A nanopore clogging self-diagnosis and recovery method and system based on pose template drift

Abstract

The present application relates to a kind of based on pose template drift Nanopore pollution block hole self-diagnosis and recovery method and system, method includes: collecting same kind of biological particles, the same kind of biological particles is introduced to the Nanopore in healthy state, generates blocking event;The pose feature of the blocking event is divided into multiple pose template categories, and the real-time pose template probability distribution is obtained;The drift vector and drift intensity of the real-time pose template probability distribution and healthy pose template probability distribution baseline are calculated, to judge the abnormal type of the Nanopore, and corresponding hierarchical recovery measures are executed.The present application improves Nanopore detection stability, pore utilization and result reliability.

Description

Technical Field

[0001] This invention relates to the field of nanopore detection and biosensing technology, and in particular to a method and system for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift. Background Technology

[0002] Solid-state nanopore detection records ion current blocking events generated when particles, nucleic acids, proteins, exosomes, virus-like particles, or other biological particles pass through nanopores. This data provides information such as blocking amplitude, relative blocking amplitude, duration, event area, multi-level blocking morphology, baseline noise, orifice current, and event frequency. Because nanopores typically have nanoscale pore sizes and high surface energy, sample aggregates, protein adsorption layers, lipid vesicles, salting-out particles, bubbles, or large impurities may remain near the pore opening or wall during detection, causing pore wall contamination, partial blockage, complete blockage, or transient pseudo-blockage. Existing systems typically rely on empirical judgment based on indicators such as decreased orifice current, increased noise, decreased event count, and increased long-stay events. Recovery is achieved through buffer rinsing, reverse voltage, pulse voltage, pressure perturbation, cleaning reagents, or pore channel replacement. More specific existing technologies mainly include the following directions: Existing technologies disclose methods to improve particle / biomolecule transport in nanopores through fluid lipid walls or coatings, thereby reducing nonspecific interactions and improving detection resolution. These methods focus on pore wall interface modification, coatings, and transport control. However, they have limitations: they do not disclose using the probability distribution of attitude templates formed by similar biological particles in healthy pores as a baseline for pore health; they do not disclose differentiating pore wall contamination, partial blockage, complete blockage, and transient interference based on the drift direction of the attitude template distribution; and they do not disclose using attitude template distribution regression as a criterion for successful loop closure.

[0003] Existing technical solutions also disclose the use of nanopores to read microbial signals, extract signal attributes, correlate different attributes, classify multiple events, and identify / quantify microorganisms. This solution demonstrates that nanopore event feature classification and machine learning for sample object identification have become an existing trend. Its limitations are: its diagnostic object is the type and quantity of microorganisms, rather than the health status of the nanopore channels; its event classification results are used to identify sample components, rather than to calculate the posture template drift of pore contamination / clogging; and it does not disclose the anomaly source typing, graded recovery, and post-recovery template regression validation.

[0004] Existing technologies also disclose methods that, after detecting a molecule passing through a nanopore, control signals, such as voltage reversal, can be used to allow the same molecule to re-pass through the nanopore, thereby achieving capture, recapture, or spatial trapping. This approach demonstrates that reverse voltage and feedback control can be used to control molecular motion near nanopores. Its limitations lie in the fact that its core focus is on molecule recapture and repeatable analysis, rather than determining contamination / blockage based on template drift in pore health, and it does not address different recovery strategies for different sources of anomalies.

[0005] Existing technologies also disclose devices, systems, and methods for nanoporous units, droplet transport, and high-throughput biomolecule processing. These methods involve arraying, droplet transport, and biomolecule analysis structures. Their limitation lies in their focus on sample processing and high-throughput analysis, rather than on the self-diagnosis and recovery loop of nanopore contamination / clogging posture template drift.

[0006] Existing technical solutions also disclose the use of deep learning to process nanopore current signals, identify events, and extract various event features. This literature indicates that automatic event recognition and machine learning classification are already in a similar background. Its limitations are: its goal is to improve the ability of nanopore event recognition and sample signal classification, not to diagnose pore contamination / clogging; and it does not organize healthy template distribution, real-time template drift, anomaly classification, graded recovery, and recovery verification into a closed-loop method.

[0007] Existing technical solutions also disclose chemically controlled solid nanopores, stable long-term collection of large amounts of DNA and protein events, and mention real-time cleaning and restoration of clogged pores. This literature is highly relevant to "pore clogging recovery." Its limitations are: it focuses on the stability of pore preparation and cleaning and restoration experiments, but does not disclose the use of attitude template distribution drift to identify contamination types, nor does it disclose attitude template regression as a condition for successful recovery.

[0008] Existing technical solutions also disclose the pore-blocking dynamics of DNA in hafnium oxide nanopores, revealing that the pore-blocking probability is related to factors such as DNA length and bias voltage. This literature can serve as background information on the pore-blocking mechanism. Its limitation is that it focuses on the pore-blocking mechanism and does not disclose online self-diagnosis, abnormal typing recovery, and array pore health weight management.

[0009] Although the aforementioned existing technologies involve orifice wall modification, event classification, reverse voltage feedback, orifice plugging mechanism, and orifice plugging recovery, they still have the following shortcomings: 1. Existing methods for determining well blockage often rely on single indicators such as well current, noise, number of events, or long-stay events, which make it difficult to distinguish between sample concentration fluctuations, bubble transients, readout noise, well wall contamination, partial blockage, and complete blockage.

[0010] 2. Existing event classification technologies mainly serve sample type identification or molecular parameter extraction, but do not convert the multi-pose event distribution of similar biological particles in healthy pore states into a pore health baseline.

[0011] 3. Existing cleaning or restoration procedures are usually performed based on experimental experience. There is no calculable mapping between the restoration intensity and the source of the anomaly, which can easily lead to excessive pulses, over-cleaning, changes in pore size, or changes in the surface properties of the pore wall.

[0012] 4. Current successful recovery is mostly based on the recovery of orifice current, the reappearance of events, or the reduction of noise, but lacks verification of whether the event attitude structure has returned to a healthy distribution, which may lead to the omission of latent contamination states.

