Device and method for nanoelectrohydrodynamics-based target detection

Abstract

The present invention provides a nanoelectrohydrodynamics-based target detection device and a target detection method using same. The nanoelectrohydrodynamics-based target detection device comprises: a first channel unit having a sample inlet at one end thereof and having m sub-channels provided such that a sample supplied through the sample inlet branches and moves through the sub-channels; a second channel unit having a sample outlet provided at one end thereof to discharge the supplied sample and having n sub-channels; and an ion-selective membrane interposed between the first channel unit and the second channel unit, wherein m and n are natural numbers greater than or equal to 2, which are equal to or different from each other.

Description

Nanoelectrohydrodynamics-based target detection device and method

[0001] The present invention relates to a nanoelectrohydrodynamics-based target detection device and method, and more specifically, to a target detection device and method that improves the precision and accuracy of target detection by utilizing multiple sub-channels.

[0002] The present invention relates to Project No. 2710003317 and Project No. 00302600, which were carried out with funding from the Ministry of Science and ICT.

[0003] A new diagnostic method with high specificity utilizing the CRISPR / Cas system, a nanoelectrohydrodynamics and gene editing technology, has recently been introduced. Material concentration technology based on the ion selectivity of nanomembranes and the ion concentration polarization (ICP) phenomenon caused by an electric field increases the concentration of the target for analysis without the addition of separate complex reagents; therefore, when combined with CRISPR, it is possible to obtain high sensitivity without amplification or enable rapid detection by significantly reducing the necessary amplification steps.

[0004] In particular, previous studies have proposed a method to detect the presence or absence of mutant genes using a dCas9 probe system. In mutant EGFR (epidermal growth factor receptor) DNA samples tagged with FAM (Fluorescein Amidite) dye, a clear positive / negative determination can be made based on the number of plugs resulting from the difference in mobility between free DNA and dCas9-bound DNA.

[0005] As shown in Figure 1, (a) when no target DNA molecule is present in the sample, one plug appears due to fluorescently labeled free DNA (negative), and (b) when a target DNA molecule is present in the sample, two plugs appear (positive) due to plug separation (also known as selective pre-enrichment) caused by the difference in mobility between free DNA and dCas9-bound DNA (Target DNA + sgRNA-dCas9). Therefore, based on ion concentration polarization, target DNA can be rapidly optically detected at low concentrations without erroneous PCR amplification.

[0006] Meanwhile, in next-generation molecular diagnostics, it is crucial to simultaneously detect multiple genes or pathogens through a single analysis (multiplexing). For example, multiplex nucleic acid detection is required to identify causative pathogens in patients with respiratory symptoms or to simultaneously screen for various mutations in tumor specimens. Although multiplexing technology is widely adopted in microfluidic platforms, it is currently not being fully utilized in ICP-based direct DNA detection due to technical limitations.

[0007] The technical problem that the technical concept of the present invention aims to solve is to provide a non-invasive liquid biopsy method that is fast, accurate, and reliable in detecting targets by integrating a CRISPR / dCas9 probe system with Ion Concentration Polarization (ICP). However, this problem is exemplary, and the technical concept of the present invention is not limited thereto.

[0008] According to one aspect of the present invention, a nanoelectrohydrodynamic-based target detection device is provided.

[0009] The above nanoelectrohydrodynamic-based target detection device comprises: a first channel section having a sample inlet at one end and m sub-channels provided so that a sample supplied through the sample inlet is branched and moves; a second channel section having a sample outlet at one end and n sub-channels provided for discharging the supplied sample; and an ion-selective membrane interposed between the first channel section and the second channel section; wherein m and n may be two or more natural numbers that are equal to or different from each other.

[0010] According to one embodiment, when an electric field is applied to the target detection device, an ion concentration polarization (ICP) phenomenon occurs in a region adjacent to the ion-selective membrane, thereby forming an ion depletion zone.

[0011] According to one embodiment, it may further include a gap portion provided to have a predetermined hollow space between the ion-selective membrane and the second channel portion.

[0012] According to one embodiment, the second channel portion may have an end open toward the hollow space.

[0013] According to one embodiment, each sub-channel provided in the first channel section and the second channel section is arranged in parallel and may have a uniform shape and dimensions.

