline jumps that do not belong to the actual chromatographic response at the interval boundary after local reprocessing.

[0132] The local re-analysis interval, the still usable peak-level reference relationship, and the interval locking results output in step S6 are used as processing conditions for re-estimating the local baseline and determining the target peak boundary in step S7.

[0133] After the local reanalysis interval is determined, the system executes step S7, which stops using peak-level reference relationships that are in a disabled state and only retains valid state relationships and restricted state relationships that are allowed to participate in auxiliary evaluation.

[0134] The local reanalysis unit first selects sampling points outside the candidate peak range where the signal changes continuously and do not fall into obvious peak areas as local baseline support points. The locally redefined baseline is continuous with the locked baseline outside the interval at the left and right ends of the interval, so as to avoid baseline jumps that do not belong to the actual chromatographic response at the interval boundary after local reprocessing.

[0135] Multiple sets of local baseline candidate results are generated based on different baseline support points. For each set of local baselines, after subtracting the baseline, the system generates multiple sets of peak boundary candidate results based on the local maximum response location of the response signal, adjacent valley points, and the intersection location of the chromatographic signal and the local baseline. Each set of candidate results includes at least the candidate peak start point, peak apex, and peak end point.

[0136] When there are shoulder peaks or overlapping peaks near the target peak, the system only generates candidate results corresponding to different common boundaries or different numbers of sub-peaks within the local reanalysis interval, without changing other chromatographic peaks outside the interval. Absolute retention time relationships or peak width relationships that are already in a deactivated state must not pre-fix the sub-peak positions; peak order relationships and peak interval relationships that are still in a valid state are only used to exclude candidate segmentation results that obviously do not conform to the current valid reference relationship.

[0137] Let the original detection response of the q-th sampling point in the local reanalysis interval be... The corresponding local baseline response is The candidate peak start point and peak end point correspond to the sampling points respectively. and The time interval between adjacent sampling points is Then the candidate peak area Calculate according to formula (8):

[0138] (8)

[0139] After obtaining multiple sets of candidate peak areas, the system checks whether the candidate results conform to the peak sequence relationship and peak interval relationship that are still in a valid state, whether the candidate peak is within the time range limited by the valid stable reference peak, whether the peak start point and peak end point correspond to local valley points or reasonable baseline junction positions, and whether the peak area change under adjacent local baseline candidate results is within the allowable range of the detection method.

[0140] If a candidate result is only valid when a reference relationship that is already in a deactivated state is continued to be invoked, the candidate result shall be excluded. For example, if a candidate peak can only correspond to the target component when the original absolute retention time that has been deactivated is forcibly adopted, and its peak order and peak interval are inconsistent with the current valid relationship, then the candidate result shall not be adopted.

[0141] If two or more sets of peak boundary candidate results simultaneously satisfy the above conditions, the system calculates the peak area change rate of the same peak boundary candidate result under different local baseline candidate results. Let the maximum peak area corresponding to the same peak boundary candidate result be... The minimum peak area is Then the peak area change rate The system prioritizes the candidate peak boundary with the smallest peak area change rate. If the peak area change rates are the same or their difference is lower than the preset resolution, the system selects the candidate peak with the smaller sum of the time distances between the peak start point and peak end point and the corresponding actual local valley point. The preset resolution is determined based on the results of repeated injections of qualified reference samples or method validation data. If a unique candidate peak cannot be obtained after the above comparison, the system does not automatically update the peak area and content of the target peak. Instead, it saves the original time-response signal, all candidate peak boundaries and the reasons for the failure to uniquely determine the peak, and calls up repeated injections, quality control samples or reruns of reference samples for supplementary confirmation.

[0142] Candidate results that satisfy the original time-response signal, local baseline continuity, and current effective peak level reference relationship are determined as the final analytical results of the target peak in the current test sample. Based on this, the system obtains the new peak start point, peak apex, peak end point, retention time, peak height, peak width, and peak area of ​​the target peak. The final peak parameters obtained in step S7 are used as input for updating the drug quantification results in step S8.

[0143] After the final peak boundary and peak area of ​​the target peak are determined, the system executes step S8, which writes the final peak area into the quantitative calculation relationship corresponding to the current detection item. This step does not change the area percentage method, external standard method, internal standard method, or correction factor method already used in the current drug detection item; it only replaces the target peak area field that is directly affected by local re-analysis.