[0013] 5. Existing array systems typically focus on parallel detection of multiple pores or sample transport, lacking methods for dynamic deweighting, isolation, and quality tracking of pores based on template drift and recovery reliability of each pore. Summary of the Invention

[0014] To address the problems existing in the prior art, the present invention aims to provide a self-diagnosis and recovery method for nanopore contamination and blockage based on attitude template drift. This method enables nanopore detection systems to no longer rely solely on opening current or event count to determine pore status. Instead, it uses the probability distribution of attitude templates formed by similar biological particles in a healthy pore state as the baseline for pore health. During the detection process, the attitude template distribution drift is calculated in real time. Based on the drift direction, drift intensity, and auxiliary electrical indicators, the method distinguishes between pore wall contamination, partial blockage, complete blockage, transient interference from bubbles / particles, and readout anomalies. It then implements graded recovery measures matched to the source of the anomaly. Finally, the regression of the attitude template distribution is used as a necessary or priority criterion for successful recovery, thereby improving the stability, pore utilization, and reliability of nanopore detection results.

[0015] To achieve the above objectives, the present invention provides the following solution: A self-diagnosis and recovery method for nanopore contamination and blockage based on attitude template drift includes: Collect similar biological particles and introduce them into nanopores in a healthy state to generate a blocking event. The attitude features of the blocking event are divided into multiple attitude template categories to obtain the real-time attitude template probability distribution; The drift vector and drift intensity of the real-time posture template probability distribution and the baseline of the healthy posture template probability distribution are calculated to determine the abnormality type of the nanopore and to implement corresponding graded recovery measures.

[0016] Optionally, calculating the drift vector and drift intensity between the baseline of the healthy posture template probability distribution and the baseline of the healthy posture template probability distribution includes: ; ; ; in, Let be the attitude template drift vector of the i-th channel within the time window t. Let be the attitude template drift vector of the i-th channel within the time window t. The baseline of the health posture template probability distribution for the i-th channel is... Let be the attitude template distribution distance of the i-th channel within the time window t. This represents the probability distribution of the real-time attitude template after smoothing. The baseline for the probability distribution of the healthy pose template after smoothing. It is an intermediate distribution. Let be the proportion of the i-th channel within the K-th pose template in the time window t. Let be the baseline of the probability distribution of the Kth type of healthy posture template for the i-th channel within the time window t. The divergence is Kullback-Leibler.

[0017] Optionally, determining the anomalous type of the nanopore includes: If the drift vector shows a continuous increase in abnormal long dwell templates and / or slow baseline recovery templates in the attitude template category, accompanied by a slow decrease in orifice current, then an orifice wall contamination trend is determined; or when the comprehensive drift score is greater than the first abnormal threshold, and the increment of orifice wall contamination-related templates is greater than the orifice wall contamination template increment threshold, and the orifice current shows a continuous decreasing trend relative to the healthy orifice current, then orifice wall contamination is determined. If the number of low-width dwell templates and / or multi-level blocking templates in the attitude template category increases, accompanied by a decrease in the number of events, or when the comprehensive drift score is greater than the first anomaly threshold. If the incremental value of some plugging-related templates is greater than the threshold value of some plugging-related templates, and the number of events per unit time is lower than the preset ratio of the number of healthy events, then it is judged as partial plugging. If the number of events per unit time is close to zero and the orifice current deviates significantly from the healthy range, or if the number of events per unit time is not greater than the target threshold or the orifice current relative to the healthy orifice current is lower than the full orifice blockage current ratio threshold, and this state is maintained continuously for a preset window number, then it is determined to be a complete orifice blockage. If a short window abruptly changes and can regress on its own, it is judged to be a transient disturbance of bubbles or particles; If the noise is abnormal but the attitude template distribution distance does not drift accordingly, it is determined to be an abnormal readout link or an external electrical disturbance.

[0018] Optionally, the comprehensive drift score includes: ; ; ; ; in, To achieve a comprehensive drift score by integrating attitude template distribution distance and auxiliary electrical parameters, This is the non-negative weighting coefficient in the overall drift score. For the deviation of the opening current, Baseline noise bias, For the event number deviation, Let be the real-time opening current of the i-th channel within the time window t. Let be the healthy opening current of the i-th channel within the time window t. , Let represent the real-time baseline noise and healthy baseline noise of the i-th channel within the time window t, respectively. This is used to avoid small positive quantities with a denominator of zero or a logarithm of zero.

[0019] Optionally, the template increment related to pore wall contamination includes: ; in, This represents the set of templates related to pore wall contamination. The increment of the cumulative proportion of the contamination-related posture template set of the i-th channel within the time window t relative to the baseline of the healthy state is given. Let be the cumulative percentage of the set of attitude templates related to the contamination of the i-th channel within the time window t. This represents the baseline cumulative percentage of the set of attitude templates related to pore wall contamination under the healthy state of the i-th pore.

[0020] Optionally, the partial plugging-related template increment includes: ; in, This is a collection of templates related to partial hole plugging. The percentage of the partially blocked attitude template for the i-th channel within the time window t is the increment relative to the baseline of the healthy state. Let be the cumulative percentage of each template in the set of attitude templates related to partial hole blocking within the time window t for the i-th channel. This represents the cumulative percentage of each template in the set of partially blocked attitude templates under the healthy state of the i-th duct.

[0021] Optionally, implementing corresponding tiered recovery measures includes: When the anomaly type is well wall contamination, a low-intensity recovery is employed; the low-intensity recovery includes buffer rinsing, reducing sample concentration, low-amplitude pulses, or short-term bypass. When the abnormality type is partial blockage, a medium-intensity recovery method is used; the medium-intensity recovery method includes reverse voltage, pulse voltage, local flushing, or reducing the flow rate; When the abnormality type is complete blockage, high-intensity recovery is used; the high-intensity recovery includes multiple rounds of reverse voltage, strong pulse, cleaning reagent flushing, channel isolation or switching to a backup channel; When the anomaly type is a bubble or particle transient interference, sampling is paused or the anomaly window is marked. When the anomaly type is a readout link anomaly or an external electrical disturbance, a circuit self-test or shielding of the abnormal time period data is used.