[0014] According to one embodiment, the ion-selective membrane may be positioned to be in contact with at least one surface of the first channel portion.

[0015] According to one embodiment, the sample outlet may be formed as a single unit so that the sample passing through the n sub-channels is integrated into a single flow and discharged.

[0016] According to one embodiment, a plurality of the target detection devices are arranged in parallel, and different samples can be supplied to each of the sample injection ports.

[0017] According to another aspect of the present invention, a nanoelectrohydrodynamic-based target detection method is provided.

[0018] The above nano-electrohydrodynamic-based target detection method is a method for detecting a target using the target detection device, and may include the step of injecting a sample into the sample injection port (S10); the step of applying an electric field to the target detection device to form a concentration plug in which the substance in the sample is concentrated (S20); and the step of determining the presence or absence of a target by monitoring the movement of the concentration plug within the second channel section (S30).

[0019] According to one embodiment, the step (S30) may be a step of determining the pattern of the concentration plug moving through each of the n sub-channels of the second channel section by combining them.

[0020] According to one embodiment, the step (S30) may be a step of automatically determining the presence or absence of a target by inputting acquired fluorescence imaging data into a 3D CNN algorithm.

[0021] According to one embodiment, the 3D CNN algorithm can determine the presence or absence of a target from the movement pattern of a concentration plug over time that has been learned.

[0022] According to one embodiment, the step (S20) may be a step in which an ion depletion zone is formed by the occurrence of an ion concentration polarization (ICP) phenomenon in a region adjacent to the ion-selective membrane.

[0023] According to one embodiment, the concentration plug may be one in which a substance within the sample is separated and concentrated by being pushed out by an electrical repulsion at the boundary of the ion depletion region.

[0024] According to the technical concept of the present invention, by utilizing multiplexing technology, the accuracy of target detection can be further improved to provide more reliable statistical results and the detection limit for EGFR mutant DNA can be raised. In addition, by introducing a gap between the ion-selective membrane and the subchannel to mitigate electrical instability in the multiplexed device, a more stable and uniform plug can be formed.

[0025] The effects of the present invention described above are illustrative and the scope of the present invention is not limited by these effects.

[0026] Figure 1 is a diagram illustrating a method for detecting a target using a dCas9 probe system.

[0027] FIG. 2 is a schematic diagram showing a target detection device according to an embodiment of the present invention.

[0028] FIG. 3 is a diagram showing the detection results using a target detection device in which two 16-subchannel target detection devices are arranged in parallel according to a second embodiment of the present invention.

[0029] Figure 4 is a diagram showing the detection results for a MYD88 L265P mutation-positive sample using the device of Figure 3.

[0030] FIGS. 5 to 8 are drawings showing a target detection device and a detection result according to the first and second embodiments of the present invention.

[0031] FIG. 9 is a diagram showing the result of applying a target detection device according to the first embodiment of the present invention and a target detection method using the same.

[0032] Hereinafter, various embodiments of the present invention will be described in detail with reference to the attached drawings. The embodiments of the present invention are provided to more fully explain the present invention to those skilled in the art, and the following embodiments may be modified in various different forms, and the scope of the present invention is not limited to the following embodiments. Rather, these embodiments are provided to make the present disclosure more faithful and complete and to fully convey the spirit of the present invention to those skilled in the art. In addition, the thickness or size of each layer in the drawings is exaggerated for convenience and clarity of explanation.

[0033] Hereinafter, a nanoelectrohydrodynamic-based target detection device according to an embodiment of the present invention will be described.

[0034] FIG. 2 is an embodiment schematically illustrating a target detection device according to an embodiment of the present invention.

[0035] Referring to FIG. 2(a), a nanoelectrohydrodynamic-based target detection device (100a) according to a first embodiment of the present invention comprises a first channel section (110), a second channel section (120), and an ion-selective membrane (130).

[0036] The first channel section (110) and the second channel section (120) include a plurality of sub-channels (114, 124) that provide a path for the sample to move. For example, the first channel section (110) may have m sub-channels (114), and the second channel section (120) may also have m sub-channels (124).

[0037] A sample inlet (112) is formed at one end of each sub-channel (114, 124), and a sample outlet (122) is formed at the other end. Each sub-channel (114, 124) may be physically separated from one another or divided by a partition wall, and may have a shape formed long in one direction to facilitate the movement of the sample. That is, it is preferable that each sub-channel (114, 124) be arranged side by side to form a parallel structure.