[0144] When using the area percentage method, the system updates the corresponding component proportions based on the effective peak range and total effective peak area specified by the current detection method. When using the external standard method, the system calculates the target component content based on the reference standard concentration, reference standard peak area, test sample peak area, sample size, and dilution factor. When using the internal standard method or correction factor method, the system continues to use the predetermined internal standard concentration, internal standard peak area, reference standard concentration, and reference standard peak area.

[0145] In one implementation using an internal standard correction relationship, the correction factor f is calculated according to equation (9):

[0146] (9)

[0147] In the formula, A 内标 C represents the internal standard peak area. 内标 For internal standard concentration, A 对照 C represents the target peak area of ​​the reference standard. 对照 For the reference standard concentration, the test sample content is calculated based on the calculation logic already determined for the current detection item, using the re-analyzed target peak area, internal standard concentration, dilution volume, and corresponding injection volume of the test sample.

[0148] When a local reanalysis changes only one target peak, the system only recalculates the content results directly associated with that target peak. The areas of other peaks that are not in the local reanalysis interval and their corresponding content results remain unchanged. When the area percentage method is used and the change in the target peak area affects the total area of ​​the effective peaks, only the relevant results using the total area of ​​the effective peaks are updated, and the correspondence between the change in the target peak area and the update of the corresponding content results is recorded.

[0149] The final peak parameters and content results obtained in step S8, along with the previously formed common drift, relation residuals, relation states, and local reanalysis process, are passed to step S9.

[0150] After updating the target peak parameters and content results, the system executes step S9 to generate a structured analytical record based on a single target peak.

[0151] Each structured analysis record includes at least the detection sequence number, sample name, sample type, injection number, instrument type, detector type, detection method and method version, target peak parameters of the reference standard, stable reference peak used for current judgment, common drift, retention time relationship residual, peak interval relationship residual, peak width relationship residual, peak order change result, peak boundary change result, status of various peak level reference relationships, local reanalysis interval, peak start point, peak apex, peak end point and peak area before and after reanalysis, as well as the final content result and the source of result change.

[0152] When a peak-level reference relationship changes from an effective state to a restricted state or a deactivated state, the system records the corresponding relationship residual, the time of state change, and the basis for judgment. When the target peak undergoes local re-analysis, the system records the effective peak-level reference relationships used in this re-analysis, the deactivated relationships that were not used, the local baseline candidate results, and the basis for determining the final peak boundary.

[0153] After the detection sequence is re-run with reference standards, quality control samples, or repeated injections, the system recalculates the peak-level reference relationships that were previously in a restricted, discontinued, or pending confirmation state. When the corresponding relationship meets the first allowable range again and remains stable in a preset number of validation injections, it is restored to an effective state. If the restoration conditions are not yet met, it continues to be in a restricted, discontinued, or pending confirmation state. The process of relationship state change and its impact on the target peak are synchronously written into the structured analysis record.

[0154] Through the above processing, the system first uses multiple stable reference peaks to identify the common drift that occurs in the current chromatographic signal. Then, based on the remaining changes of the target peak relative to the common drift, it determines whether the specific peak-level reference relationship is still applicable. For the retention time, peak order, peak interval, peak width, and peak boundary relationship that is still valid, it continues to be used. For the relationship that has deviated, it restricts or stops its participation in the target peak analysis, and only redetermines the baseline, peak boundary, and peak area for the affected local chromatographic signal.

[0155] This process avoids directly judging normal changes such as multiple chromatographic peaks advancing or delaying as a failure of the reference relationship, and also prevents the reference relationship that has already experienced local mismatch from continuing to fix the position and boundary of the current test sample target peak. The entire process directly affects the chromatographic detection signal, retention time, peak boundary and peak area, and the resulting structured results are used for drug component qualitative analysis and content calculation.

[0156] Example 2

[0157] This embodiment performs offline validation of the processing results of the above method under conditions of common drift, local shift of the target peak, peak shape change, and local baseline change. Using the basic chromatographic parameters of a target peak retention time of 8.500 min and a control peak area of ​​1,200,000 as the validation benchmark, four types of conditions were set: common drift, common drift superimposed with local shift of the target peak, common drift superimposed with peak broadening or shoulder peaks, and baseline drift superimposed with abnormal reference peaks. Six groups were set for each type, forming a total of twenty-four validation sequences. The original time-response sequence was retained for each group, and the target peak retention time, peak start point, peak end point, and peak area determined through verification were used as the error calculation benchmark.