[0022] Optionally, the method further includes: After the graded recovery measures are completed, if the recovered aperture current meets the healthy aperture current range, but the drift intensity of the real-time attitude template probability distribution after recovery relative to the baseline of the healthy attitude template probability distribution is still higher than the recovery threshold, then the nanopore is marked as a latent unrecovered state, and the recovery measures continue to be implemented. At the same time, the pore health is calculated, and the pore weight is reduced or the pore is isolated based on the pore health.

[0023] Optionally, calculating the health of the orifice includes: ; ; in, This represents the health status of the i-th channel. This is the drift penalty coefficient. To restore reliability, The number of successful recovery attempts up to time window t. This represents the total number of recovery attempts up to time window t.

[0024] To achieve the above objectives, the present invention also provides a self-diagnosis and recovery system for nanopore contamination and blockage based on attitude template drift, comprising: The signal acquisition module is used to acquire similar biological particles, introduce the similar biological particles into the nanopores in a healthy state, and generate a blocking event. The posture template recognition module is used to classify the posture features of the blocking event into multiple posture template categories to obtain a real-time posture template probability distribution. The anomaly classification and recovery decision module is used to calculate the drift vector and drift intensity of the real-time posture template probability distribution and the baseline of the healthy posture template probability distribution, in order to determine the anomaly type of the nanopore and execute the corresponding graded recovery measures.

[0025] The beneficial effects of this invention are as follows: 1. This invention separates "well health status" from "sample type identification" and uses the probability distribution drift of the posture template of similar biological particles to identify well contamination / clogging, avoiding the simple application of event classification technology to sample classification.

[0026] 2. This invention not only observes the opening current, noise, or number of events, but also observes the directional changes in the proportion of templates with different postures, which can detect mild hole wall contamination and hidden blockages earlier.

[0027] 3. This invention links the source of the anomaly with the recovery measures, and selects recovery strategies according to the hole wall contamination, partial hole blockage, complete hole blockage, transient interference and readout anomaly, thereby reducing excessive cleaning and excessive pulses.

[0028] 4. This invention uses attitude template distribution regression as an important condition for successful recovery, and can identify the latent unrecovered state of "current recovery but template failure".

[0029] 5. This invention is applicable to nanopore arrays, and can deweight or isolate pores based on health and recovery reliability, thereby reducing the impact of abnormal pores on the overall detection results.

[0030] 6. Compared with hole wall modification, event classification, reverse voltage control and ordinary hole plugging and cleaning methods, this invention forms a complete closed loop of healthy baseline, drift calculation, anomaly classification, graded recovery and regression verification. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a flowchart of a self-diagnosis and recovery method for nanopore contamination and blockage based on attitude template drift according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the construction of blocking event posture features and posture template categories according to an embodiment of the present invention; Figure 3 This is a diagram showing the relationship between the probability distribution of the healthy posture template, the probability distribution of the real-time posture template, and the drift vector in an embodiment of the present invention. Figure 4 This is a logic diagram illustrating the anomaly source classification in an embodiment of the present invention. Figure 5 This is a diagram showing the selection of graded recovery measures according to an embodiment of the present invention; Figure 6 This is a schematic diagram of template regression verification after recovery according to an embodiment of the present invention; Figure 7 This is a schematic diagram of template regression verification after recovery according to an embodiment of the present invention; Figure 8 This is a schematic diagram comparing the present invention with the prior art in an embodiment of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] The diagnostic object of this embodiment is the health status of the nanopore channels, rather than the type of sample being tested. The attitude template is not a simple sample classification label, but a set of attitude-related current event categories formed by the same type or class of biological particles in healthy pores; when pore wall contamination, pore blockage, or external disturbances occur, the probability distribution of the attitude template of the same type of particles will show directional drift, and this drift can serve as a statistical fingerprint of the health status of the channels.

[0036] The diagnostic object of this embodiment is the health status of the nanopore channels, rather than the type of sample being tested. The attitude template is not a simple sample classification label, but a set of attitude-related current event categories formed by the same type or class of biological particles in healthy pores; when pore wall contamination, pore blockage, or external disturbances occur, the probability distribution of the attitude template of the same type of particles will show directional drift, and this drift can serve as a statistical fingerprint of the health status of the channels.

[0037] like Figure 1 As shown, this embodiment discloses a self-diagnosis and recovery method for nanopore contamination and blockage based on attitude template drift, including: when the nanopore is in a healthy state, collecting blocking events caused by similar biological particles passing through the nanopore, and establishing a baseline for the probability distribution of a healthy attitude template. During the detection process, nanopore ion current signals are acquired and blocking event attitude features are extracted. The blocking event posture features Mapped to multiple pose template categories The real-time attitude template probability distribution is obtained. ; Calculate the probability distribution of the real-time attitude template Relative to the baseline of the probability distribution of the healthy posture template drift vector and / or drift intensity According to the drift vector The drift direction of the attitude template probability distribution is determined, and the degree of anomaly is determined based on the drift intensity, comprehensive drift score, and at least one auxiliary indicator among orifice current, baseline noise, and event count. The drift intensity is then used to determine the degree of anomaly. Comprehensive drift score and orifice current Baseline noise , number of events At least one auxiliary indicator is used to determine the degree of anomaly, and the incremental changes of different attitude template categories are combined to identify at least one type of anomaly, including orifice wall contamination, partial or complete blockage, transient interference from bubbles or particles, and readout anomalies. Corresponding graded recovery measures are implemented according to the anomaly type. After the graded recovery measures, the attitude template probability distribution drift is recalculated to determine whether the recovered attitude template probability distribution returns to the baseline of the healthy attitude template probability distribution. As one of the criteria for successful recovery.

[0038] Specifically, this embodiment discloses a self-diagnosis and recovery method for nanopore contamination and blockage based on attitude template drift, including: S1, Pore health verification. Initial opening current, baseline noise, and blank buffer are used to verify the nanopores, and pores with unstable opening current, abnormal noise, or initial blockage are removed.

[0039] S2, Establish a healthy posture template baseline. Introduce similar biological particle samples into the healthy channel, collect a preset number of blocking events, classify the blocking events into multiple posture template categories, and obtain the healthy posture template probability distribution baseline, healthy opening current, healthy baseline noise, and number of healthy events.