[0038] Each sub-channel (114, 124) may have a uniform shape and dimensions. This minimizes structural deviations of the device, thereby increasing the reliability of target detection. For example, each sub-channel (114, 124) may have substantially the same diameter, thereby making the flow velocity distribution of the sample flowing through each channel uniform.

[0039] Electrodes may be disposed respectively in the first channel section (110) and the second channel section (120). For example, electrodes may be disposed respectively at the end of the sub-channel (114, 124) on the side of the first channel section (110) and the end of the sub-channel (120) on the side of the second channel section (120). Accordingly, at one end of the first channel section (110), a control voltage (V) that changes according to the control of the control unit is applied. CTRL ) can be applied. A reference voltage (V) can be applied to the second channel section (120). H ) can be applied. When an electric field is applied, an ion concentration polarization (ICP) phenomenon occurs in the region adjacent to the ion-selective membrane (130) interposed between the first channel portion (110) and the second channel portion (120), thereby forming an ion depletion zone.

[0040] The ion depletion region can be formed in a direction toward the second channel portion (120) of the ion-selective permeable membrane (130). The ion depletion region can be formed larger as the strength of the applied electric field increases.

[0041] The target detection device (100a) of FIG. 2(a) can perform the same multiple experiments at once by, for example, having a plurality of sub-channels (114, 124). For example, when the number of sub-channels (114, 124) is m, if the same specimen is introduced into the m sub-channels (114, 124) simultaneously, the effect of performing m experiments at once is obtained without the need to repeat m times.

[0042] In addition, since there is no risk of the samples mixing with each other even when multiple different samples are injected simultaneously into the sample injection ports (112) formed in multiple numbers, there is an advantage of being able to inject multiple samples simultaneously.

[0043] Referring to FIG. 2(b), a nanoelectrohydrodynamic-based target detection device (100b) according to a second embodiment of the present invention comprises a first channel portion (110), a second channel portion (120), an ion-selective membrane (130), and a gap portion (140).

[0044] The first channel section (110) may be provided with a sample inlet (112) at one end for injecting a sample for separation and concentration. The sample inlet (112) may be formed as a single unit so that the sample is injected into the first channel section (110) in a single flow.

[0045] The first channel section (110) and the second channel section (120) include a plurality of sub-channels (114, 124) that provide a path for the sample to move. The sub-channels (114, 124) may be physically separated or divided by a partition (150) and may have a shape formed long in one direction to facilitate the movement of the sample. That is, it is preferable that each sub-channel (114, 124) be arranged side by side to form a parallel structure.

[0046] Each sub-channel (114, 124) may have a uniform shape and dimensions. This minimizes structural deviations of the device, thereby increasing the reliability of target detection. For example, by having each sub-channel (114, 124) substantially the same diameter, the flow velocity distribution of the sample flowing through each channel becomes uniform, and the phenomenon of the flow rate being concentrated in only a specific channel can be prevented.

[0047] In the first channel section (110), m sub-channels (114) may be formed, and a sample may branch out and move through the m sub-channels (114). The second channel section (120) may have n sub-channels (124) formed therein.

[0048] In this case, m and n are natural numbers greater than or equal to 2, such as 4, 8, 16, and 32. A larger number of subchannels is advantageous for high-speed, high-volume analysis of samples, improves statistical reliability, and minimizes the impact on overall performance even if some channels are clogged or defective. However, the manufacturing process of the device is complex, and it may be difficult to uniformly control the electric field and fluidity between channels. Therefore, it is desirable to appropriately adjust the number of subchannels (m, n) according to the size of the device, the manufacturing process, and the type of target.

[0049] The above m and n may be the same or different, but it is desirable for them to be the same for the convenience of device design.

[0050] Electrodes may be disposed in the first channel section (110) and the second channel section (120), respectively. Accordingly, a control voltage (V) that changes according to the control of the control unit is provided at one end of the first channel section (110). CTRL ) can be applied. A reference voltage (V) can be applied to the second channel section (120). H) can be applied. When an electric field is applied, an ion concentration polarization (ICP) phenomenon occurs in the region adjacent to the ion-selective membrane (130) interposed between the first channel portion (110) and the second channel portion (120), thereby forming an ion depletion zone.