[0158] A fixed retention time window and a processing method using full-segment baseline integration were used as a control method. The method described in this embodiment follows the order of common drift determination, peak-level relation residual calculation, relation status judgment, and local re-analysis. The basic parameters used for verification are shown in Table 1.

[0159] Table 1 Verification of Basic Parameters

[0160]

[0161] Among them, the allowable value for the stable reference peak position and the allowable value for common drift consistency are used to determine whether the reference peak can participate in the common drift calculation; the first allowable value is used to distinguish between the effective state and the restricted state, and the second allowable value is used to distinguish between the restricted state and the deactivated state.

[0162] In Table 2, condition A represents common drift, condition B represents common drift and local shift, condition C represents common drift and peak broadening or shoulder peak, and condition D represents baseline drift and reference peak anomaly.

[0163] Table 2 Retention time data of twenty-four validation sequences

[0164]

[0165] Table 3. Peak area error and state judgment results of 24 sets of verification sequences.

[0166]

[0167] In Table 3, E indicates that all peak-level reference relationships are valid; R1 indicates that the retention time relationship is restricted and local reanalysis is initiated; R2 indicates that the peak width relationship is disabled and local reanalysis is initiated; R3 indicates that the peak width relationship is disabled and local reanalysis is performed on the shoulder peak region; R4 indicates that abnormal relationships are restricted and the local baseline is re-estimated; R5 indicates that abnormal relationships are restricted and local reanalysis is initiated; P indicates that the status is pending confirmation and peak area and content are not updated.

[0168] In Table 3, “—” indicates that the corresponding verification sequence has entered the pending confirmation state due to insufficient consistency of the stable reference peak or the inability to uniquely determine the local peak boundary. The system has not automatically updated the peak area and content results of the target peak.

[0169] Table 4 Summary of errors under different verification conditions

[0170]

[0171] As shown in Table 4, among the twenty-four verification sequences, the average error of the target peak retention time of the control method was 0.083 min, and the average relative error of the peak area was 8.037%. After adopting the method described in this embodiment, the average error of the target peak retention time was reduced to 0.014 min. Among the twenty-two verification sequences that generated automatically updated results, the average relative error of the peak area was 1.725%. The other two sequences entered the pending confirmation state, and peak boundaries that could not be uniquely confirmed were not used to update the quantitative results.

[0172] The following uses sample S09 to illustrate the calculation process of common drift, relational residuals, and state judgment. The retention time data of the three stable reference peaks are shown in Table 5.

[0173] Table 5. Retention time variation of stable reference peak in sample S09

[0174]

[0175] The consistency of the retention time changes of the three stable reference peaks satisfies the common drift calculation conditions, and the common drift of sample S09 is determined using the median value:

[0176]

[0177] In sample S09, the change in total retention time of the candidate target peak relative to the control target peak is:

[0178]

[0179] After deducting common drift, the retained time relationship residuals are:

[0180]

[0181] The residual is greater than the first allowable value of 0.030 min for the retention time residual, but does not exceed the second allowable value of 0.100 min. Therefore, the retention time relationship is determined to be in a restricted state. The retention time relationship in the restricted state no longer directly fixes the target peak position, but is only used to expand the candidate peak search range. The system continues to use the peak order relationship and peak interval relationship that are still in a valid state to determine the local re-analysis interval.

[0182] The control method directly uses the original retention time window and the entire baseline to determine the target peak boundary of sample S09, and its peak area error relative to the verification benchmark is 9.45%. This method re-estimates the local baseline and determines the final peak boundary from the candidate results that meet the valid peak sequence relationship, peak interval relationship and local baseline continuity, and the peak area error is reduced to 1.74%.

[0183] The above verification results show that common drift subtraction can distinguish between the synchronous movement of multiple chromatographic peaks and the local shift of the target peak; peak-level reference relationship status can control the subsequent use rights of invalid relationships; local reanalysis can redetermine the baseline and boundary of the affected target peak without changing the peak parameters outside the interval; for verification sequences that cannot obtain a unique boundary, the pending confirmation status can prevent unreliable areas from entering the content calculation.

[0184] In this embodiment, common drift, local shift, peak broadening, shoulder peak, and baseline change are all formed under set conditions based on the basic chromatographic parameters, and are used to illustrate the calculation process, state judgment conditions, and repeatability verification method of this method.