[0040] S3, Real-time Acquisition and Template Mapping. During the detection process, pose features are extracted from blocking events within each sliding time window or event window, and each blocking event is mapped to a pose template category.

[0041] S4, Calculate the real-time attitude template distribution. Calculate the proportion of each attitude template category in the current window to obtain the real-time attitude template probability distribution.

[0042] S5, calculate attitude template drift. Compare the real-time attitude template probability distribution with the baseline of the healthy attitude template probability distribution to obtain the drift vector and drift intensity; if there is no effective blocking event in the current window, enter the no-event branch, and combine the orifice current and noise to determine whether the orifice is completely blocked, the sample is exhausted, or there is transient interference, without forcibly calculating the probability distribution distance.

[0043] S6. Perform anomaly classification. If the drift vector shows a continuous increase in abnormally long dwell templates and / or slow baseline recovery templates, accompanied by a slow decrease in orifice current, it is judged as orifice wall contamination; if the low-width dwell templates and / or multi-level blocking templates increase, accompanied by a decrease in the number of events, it is judged as partial orifice blockage; if the number of events is close to zero and the orifice current deviates significantly from the healthy range, it is judged as complete orifice blockage; if there is only a short window abrupt change and it can return to normal on its own, it is judged as bubble or particle transient interference; if the noise is abnormal but the attitude template distribution has no corresponding drift, it is judged as readout link anomaly or external electrical disturbance.

[0044] S7, Perform graded recovery. Select the recovery strategy based on the anomaly type and drift intensity. For well wall contamination, prioritize low-intensity recovery, including buffer flushing, reducing sample concentration, low-amplitude pulses, or short-term bypass; for partial blockage, use medium-intensity recovery, including reverse voltage, pulsed voltage, local flushing, or reduced flow rate; for complete blockage, use high-intensity recovery, including multiple rounds of reverse voltage, strong pulses, flushing with cleaning reagents, well isolation, or switching to a backup well; for transient interference, pause sampling or mark the abnormal window; for readout anomalies, use circuit self-test or shield data during the abnormal time period.

[0045] S8, Closed-loop verification after recovery. After the recovery measures are completed, events are reacquired and the attitude template drift is recalculated. If the drift intensity after recovery is lower than the recovery threshold, and the number of events, orifice current, or noise meets the recovery conditions, the recovery is considered successful; if the orifice current has been recovered but the attitude template distribution has not returned, it is considered as implicitly unrecovered and the next level of recovery is performed or the orifice weight is reduced.

[0046] S9, Array Channel Health Management. For nanopore arrays, channel health is calculated based on the drift intensity, anomaly type, recovery count, and recovery success rate of each channel. Channels with low health are downweighted or isolated, so that the overall identification results are primarily contributed by healthy channels.

[0047] S10 outputs diagnostic and recovery records. It records the anomaly type, drift vector, drift intensity, recovery measures, template distribution before and after recovery, and recovery results for each channel. like Figure 2 As shown, Figure 2 It displays current blocking amplitude, relative amplitude, duration, event area, entry edge slope, exit edge slope, multi-level blocking, platform flatness, and baseline recovery time, and shows multiple attitude template categories.

[0048] The blocking event posture characteristics Including blocking amplitude Relative blocking amplitude Duration Event area , enter along the slope Exit along the slope Multi-level blocking characteristics Platform flatness Baseline recovery time Disturbance response characteristics At least three of them.

[0049] Specifically, the pose feature vector of the i-th channel during the n-th event within the time window t is: (1); in, Let be the attitude feature vector of the nth blocking event in the i-th channel, where i is the i-th channel number in the nanopore or nanopore array, n is the n-th blocking event number within the time window t, and t is the sliding time window or event window number. These represent the blocking amplitude, relative blocking amplitude, duration, event area, entry slope, exit slope, multi-level blocking characteristics, plateau flatness, baseline recovery time, and disturbance response characteristics, respectively.

[0050] Furthermore, the blocking event pose features Mapped to multiple pose template categories The attitude template mapping function is: (2); in, For attitude template mapping function, Let K be the k-th pose template, where K is the total number of pose template categories.

[0051] The number of template events of the kth term is: (3); in, The i-th channel is mapped to the k-th pose template within the time window t. The number of events, Let be the set of valid blocking events detected in the i-th channel within the time window t. Let represent the attitude feature vector of the i-th channel during the n-th blocking event within the time window t. This represents the k-th pose template.

[0052] when At that time, the proportion of the k-th type template is: (4); in, Let be the proportion of the i-th channel within the k-th pose template in the time window t. Let be the number of effective blocking events for the i-th channel within the time window t. .

[0053] when At that time, the probability distribution of equation (4) is not calculated, and the process enters the event-free branch; this branch combines the real-time opening current of the i-th channel within the time window t. Real-time baseline noise of the i-th channel within time window t The system determines whether the sample is completely blocked, exhausted, or subject to transient interference based on the sample injection status.

[0054] Furthermore, the posture template category This includes at least four of the following: fast pass template, normal short dwell template, normal long dwell template, low swath dwell template, multi-level blocking template, abnormally long dwell template, slow baseline recovery template, noise-accompanied template, and disturbance response template; wherein, the attitude template category It is used to characterize changes in the health status of nanopore channels, rather than just to identify the type of sample being tested.

[0055] like Figure 3 As shown, Figure 3 Display using bar charts, radar charts, or probability distribution maps. and .

[0056] Furthermore, the drift vector satisfies ,in, Let represent the real-time attitude template probability distribution of the i-th nanopore within the time window t. represents the baseline of the healthy posture template probability distribution for the i-th nanopore.

[0057] Specifically, the baselines for the real-time pose template probability distribution and the healthy pose template probability distribution are as follows: (6); in, Let represent the real-time attitude template probability distribution of the i-th channel within the time window t. Let be the baseline of the health posture template probability distribution for the i-th channel.

[0058] The attitude template drift vector is: (7); in, Let be the attitude template drift vector of the i-th channel within the time window t.

[0059] Furthermore, the drift intensity Based on the probability distribution of real-time attitude template Baseline of probability distribution of healthy posture template The drift intensity is obtained from the Jensen-Shannon distance, total variation distance, Wasserstein distance, cosine distance, or a weighted combination thereof; or, the drift intensity is further correlated with the orifice current deviation. Baseline noise deviation , event number deviation At least one of the following is used to obtain the comprehensive drift score. .