[0051] As the number of subchannels (114, 124) increases, active electric convection may be formed near the ion-selective membrane (130). Backflow in the opposite direction may occur at the interface of the ion-depleted region, and consequently, vortices (eddies) may be generated. Since the characteristic length scale is one of the most important variables determining the stability of the target detection device, inherent electroconvective instability may be a problem. In particular, random vortices induced by changes in local ion concentration can have a significant effect on plug formation and its reproducibility.

[0052] In order to mitigate electrical instability in a multi-channel device, a gap portion (140) may be formed between the ion-selective membrane (130) and the second channel portion (120). By making the ion depletion region have a hollow space (empty space) through the gap portion (140), the selective membrane (130) and the second channel portion (120) are spaced apart from each other, and target substances can be continuously separated and concentrated along the boundary of the ion depletion region.

[0053] If there is no gap section (140), it becomes difficult to separate and concentrate the target substance, resulting in uneven movement of the concentration plug and a significant deviation in the distance traveled. In the second embodiment of the present invention, by providing the gap section (140), the concentration plug moves more uniformly, and the standard deviation of the plug's distance traveled can be significantly reduced. Additionally, while the influence of the vortex is strong near the ion-selective membrane (130), the influence of the vortex can be reduced as it approaches the second channel section (120).

[0054] When multiple samples are injected into the sample injection ports (112) formed in multiple ways, there is a risk that the samples will mix in the gap section (140), so it is preferable to inject a single solution.

[0055] The ion-selective membrane (130) may be a cation-permeable membrane or an anion-permeable membrane. In one embodiment, the ion-selective membrane (130) may be Nafion. The ion-selective membrane (130) may be positioned to be in contact with at least one surface of the first channel portion (110).

[0056] The second channel section (120) may have an end open toward the hollow space. The sample outlet (122) formed at one end of the second channel section (120) is preferably formed as a single unit so that the sample passing through n sub-channels (124) is integrated into a single flow and discharged.

[0057] As a preferred embodiment of the present invention, a single device capable of analyzing multiple samples simultaneously can be implemented by integrating multiple target detection devices described above.

[0058] In one embodiment, a plurality of target detection devices may be arranged in parallel and manufactured so that different samples are supplied to each sample injection port. For example, two target detection devices, a first target detection device and a second target detection device having 16 sub-channels, may be arranged in parallel, and by injecting sample A into the first target detection device and sample B into the second target detection device, it may be possible to analyze samples A and B simultaneously.

[0059] The first target detection device and the second target detection device do not share a sample injection port, so that all analysis targets can be injected individually into each device, and a single solution can be measured four times repeatedly.

[0060] In addition, ion concentration polarization phenomena can be observed in a total of 32 channels with a single sample injection. This is because, due to the characteristics of microfluidic channels, channels can become clogged by dust or particles of unknown origin; however, by observing the concentration phenomenon occurring in a large number of channels, it enables statistical analysis even without concentration failures.

[0061] According to another aspect of the present invention, a nanoelectrohydrodynamics-based target detection method is provided. Hereinafter, a nanoelectrohydrodynamics-based target detection method according to one embodiment of the present invention will be described.

[0062] The above nano-electrohydrodynamic-based target detection method is a method for detecting a target using a target detection device (100) according to the embodiment of the present invention described above, and may include the steps of: injecting a sample into a sample injection port (S10); applying an electric field to the target detection device to form a concentration plug in which a substance in the sample is concentrated (S20); and monitoring the movement of the concentration plug within a second channel to determine the presence or absence of a target (S30).

[0063] First, a sample is injected through the sample injection port (112) (S10).

[0064] Next, an electric field is applied to the target detection device to form a concentrated plug in which the substance in the sample is concentrated (S20).

[0065] Step (S20) may be a step in which an ion depletion zone is formed and a concentration plug is formed by the occurrence of an ion concentration polarization (ICP) phenomenon in a region adjacent to the ion-selective membrane (130).

[0066] In this case, the concentration plug may be a substance within the sample that has been separated and concentrated by being pushed out by electrical repulsion at the boundary of the ion depletion region.

[0067] Next, the movement of the concentration plug within the second channel section (120) is monitored to determine the presence or absence of a target (S30).