[0185] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A structured analysis method for drug chromatographic signals based on an intelligent sensing system, characterized in that, Includes the following steps: S1: Read the original time-response sequences of the reference standard and the test sample under the same detection method and method version, and establish the drug chromatographic detection sequence; S2: Analyze the reference standard, extract the retention time, peak order, peak interval, peak width and peak boundary of the target peak, and establish a peak-level reference relationship set; S3: Based on the peak sequence and peak interval, form a candidate correspondence in the current test sample, determine at least two stable reference peaks from the known component peaks, and determine the common drift based on their retention time changes; S4: Compare the changes of the candidate target peak relative to the control target peak with the common drift to obtain the retention time relationship residual, peak interval relationship residual and peak width relationship residual, and determine the peak order and peak boundary change results; S5: Based on the residuals of the retention time relationship, the residuals of the peak interval relationship, the residuals of the peak width relationship, the results of the peak sequence and peak boundary changes, determine each peak level reference relationship as an effective state, a restricted state, a disabled state or a pending confirmation state, and determine its usage rights in candidate peak positioning, peak boundary determination and peak area integration. S6: When all peak-level reference relationships are valid, determine the final peak boundary and peak area based on the peak-level reference relationships of the common drift and valid states. When there are restricted or deactivated states, determine the local reanalysis interval, lock the baseline and peak parameters outside the interval, re-estimate the local baseline within the interval and generate peak boundary candidate results, exclude candidate results that depend on the deactivated state, determine the final peak boundary and peak area based on the peak-level reference relationships of the valid states, and write them into the structured analysis record.

2. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 1, characterized in that: When establishing the chromatographic detection sequence of the drug, the original time-response sequence of each injection is associated with the detection method identifier, method version, sample name, sample type, injection number, injection time and data file identifier; each injection retains its own independent chromatographic running time axis, and the original time-response sequences of different injections are not spliced ​​together end to end; Only data obtained using the same detection method and the same method version are included in the same drug chromatographic detection sequence, and the sample types include at least reference standards and test samples.

3. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 2, characterized in that: The retention time relationship records the initial peak time of the target peak in the reference standard; the peak sequence relationship records the peak exit order between the target peak and the known peaks before and after it; the peak interval relationship records the relative time distance between the target peak and one or more reference peaks; the peak width relationship records the baseline peak width or half-peak width of the target peak; the peak boundary includes the peak start point and peak end point of the target peak; the peak boundary relationship records the positional relationship of the peak start point and peak end point relative to the local valley points on the left and right sides of the target peak, the peak boundaries of adjacent known peaks, and the corresponding local baselines.

4. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 3, characterized in that: When establishing the candidate correspondence, the known peaks in the reference standard are used as candidate objects, and candidate corresponding peaks are generated in the current test sample; When the elution order of the candidate corresponding peak is consistent with that of at least one adjacent known peak with which a candidate correspondence has been established, and the difference between the relative time interval between the two peaks in the current test sample and the corresponding relative time interval in the control sample is within a preset initial range, a candidate correspondence is established. The preset initial range is determined based on the system applicability requirements of the corresponding detection method, method validation data, or repeated injection results of qualified reference standards; when one reference standard peak corresponds to multiple candidate corresponding peaks of the test sample, or multiple reference standard peaks correspond to the same candidate corresponding peak of the test sample, the corresponding candidate corresponding peaks are not included in the stable reference peak set.

5. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 4, characterized in that: The stable reference peak has a clear peak in both the reference standard and the current test sample. The change in the peak position under different local baseline support point conditions is within the allowable range of the stable reference peak position, and there is no change in peak order or serious overlap. Let the retention time of the i-th stable reference peak in the reference standard be . The retention time of the corresponding peak in the k-th sample injection is The change in retention time of the stable reference peak. for: When the number of stable reference peaks m is no less than three, the common drift is... for: In the formula, m is the number of stable reference peaks involved in the calculation. When there are only two stable reference peaks, the difference in their retention time changes is compared. When the difference in the retention time changes of the two stable reference peaks does not exceed the allowable value for common drift consistency, the common drift is determined by the average value of their retention time changes. When it exceeds the allowable value for common drift consistency, the common drift is not determined, and the corresponding peak-level reference relationship is determined to be in a state of pending confirmation. The allowable values ​​for the position of stable reference peaks and the allowable values ​​for common drift consistency are determined based on the method validation data or the results of repeated injections of qualified reference standards.

6. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 5, characterized in that: Let the retention time of the target peak of the reference standard be... The retention time of the candidate target peak in the current test sample is The retention time relationship residual for: Let the retention times of the nearest stable reference peak in the reference standard and the current test sample be respectively. and Then the peak interval relationship residual for: Let the peak widths of the target peak in the reference standard and the current test sample be defined under the same peak width parameter. , Then the peak width relationship residual for: 。 7. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 6, characterized in that: For the h-th type of peak-level reference relationship that can be numerically expressed, excluding the peak width relationship, let its peak-level relationship residual be... : The first allowed value is The second allowed value is ,and < ; when When this happens, the corresponding peak-level reference relationship is determined to be in a valid state; when When this occurs, it is determined to be in a restricted state; when When the peak reference relationship is insufficient, it is determined to be in a disabled state; when the verification data of the corresponding peak reference relationship is insufficient, it is determined to be in a pending confirmation state.

8. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 7, characterized in that: When the peak order change results in the peak order being maintained, the peak order relationship is determined to be a valid state; When the peak order cannot be reliably confirmed, the peak order relationship is determined as a restricted state; When the peak order changes and conflicts with the known peak relationship of the target component, the peak order relationship is determined to be in a deactivated state; the retention time relationship in a valid state is used for candidate position ranking, the peak order relationship and peak interval relationship in a valid state are used to determine the candidate interval, and the peak width relationship and peak boundary relationship in a valid state are used to evaluate the peak boundary candidate results. Peak-level reference relationships in a restricted state are only used to expand the search range of candidate peaks or to assist in the evaluation of candidate results. Peak-level reference relationships in a deactivated state do not participate in peak location, peak boundary determination, or peak area integration. The peak width relationship residual is not used as the sole basis for determining the peak width relationship to be in a disabled state. When the peak width relationship residual exceeds the corresponding second allowable value, and there is a change in the peak boundary, an abnormal peak spacing relationship, or an increase in the degree of overlap with adjacent peaks, the peak width relationship is determined to be in a disabled state.

9. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 8, characterized in that: The local reanalysis interval is determined based on the relationship between adjacent peaks that are still in a valid state, stable valley points, peak boundaries of adjacent confirmed peaks, and the continuous range of local baselines. When the target peak is located between two stable reference peaks, the time range between the right boundary of the previous stable reference peak and the left boundary of the next stable reference peak is defined as the local reanalysis interval. Within the local reanalysis interval, local baseline support points are selected from the sampling points that have never fallen into the candidate peak interval, and the local baseline support points located on both sides of the candidate peak are linearly connected to form local baseline candidate results. Peak boundary candidate results are generated based on the local maximum response location, adjacent valley points, and the intersection location of the chromatographic signal and the local baseline of the response signal after removing the local baseline. Let the original detection response of the q-th sampling point be... The corresponding local baseline response is The candidate peak start point and peak end point correspond to the sampling points respectively. and The time interval between adjacent sampling points is Then the candidate peak area for: 。 10. The method for structured analysis of drug chromatographic signals based on an intelligent sensing system according to claim 9, characterized in that: When there are two or more sets of peak boundary candidate results that simultaneously meet the peak order relationship, peak interval relationship and local baseline continuity in a valid state, calculate the peak area change rate of the same peak boundary candidate result under different local baseline candidate results. Let the maximum peak area corresponding to the candidate results of the same peak boundary be . The minimum peak area is The peak area change rate ; The peak boundary candidate with the smallest peak area change rate is selected; when the peak area change rates are the same or their difference is lower than the preset resolution value, the peak boundary candidate with the smaller sum of the time distances between the peak start point and the peak end point and the corresponding local valley point is selected. The preset resolution value is determined based on the repeated injection results of qualified reference standards or method verification data. If a unique peak boundary candidate result cannot be obtained after comparison, the peak area and drug content results of the target peak are not automatically updated; if a unique peak boundary candidate result is obtained, the drug content result directly associated with the target peak is updated according to the corresponding final peak area, and the drug chromatographic detection sequence, stable reference peak, common drift, peak level relationship residual, current status of each peak level reference relationship, local reanalysis interval, final peak parameters and drug content results are written into the structured analysis record.

Images