[0060] Specifically, to avoid zero probability in the logarithmic terms, the probability distribution can be smoothed first: (8); in, This represents the number of template events of type k in the health baseline. To avoid small positive quantities with a denominator of zero or a logarithm of zero, Given the number of baseline events in the window corresponding to the i-th channel in a healthy state, the smoothed probability distribution of formula (8) further represents... and .

[0061] The intermediate distribution is as follows: (9); in, It is an intermediate distribution.

[0062] The distance between Jensen and Shannon is: (10); in, The distance of the attitude template distribution of the i-th channel within the time window t.

[0063] The Kullback-Leibler distance is defined as: (11); in, The Kullback-Leibler distance. Let be the real-time attitude template probability distribution of the i-th channel after smoothing within the time window t. Let i be the baseline of the healthy posture template probability distribution after smoothing the i-th channel. Let be the probability value of the k-th type of pose template in the real-time pose template probability distribution. Let K be the probability value of the k-th type of posture template in the baseline of the healthy posture template probability distribution, where K represents the number of posture template categories.

[0064] The auxiliary electrical parameters are normalized as follows: (12); (13); (14); in, For the deviation of the opening current, Baseline noise bias, For the event number deviation, Let be the real-time opening current of the i-th channel within the time window t. Let be the healthy opening current of the i-th channel within the time window t. , Let represent the real-time baseline noise and healthy baseline noise of the i-th channel within the time window t, respectively. This is used to avoid small positive quantities with a denominator of zero or a logarithm of zero.

[0065] Overall drift rating: (15); in, To achieve a comprehensive drift score by integrating attitude template distribution distance and auxiliary electrical parameters, ,and: (16); in, This represents the non-negative weighting coefficient in the overall drift score.

[0066] The drift vector is used to characterize the direction of change of the attitude template probability distribution and to determine the trend of change of the anomaly-related template; the comprehensive drift score, the increment of the related template, and the auxiliary electrical index are used to further confirm the anomaly type.

[0067] like Figure 4-5 As shown, Figure 4 The system takes drift direction, drift intensity, orifice current, number of events, and noise as inputs and outputs orifice wall contamination, partial orifice blockage, complete orifice blockage, transient interference, and readout anomalies. Figure 5 Arranged from low to high recovery strength: buffer flushing, reducing sample concentration, low-amplitude pulse, reverse voltage, local flushing, strong pulse, cleaning reagent, channel isolation, and switching to a spare channel.

[0068] Furthermore, when the overall drift score Greater than the first abnormal threshold Furthermore, the increase in template related to pore wall contamination Greater than the template increment threshold for hole wall contamination Simultaneous opening current Relative to healthy opening current When the trend is continuously decreasing, it is determined that the nanopores are contaminated with pore walls; the corresponding graded recovery measures include buffer rinsing, reducing sample concentration, low-amplitude pulse voltage, short-term bypass, or a combination thereof.

[0069] When the overall drift score Greater than the first abnormal threshold Furthermore, the templates related to some plugging holes have increased in number. Greater than the incremental threshold of partial plugging template At the same time, the number of events per unit time Below the number of health events When the preset ratio is reached, it is determined that the nanopores are partially blocked; the corresponding graded recovery measures include reverse voltage, pulse voltage, local flushing, reducing flow rate, or a combination thereof.

[0070] Number of events per unit time Or the current through the hole Relative to healthy opening current Below the threshold for the proportion of fully plugged current If this state is maintained for a preset number of windows, it is determined that the nanopore is completely blocked; the corresponding graded recovery measures include multiple rounds of reverse voltage, strong pulse voltage, rinsing with cleaning reagent, pore isolation, switching to a backup pore or a combination thereof.

[0071] Specifically, for any set of templates Its cumulative percentage is: (17); in, The template set for the i-th channel within the time window t The cumulative percentage, This represents the proportion of the i-th channel within the k-th pose template in the time window t.

[0072] The template increment related to well wall contamination is: (18); in, This represents the set of templates related to pore wall contamination. The increment of the cumulative proportion of the contamination-related posture template set of the i-th channel within the time window t relative to the baseline of the healthy state is given. Let be the cumulative percentage of the set of attitude templates related to the contamination of the i-th channel within the time window t. This represents the baseline cumulative percentage of the set of attitude templates related to pore wall contamination under the healthy state of the i-th pore.

[0073] The incremental values ​​for some plugging-related templates are: (19); in, This is a collection of templates related to partial hole plugging. The increment of the cumulative proportion of the partial blockage-related attitude template set of the i-th channel within the time window t relative to the healthy state baseline. This represents the cumulative percentage of the set of attitude templates related to partial hole blocking for the i-th channel within the time window t. This represents the baseline cumulative percentage of the set of attitude templates related to partial blockage of the i-th channel in a healthy state.

[0074] When L consecutive windows satisfy the following formula, it is determined that the hole wall is contaminated: (20); in, The first abnormal threshold, The threshold for template increment due to pore wall contamination. This is the threshold value for the opening current deviation.

[0075] When L consecutive windows satisfy the following formula, it is determined that the hole is partially blocked: (twenty one); in, This is the incremental threshold for partially plugging the template. This is the threshold value for the opening current deviation.

[0076] A candidate state for complete hole plugging is determined when one of the following conditions is met: (twenty two); in, This is the minimum threshold for the number of effective blocking events allowed per unit time under the complete hole-blocking candidate state. The threshold value for the ratio of the opening current to the healthy opening current under fully blocked conditions.

[0077] If equation (22) is accompanied by If a valid blocking event cannot be recovered after L consecutive windows, it is determined that the hole is completely blocked. in, Noise deviation threshold Let be the baseline noise level of the i-th channel in a healthy state.