[0068] For example, a single enrichment plug may be separated into two plugs, one of which is a enrichment plug of a free labeled substance, i.e., a probe not bound to a target substance (hereinafter, the first enrichment plug), and the other may be a enrichment plug of a probe substance bound to a target substance (hereinafter, the second enrichment plug).

[0069] The vicinity of the ion depletion region where such a concentration plug can be generated can be observed, and an image of the vicinity of the ion depletion region can be obtained according to a preset period. For example, an image of the portion where the sample is being concentrated within the second channel portion (120) can be obtained using an optical device such as a microscope, and an image of the portion where the sample is being concentrated can be obtained by capturing an image in real time from this image. When a fluorescent concentration target is concentrated to a concentration above a certain level, the location of the plug can be identified in real time through an optical system.

[0070] Alternatively, by inputting fluorescence imaging data acquired via a microscope into a 3D CNN algorithm, it is possible to implement automatic classification of the presence or absence of dCas9-target coupled signals. This eliminates judgment errors caused by user subjectivity and enables the AI ​​to recognize patterns and detect even weak positive signals, even at low signal-to-noise ratios. In other words, it can learn and utilize subtle spatiotemporal changes over time, such as plug movement, diffusion, and pattern segmentation. For instance, even in images that appear similar to negative to the naked eye, the CNN can detect signs of delayed movement or segmentation at the tip of the plug to classify them as positive.

[0071] At this time, the reliability of the monitoring results may be reduced due to random noise and experimental hands-on errors. For example, a single concentration plug may not be separated into a first concentration plug and a second concentration plug, or separation may not occur clearly. Accordingly, according to an embodiment of the present invention, if the device is designed to have n sub-channels (124) in the second channel section, the accuracy and reliability of the detection results can be secured. For example, since there are many sub-channels that can be measured repeatedly, the result values ​​are averaged, and even if some channels are blocked or defective, the impact on the overall performance is not significant. Therefore, it is desirable to determine the presence or absence of a target by synthesizing the movement patterns of the concentration plug through each of the n sub-channels (124).

[0072] Thus, the nanoelectrohydrodynamic-based high-concentration detection method of a target according to an embodiment of the present invention presents a noninvasive liquid biopsy method designed for rapid and accurate molecular detection in bloodstream biomolecules such as DNA.

[0073] Experimental examples to aid in understanding the present invention are described below. These experimental examples are provided to aid in understanding the present invention, and it should be understood that the present invention is not limited to these experimental examples.

[0074] Experimental Example 1

[0075] As a target detection device, two 16-subchannel target detection devices according to the second embodiment were arranged in parallel (Fig. 3(a)).

[0076] To verify whether single nucleotide mutations could be distinguished in multiple formats, two well-known cancer-related mutations were selected as experimental subjects: EGFR L858R (a substitution mutation in the EGFR gene that confers resistance to lung cancer treatments) and MYD88 L265P (a B-cell malignancy-associated repeat mutation found in more than 90% of all cases of Waldenstrup macroglobulinemia). CRISPR dCas9 RNPs specific to the DNA sequences of each mutation were designed. To confirm the presence and enrichment of each analyte, dCas9 RNPs were tagged with mCherry, and DNA samples (mutation-positive or wild-type negative) were tagged with FAM dye.

[0077] Fluorescence bands began to be observed within about 30 seconds after applying an electric field, and separation and concentration were observed for up to 30 minutes.

[0078] Figure 3(b) shows the results of nanoelectrohydrodynamic enrichment for the detection of EGFR mutations. It shows fluorescence images taken 30 minutes after the start of ICP separation enrichment for a positive sample containing the EGFR L858R mutant gene and dCas9, and a negative sample containing the EGFR wild-type gene and dCas9 RNP. Green fluorescence represents DNA labeled with FAM (for visualization purposes), and red represents dCas9 RNP labeled with mCherry. Although there is channel-specific variation in the location where enrichment proceeds within each subchannel, generally, in the positive sample, two enrichment plugs are observed in each fluorescence channel when observing the mCherry fluorescence signal and the FAM fluorescence signal, respectively, while in the negative sample, one enrichment plug is observed per subchannel in each fluorescence channel. This confirms that the pattern is identical to that of previous studies confirming the presence or absence of target genes through DNA separation enrichment, where two DNA enrichment plugs were observed in the positive sample and one in the negative sample.

[0079] Upon magnifying the subchannels (Fig. 3(c)), the two DNA enrichment plugs in the positive sample are relatively far apart (yellow and green dotted lines), whereas the dCas9 RNP enrichment plugs are closely attached (orange and yellow dotted lines). In this case, among the two dCas9 enrichment plugs, only the right-hand enrichment plug overlaps with the DNA enrichment plug, confirming that this is the region where molecules formed by the binding of dCas9 RNP and the target DNA, EGFR mutant DNA, are enriched. In contrast, in the negative sample containing EGFR wild-type DNA, one dCas9 enrichment plug and one DNA enrichment plug were observed. Furthermore, analysis of the two fluorescence signals revealed that their positions did not overlap, confirming that EGFR wild-type DNA and dCas9 RNP enriched independently without binding to each other. This trend was confirmed to appear repeatedly across all 16 subchannels.

[0080] When analyzing the brightness profile (A-A') of mCherry fluorescence extracted along the microfluidic channel (Fig. 3(d)), the positive sample (blue curve) shows two peaks, one representing the free dCas9 RNP enrichment group and the other representing the dCas9 RNP enrichment group bound to the EGFR Mutant, while the negative sample (red curve) shows a single peak at approximately 250 μm. This indicates that dCas9 failed to bind to the wild-type sequence, resulting in the presence of only the free dCas9 RNP enrichment group.

[0081] Therefore, analysis is possible even without fluorescent labeling on the DNA, and the fluorescence originates solely from dCas9-mCherry. In this case, two dCas9 enrichment plugs are formed in positive samples, and one enrichment plug is formed in negative samples. The DNA fluorescence is merely a reference for visualizing binding in Figure 3, and the actual distinction is reliably performed using only dCas9 fluorescence.

[0082] After confirming the detection of EGFR mutations, the multi-detection performance targeting the MYD88 L265P mutation was further evaluated. The MYD88 analysis was also performed in parallel channels under the same conditions, and dCas9-mCherry RNP was designed to specifically bind to the L265P site. Figure 4 summarizes the results for MYD88 L265P mutation-positive samples. Similar to the case of EGFR, enriched free dCas9 RNP plugs that did not participate in binding with the dCas9-DNA complex were observed in the mutation-positive samples. These results were observed similarly in all 16 subchannels injected with the same sample; specifically, analysis of the fluorescence intensity profiles of channels 1 and 16 clearly shows the complex plugs (orange arrows) and free dCas9 RNP plugs (yellow arrows). However, it was confirmed that the ratio of dCas9 and DNA participating in binding was lower compared to previous studies.

[0083] In summary, the CRISPR / dCas9-ICP method was demonstrated to enable reliable binary analysis using two oncogene mutation models. Samples containing target mutations formed additional fluorescent plugs distinct from the wild type. The distance between plugs (tens to hundreds of microns) formed by nanoelectronic kinetic focusing was easily distinguishable on the fluorescence profile. Reproducibility and consistency were confirmed by performing repeated experiments in parallel channels, which compensated for the variability and channel blockage issues of the existing single-channel method. Simultaneous analysis of EGFR and MYD88 mutations was possible on the same chip, demonstrating the potential for high-multiple gene diagnosis of the target detection device of the present invention. dCas9 RNPs confer specificity, and ICP ensures sensitivity by separating and concentrating molecules within 30 minutes.

[0084] Experimental Example 2

[0085] FIG. 5 is a schematic diagram of a target detection device according to the first and second embodiments of the present invention, and the detection results according to the same are shown in FIG. 6 to 8.

[0086] As illustrated in FIG. 5, an electric field was applied to a device (a) in which no gap is formed between the ion-selective membrane (130) and the second channel (120) according to the first embodiment, and to a device (b) in which a gap is formed according to the second embodiment, and a concentration plug formed was observed.

[0087] As a result, it can be seen that the plug moves more uniformly when a gap is formed (b) compared to when no gap is formed (a) (see Fig. 6).

[0088] Referring to Fig. 7, it can be seen that the standard deviation of the travel distance in the 10-minute test can be reduced by 57% when a gap is formed (b) compared to when no gap is formed (a).

[0089] In addition, in FIG. 8, it can be seen that the speed deviation of the plug is smaller when a gap is formed (b) compared to when a gap is not formed (a).