[0078] If there is only a short window abrupt change and it can regress on its own, it is judged to be a bubble or particle transient interference; if the noise is abnormal but the attitude template distribution has no corresponding drift, it is judged to be a readout link abnormality or external electrical disturbance. If the orifice current, number of events, baseline noise, or attitude template distribution abruptly changes within a short time window and recovers to a healthy range on its own without implementing recovery measures, it is identified as a bubble or particle transient interference. The aforementioned transient interference and readout anomaly detections correspond to different anomaly branches. Transient interference manifests as a sudden change in electrical parameters and / or attitude template distribution within a short period, followed by recovery. Readout link anomalies manifest as an abnormal increase in noise, but no corresponding drift in the attitude template probability distribution. If the anomaly lasts only a short time... A window, and then satisfying: (twenty three); in, To restore the threshold, The length of the time window for determining transient interference; This is then determined to be transient interference from bubbles or particles.

[0079] like but and If so, it is determined that the readout link is abnormal or there is an external electrical disturbance.

[0080] like Figure 6 As shown, Figure 6 Display before, during, and after recovery and It also specifically marks the implicit unrecovered state as "the hole current has been restored but the template drift has not been restored".

[0081] Furthermore, if the recovered opening current It meets the healthy aperture current range, but the recovered attitude template probability distribution is relative to the baseline of the healthy attitude template probability distribution. drift intensity Still above the recovery threshold If the anomaly is not found, the nanopore will be marked as a latent unrecovered state, and corresponding recovery measures will be implemented according to the current anomaly type, such as reducing the channel weight or isolating the channel.

[0082] like Figure 7 As shown, Figure 7 Display the health status of multiple channels , channel weight The process of prioritizing the participation of deweighted channels, isolated channels, and healthy channels in the overall calculation.

[0083] When the nanopore is a channel in a nanopore array, the overall drift score of each channel is used. The health status of the orifice is calculated based on the type of abnormality, the number of recovery attempts, and the recovery success rate. And based on the health of the duct Update channel weight Alternatively, low-health channels can be isolated, allowing array detection results to be preferentially used by healthy or high-weight channels in the calculation.

[0084] Specifically, the recovery time window is denoted as The conditions for successful recovery are: (twenty four); And at least two of the following conditions must be satisfied by formulas (25)-(27): (25); (26); (27); in, These represent the recovery rate of the number of events after recovery, the allowable deviation rate of the opening current, and the allowable multiple of the noise, respectively.

[0085] If at least two of equations (25) to (27) are satisfied, but equation (24) is not satisfied, then it is marked as a hidden unrecovered state, and the next level of recovery measures are implemented or the channel weight is reduced.

[0086] The recovery reliability of the i-th channel is given by formula (28): (28); in, The number of successful recovery attempts up to time window t. This represents the total number of recovery attempts up to time window t.

[0087] The health of the i-th channel is given by formula (29): (29); in, Indicates the health of the orifice. This is the drift penalty coefficient. To restore reliability.

[0088] The array weight of the i-th channel is given by formula (30): (30); in, Let be the array weight of the i-th channel.

[0089] when When the channel is below the preset isolation threshold, it is downweighted or isolated.

[0090] In one embodiment, the parameter range includes: K can be 4 to 12; the number of baseline health events. It can be from 100 to 100,000; the time window can be... to The event window can hold 20 to 5000 events; the first anomaly threshold. It can be between 0.08 and 0.25; the second abnormal threshold. It can be between 0.15 and 0.45; recovery threshold It can be 0.03 to 0.15; the number of consecutive abnormal windows L can be 2 to 20; the number of transient interference windows It can be 1 to 5; the low-amplitude pulse voltage can be... to Strong pulse voltage can be to The reverse voltage can be 20% to 200% of the forward detection voltage; the pulse duration can be... to The rinsing time can range from 0.5 seconds to 300 seconds. It can be 0.5 to 1.2; It can be between 0.02 and 0.3; It can be 1.0 to 3.0; It is a positive small quantity, and its value is used to avoid division by zero or logarithmic zero value, without changing the physical meaning of the fractal rule.

[0091] This embodiment also discloses a self-diagnosis and recovery system for nanopore contamination and blockage based on attitude template drift, including: The signal acquisition module is used to acquire similar biological particles, introduce the similar biological particles into the nanopores in a healthy state, and generate a blocking event. The posture template recognition module is used to classify the posture features of the blocking event into multiple posture template categories to obtain a real-time posture template probability distribution. The anomaly classification and recovery decision module is used to calculate the drift vector and drift intensity of the real-time posture template probability distribution and the baseline of the healthy posture template probability distribution, in order to determine the anomaly type of the nanopore and execute the corresponding graded recovery measures.

[0092] Specifically, this embodiment also discloses a self-diagnosis and recovery system for nanopore contamination and blockage based on attitude template drift, including: It includes a signal acquisition module, a health template baseline module, an attitude template recognition module, a drift calculation module, an anomaly classification module, a recovery decision module, a recovery verification module, and a recording output module. In array scenarios, it also includes a channel health weight management module.

[0093] The signal acquisition module acquires the time series of ion currents in nanopores and detects the opening current, baseline noise, blocking events, event count, and recovery response. The healthy template baseline module obtains the baseline of the attitude template probability distribution of similar biological particles under healthy pore conditions. The attitude template identification module maps blocking events to several attitude template categories. The drift calculation module calculates the drift vector and drift intensity between the real-time attitude template distribution and the healthy template baseline. The anomaly classification module identifies pore wall contamination, partial pore blockage, complete pore blockage, transient interference, or readout anomalies based on drift direction, drift intensity, changes in opening current, noise, and event count. The recovery decision module selects recovery measures and recovery intensity based on the anomaly type. The recovery verification module determines whether the attitude template distribution returns to the healthy baseline after recovery. The pore health weight management module updates the array pore weights based on drift intensity and recovery reliability.

[0094] Machine learning models, clustering models, rule thresholds, template matching, statistical discrimination models, or combinations thereof can all be used for pose template recognition. However, the core of this invention does not rely on a specific algorithm, but rather on using the probability distribution of healthy pose templates, real-time pose template distribution, template drift, anomaly classification, hierarchical recovery, and template regression verification to form a closed loop.

[0095] like Figure 8 As shown, Figure 8 The differences in judgment capabilities between the orifice current threshold method, the event number threshold method, the simple event classification method, the ordinary orifice cleaning method, and the present invention in six scenarios: mild contamination, partial orifice blockage, complete orifice blockage, transient interference, readout anomaly, and latent unrecovered.