[0090] Experimental Example 3

[0091] As an AI-based target detection device, a 4-subchannel target detection device equipped with four sample inlets according to the first embodiment was prepared (Fig. 9(a)). Since these independent channels do not share sample inlets, the solution to be analyzed can be individually injected into each channel, or a single solution can be measured four times repeatedly. In this experiment, by injecting each of the four channels with one sample, four experimental results could be obtained from a single chip.

[0092] SARS-CoV-2, which causes COVID-19, and M. tuberculosis, a type of tuberculosis pathogen, were selected as targets. Unique dCas9 RNPs were designed for each pathogen, targeting the conserved region of the N gene for SARS-CoV-2 and the rpoB sequence for M. tuberculosis.

[0093] Clinical samples pre-identified for SARS-CoV-2 via RT-PCR and for tuberculosis via GeneXpert or culture methods were processed. The validation set consisted of 40 tuberculosis samples (15 positive, 25 negative) and 50 SARS-CoV-2 samples (20 positive, 30 negative) with known clinical test results. After nucleic acid extraction, isothermal amplification, and binding with dCas9 RNP, the samples were injected into the chip. Each sample was injected into four subchannels, and fluorescence patterns were observed during ICP enrichment.

[0094] Visually, the positive sample showed a slight change in fluorescence distribution compared to the negative sample, but the difference was subtle compared to the two distinct bands observed in the single nucleotide mutation experiment (Fig. 9(b)).

[0095] To overcome this, fluorescence image sequences were classified using a trained 3D CNN. To enhance the reliability of result interpretation, a 3D Convolutional Neural Network (3D-CNN) was developed to classify fluorescence patterns as "positive" or "negative." The model's input consists of fluorescence image sequences over time, forming a 3D data block (2D space + 1D time). Hundreds of ICP image sequences targeting SARS-CoV-2 and M. tuberculosis were collected, and a training dataset was constructed by assigning positive / negative labels based on PCR. Specifically, 304 sequences were collected from SARS-CoV-2 clinical samples and 252 sequences from M. tuberculosis. For each sequence, the region of interest (enrichment area) was clipped, normalized, and interpolated, and data augmentation techniques such as rotation and intensity transformation were applied. The model consisted of an input layer of size (XXYXT), a convolutional layer for extracting spatiotemporal features, and a fully connected layer for final binary classification (Fig. 9(c)).

[0096] Supervised learning was performed using the Adam optimizer to minimize binary cross-entropy loss. Twenty percent of the total data was used for validation, and model performance was evaluated on separately stored test images. The trained model demonstrated high accuracy by effectively distinguishing subtle differences in enrichment plug formation dynamics, and in actual operation, it outputs positive / negative judgments and confidence scores for new sample images. This automated interpretation serves as a foundation for reducing subjective user intervention and expanding multiplexing capabilities. Theoretically, it is possible to identify multiple targets simultaneously or to identify specific targets based on the characteristics of plug patterns. Model training and inference were performed on an A100 GPU server using Python (PyTorch), and the trained model can be deployed on smartphone apps or small devices for point-of-care diagnosis.

[0097] The performance of deep learning models was compared by evaluating unused test data on models trained from 1 to 25 minutes from the start of full image enrichment. These deep learning models demonstrated classification accuracy of over 90%, showing superior performance compared to observers.

[0098] As shown in Fig. 9(d), in tuberculosis detection, accuracy 92.3%, precision 92.3%, recall 92.3%, and Area under curve 98% were recorded using only data observed for about 2 minutes after concentration. Excellent performance was maintained even as the time used for image analysis increased, and the highest performance was recorded in the model trained on 5 minutes of measurement data. This means that results can be obtained in a very short time without the need for a long analysis period.

[0099] Meanwhile, as shown in Fig. 9(e), SARS-CoV-2 was distinguished between positive and negative with an accuracy of over 90%. In the training model using 3- and 4-minute measurement data, accuracy and precision achieved over 80% and 100%, respectively, but recall fell short of 70%. When the training video duration was increased to 25 minutes, the highest performance was demonstrated, recording an accuracy of 93.6%, precision of 100%, recall of 88.2%, and area under curve of 98%.