[0096] In one implementation, silicon nitride solid-state nanopores are used to detect the same protein particles. Under healthy conditions, the proportion of normal short-dwelling templates is 48%, normal long-dwelling templates is 26%, abnormally long-dwelling templates is 5%, and slow baseline recovery templates are 3%. After continuous detection for 20 minutes, the proportion of abnormally long-dwelling templates increases to 22%, and the proportion of slow baseline recovery templates increases to 15%, with the opening current... Healthier opening current A decrease of 8%, number of events For health event numbers 78% of the overall drift score It is higher than the first abnormal threshold. And below the second abnormal threshold The system determined the issue to be well wall contamination, not complete well blockage. The system performed a 10-second buffer rinse, reduced the sample concentration to 50% of its original level, and applied five low-amplitude pulses of 50 mV for 100 ms. After recovery... If the opening current deviation is less than 3% and the abnormally long dwell time template drops to 8%, the recovery is considered successful.

[0097] In one possible implementation, a glass nanopore is used to detect protein complexes. Under healthy conditions, the proportion of low-bandwidth retention templates is 7%, and the proportion of multilevel blocking templates is 6%. Within a certain time window, the proportion of low-bandwidth retention templates rises to 31%, and the proportion of multilevel blocking templates rises to 18%, with the number of events per unit time... Down to 40% of the opening current A decrease of 18%, overall drift score The system detected partial blockage and applied a reverse voltage at 80% of the forward detection voltage amplitude for 500ms, followed by three 100mV, 200ms pulses, while simultaneously reducing the injection flow rate. After recovery, the low-width retention time on the template decreased to 12%, and the event count recovered. 82%, The recovery was deemed successful.

[0098] In one implementation, a 20nm solid-state nanopore is used to detect extracellular vesicle samples. Three consecutive time windows of a certain pore channel satisfy... Opening current Below 50% of the baseline noise Exceeding the noise threshold, the overall drift score is... Above the second abnormal threshold The system determined the well to be completely blocked and sequentially performed buffer flushing, reverse voltage, strong pulse voltage, and cleaning reagent flushing. The opening current after the first round of recovery was... The baseline returned to a healthy range, but the abnormally long dwell time template and the slow baseline recovery template remained significantly higher than the baseline. Above the recovery threshold The system marks the orifice as hidden and unrecovered, continues low-intensity flushing, and reduces the orifice weight until... It was later restored to use.

[0099] In one possible implementation, during nanopore detection, the number of events occurring within a certain time window... Sudden drop and noise Sudden rise, However, this anomaly only lasts for one window; the next window will not see any further recovery measures implemented. Opening current and number of events The system returns to a healthy range. If the anomaly is determined to be transient interference from bubbles or particles, sample advance is paused for 2 seconds, the data in that window is marked as invalid, and strong pulse and cleaning reagent recovery are not performed.

[0100] In one possible implementation, within the same detection system, the baseline noise of a certain channel... Suddenly rises to healthy baseline noise 2.5 times that, but the real-time attitude template probability distribution With health baseline near, Opening current No sustained decline, number of events No significant decrease was observed. The system determined that the issue stemmed from a readout link malfunction or external electrical disturbance, rather than via wall contamination or blockage. The system performed a circuit self-check, shielded the data for that time period, and paused the recovery control module to prevent the accidental application of reverse voltage or strong pulses.

[0101] In one possible implementation, in a 16-channel nanopore array, the overall drift score of channel 5 is... Four consecutive windows showed a value higher than 0.25, and even after two rounds of partial plugging recovery, the value still did not fall below the recovery threshold. The system is based on The health of channel 5 decreases, and the channel weight of channel 5 is adjusted accordingly. The weight was reduced from 1.0 to 0.2. During the overall array status identification, the event statistics of channel 5 were only included in the calculation with a weight of 0.2; if the situation could not be recovered later, channel 5 was isolated to avoid abnormal template distribution from contaminating the overall identification results.

[0102] In one possible implementation, in the detection of the same protein sample, the traditional aperture current threshold method only applies to the aperture current. An alarm is triggered when the concentration drops by more than 15%, and the mild pore wall contamination described in the above-described implementation method cannot be identified; the event number threshold method misjudges sample concentration fluctuations as pore blockage; the simple machine learning event classification method can identify different event patterns, but it lacks a healthy template baseline. The regression criterion cannot determine whether the hole has been recovered. This invention utilizes the attitude template drift vector. Drift intensity The system uses a combination of abnormal classification and template regression after recovery to identify hole wall contamination when the opening current drops by only 8%, and to identify latent unrecovered state when the opening current recovers but the template does not regress.

[0103] In summary, the key technical points of this invention are: 1. The probability distribution of the posture template formed by similar biological particles in healthy channels is used as the baseline of channel health.

[0104] 2. Determine the contamination / blockage status of nanopores by the drift vector and / or drift intensity between the real-time attitude template probability distribution and the healthy baseline.

[0105] 3. Correspond the drift direction to the source of the anomaly: abnormal long dwell template and slow baseline recovery template increase the corresponding hole wall contamination, low width dwell template and multi-level blocking template increase the corresponding partial hole blockage, and near-zero event and abnormal opening current correspond to complete hole blockage.

[0106] 4. Graded recovery is performed based on the source of the anomaly, and reverse voltage, pulse voltage, buffer flushing, cleaning reagents and channel isolation are optional recovery methods, rather than the only core of this invention.

[0107] 5. The probability distribution of the recovered posture template is used to regress the healthy baseline as the priority criterion for successful recovery, and latent unrecovered states are identified.

[0108] 6. In the nanopore array, the pore health is calculated based on the drift intensity, recovery number, and recovery reliability, and the pore weight is dynamically updated.

[0109] 7. Isolate transient interference from bubbles / particles and readout anomalies from actual contamination / clogging to avoid unnecessary strong recovery operations.

[0110] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift, characterized in that, include: Collect similar biological particles and introduce them into nanopores in a healthy state to generate a blocking event. The attitude features of the blocking event are divided into multiple attitude template categories to obtain the real-time attitude template probability distribution; The drift vector and drift intensity of the real-time posture template probability distribution and the baseline of the healthy posture template probability distribution are calculated to determine the abnormality type of the nanopore and to implement corresponding graded recovery measures.

2. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 1, characterized in that, Calculating the drift vector and drift intensity between the baseline of the healthy posture template probability distribution and the baseline of the healthy posture template probability distribution includes: ； ； ； in, Let be the attitude template drift vector of the i-th channel within the time window t. Let be the attitude template drift vector of the i-th channel within the time window t. The baseline of the health posture template probability distribution for the i-th channel is... Let be the attitude template distribution distance of the i-th channel within the time window t. This represents the probability distribution of the real-time attitude template after smoothing. The baseline for the probability distribution of the healthy pose template after smoothing. It is an intermediate distribution. Let be the proportion of the i-th channel within the K-th pose template in the time window t. Let be the baseline of the probability distribution of the Kth type of healthy posture template for the i-th channel within the time window t. The divergence is Kullback-Leibler.

3. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 1, characterized in that, The abnormality type of the nanopore is determined by: If the drift vector shows a continuous increase in abnormal long dwell templates and / or slow baseline recovery templates in the attitude template category, accompanied by a slow decrease in orifice current, then an orifice wall contamination trend is determined; or when the comprehensive drift score is greater than the first abnormal threshold, and the increment of orifice wall contamination-related templates is greater than the orifice wall contamination template increment threshold, and the orifice current shows a continuous decreasing trend relative to the healthy orifice current, then orifice wall contamination is determined. If the number of low-width dwell templates and / or multi-level blocking templates in the attitude template category increases, accompanied by a decrease in the number of events, or when the comprehensive drift score is greater than the first anomaly threshold. If the incremental value of some plugging-related templates is greater than the threshold value of some plugging-related templates, and the number of events per unit time is lower than the preset ratio of the number of healthy events, then it is judged as partial plugging. If the number of events per unit time is close to zero and the orifice current deviates significantly from the healthy range, or if the number of events per unit time is not greater than the target threshold or the orifice current relative to the healthy orifice current is lower than the full orifice blockage current ratio threshold, and this state is maintained continuously for a preset window number, then it is determined to be a complete orifice blockage. If a short window abruptly changes and can regress on its own, it is judged to be a transient disturbance of bubbles or particles; If the noise is abnormal but the attitude template distribution distance does not drift accordingly, it is determined to be an abnormal readout link or an external electrical disturbance.

4. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 3, characterized in that, The overall drift score includes: ； ； ； ； in, To achieve a comprehensive drift score by integrating attitude template distribution distance and auxiliary electrical parameters, This is the non-negative weighting coefficient in the overall drift score. For the deviation of the opening current, Baseline noise bias, For the event number deviation, Let be the real-time opening current of the i-th channel within the time window t. Let be the healthy opening current of the i-th channel within the time window t. , Let represent the real-time baseline noise and healthy baseline noise of the i-th channel within the time window t, respectively. This is used to avoid small positive quantities with a denominator of zero or a logarithm of zero.

5. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 3, characterized in that, The template increments related to well wall contamination include: ； in, This represents the set of templates related to pore wall contamination. The increment of the cumulative proportion of the contamination-related posture template set of the i-th channel within the time window t relative to the baseline of the healthy state is given. Let be the cumulative percentage of the set of attitude templates related to the contamination of the i-th channel within the time window t. This represents the baseline cumulative percentage of the set of attitude templates related to pore wall contamination under the healthy state of the i-th pore.

6. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 3, characterized in that, The incremental portion of the plugging-related template includes: ； in, This is a collection of templates related to partial hole plugging. The percentage of the partially blocked attitude template for the i-th channel within the time window t is the increment relative to the baseline of the healthy state. Let be the cumulative percentage of each template in the set of attitude templates related to partial hole blocking within the time window t for the i-th channel. This represents the cumulative percentage of each template in the set of partially blocked attitude templates under the healthy state of the i-th duct.

7. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 1, characterized in that, The corresponding tiered recovery measures include: When the anomaly type is well wall contamination, a low-intensity recovery is employed; the low-intensity recovery includes buffer rinsing, reducing sample concentration, low-amplitude pulses, or short-term bypass. When the abnormality type is partial blockage, a medium-intensity recovery method is used; the medium-intensity recovery method includes reverse voltage, pulse voltage, local flushing, or reducing the flow rate; When the abnormality type is complete blockage, high-intensity recovery is used; the high-intensity recovery includes multiple rounds of reverse voltage, strong pulse, cleaning reagent flushing, channel isolation or switching to a backup channel; When the anomaly type is a bubble or particle transient interference, sampling is paused or the anomaly window is marked. When the anomaly type is a readout link anomaly or an external electrical disturbance, a circuit self-test or shielding of the abnormal time period data is used.

8. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 1, characterized in that, The method also includes: After the graded recovery measures are completed, if the recovered aperture current meets the healthy aperture current range, but the drift intensity of the real-time attitude template probability distribution after recovery relative to the baseline of the healthy attitude template probability distribution is still higher than the recovery threshold, then the nanopore is marked as a latent unrecovered state, and the recovery measures continue to be implemented. At the same time, the pore health is calculated, and the pore weight is reduced or the pore is isolated based on the pore health.

9. The method for self-diagnosis and recovery of nanopore contamination and blockage based on attitude template drift according to claim 8, characterized in that, Calculating the health of the orifice includes: ； ； in, This represents the health status of the i-th channel. This is the drift penalty coefficient. To restore reliability, The number of successful recovery attempts up to time window t. This represents the total number of recovery attempts up to time window t.

10. A self-diagnosis and recovery system for nanopore contamination and blockage based on attitude template drift, implemented according to any one of claims 1-9, characterized in that, include: The signal acquisition module is used to acquire similar biological particles, introduce the similar biological particles into the nanopores in a healthy state, and generate a blocking event. The posture template recognition module is used to classify the posture features of the blocking event into multiple posture template categories to obtain a real-time posture template probability distribution. The anomaly classification and recovery decision module is used to calculate the drift vector and drift intensity of the real-time posture template probability distribution and the baseline of the healthy posture template probability distribution, in order to determine the anomaly type of the nanopore and execute the corresponding graded recovery measures.

Images