[0100] According to the embodiment of the present invention as described above, multiple targets can be analyzed simultaneously with a single chip, and there is potential to implement multiplex detection of more than 100 types. The PDMS chip-based system is easy to manufacture and allows for the expectation of scaling up to mass production through parallel production, and it also possesses versatility that enables rapid response to new pathogens or mutations by simply changing the gRNA sequence. The detection sensitivity and reliability of the target detection device of the present invention are similar to those of standard PCR testing, and it can be usefully utilized in point-of-care (POC) environments.

[0101] In addition, since deep learning-based readings can reduce errors and increase sensitivity, in actual applications, samples can be processed on a compact device, targets can be concentrated via ICP, and fluorescence images can be analyzed to provide automated diagnostic results. Multiplexing expansion is also possible, allowing for the analysis of multiple infectious agents in parallel channels or the classification of multiple pathogens with a single model.

[0102] The present invention has been described with reference to the embodiments illustrated in the drawings, but this is merely illustrative, and those skilled in the art will understand that various modifications and equivalent alternative embodiments are possible therefrom. Accordingly, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.

Claims

1. A first channel section having a sample injection port at one end and m sub-channels provided so that a sample supplied through the sample injection port can branch and move; A second channel section having a sample discharge port at one end for discharging the supplied sample and n sub-channels provided; and It includes an ion-selective membrane interposed between the first channel portion and the second channel portion; and The above m and n are natural numbers of 2 or more that are equal to or different from each other, Nanoelectrohydrodynamics-based target detection device.

2. In Paragraph 1, When an electric field is applied to the target detection device, an ion concentration polarization (ICP) phenomenon occurs in the region adjacent to the ion-selective membrane, thereby forming an ion depletion zone. Nanoelectrohydrodynamics-based target detection device.

3. In Paragraph 1, A gap portion further comprising a predetermined hollow space between the ion-selective membrane and the second channel portion. Nanoelectrohydrodynamics-based target detection device.

4. In Paragraph 3, The second channel portion has an end open toward the hollow space, Nanoelectrohydrodynamics-based target detection device.

5. In Paragraph 1, Each sub-channel provided in the first channel section and the second channel section is arranged in parallel and has a uniform shape and dimensions. Nanoelectrohydrodynamics-based target detection device.

6. In Paragraph 1, The ion-selective membrane is disposed to be in contact with at least one surface of the first channel portion, Nanoelectrohydrodynamics-based target detection device.

7. In Paragraph 1, The sample outlet is formed as a single unit so that the sample passing through the n sub-channels is integrated into a single flow and discharged. Nanoelectrohydrodynamics-based target detection device.

8. A device comprising a plurality of target detection devices of claim 1, A plurality of the above target detection devices are arranged in parallel, and Different samples are supplied to each of the above sample inlets, Nanoelectrohydrodynamics-based target detection device.

9. A method for detecting a target using a target detection device according to any one of claims 1 to 8, wherein Step of injecting a sample into the sample injection port (S10); A step (S20) of applying an electric field to the target detection device to form a concentration plug in which the substance in the sample is concentrated; and A step (S30) of determining the presence or absence of a target by monitoring the movement of the concentration plug within the second channel section; Nanoelectrohydrodynamics-based target detection method.

10. In Paragraph 9, The above step (S30) is, A step of determining by synthesizing the movement patterns of each of the n sub-channels of the second channel section of the above concentration plug, Nanoelectrohydrodynamics-based target detection method.

11. In Paragraph 9, The above step (S30) is, A step of automatically determining the presence or absence of a target by inputting acquired fluorescence imaging data into a 3D CNN algorithm, Nanoelectrohydrodynamics-based target detection method.

12. In Paragraph 11, The above 3D CNN algorithm determines the presence or absence of a target from the movement pattern of a pre-learned enrichment plug over time, Nanoelectrohydrodynamics-based target detection method.

13. In Paragraph 9, The above step (S20) is, A step in which an ion depletion zone is formed by the occurrence of ion concentration polarization (ICP) in a region adjacent to the ion-selective membrane, Nanoelectrohydrodynamics-based target detection method.

14. In Paragraph 12, The above concentration plug is one in which a substance within the sample is separated and concentrated by being pushed out by electrical repulsion at the interface of the ion depletion region. Nanoelectrohydrodynamics-based target detection method.

Images