Ai-chip-on-chip, clinical prediction engine

Abstract

To provide a method for assessing a characteristic of a drug.SOLUTION: The present invention provides a method comprising the steps for: treating a tissue sample with a drug; extracting from the tissue sample at least one feature of the tissue sample as treated; providing the at least one feature to an engine; obtaining a prediction from the engine; and associating the prediction with the drug. In one embodiment, the at least one feature comprises a plurality of features, and each feature from the plurality of features is extracted from the tissue sample at a predetermined time after treating the tissue with the drug.SELECTED DRAWING: Figure 1A

Description

[Technical field]

[0001] The present disclosure relates to methods for improving the efficiency of drug development. [Background technology]

[0002] The process of developing pharmaceuticals relies on testing in two-dimensional tissue cultures and animal models. This process is highly ineffective, which leads to over 90% failure in human clinical trials. There is an unmet need for improved methods for developing and testing pharmaceuticals. Summary of the Invention [Means for solving the problem]

[0003] A preferred embodiment of the present invention provides a genomic sensing method comprising detecting the presence of an RNA molecule expressed by a biological tissue, the detecting comprising a first detection at a first time and a second detection at a second time, the second time following the first time, and the second detection being unaffected by the first detection.

[0004] Another preferred embodiment of the present invention provides a genomic sensing method comprising detecting the presence of an RNA molecule expressed by a biological tissue, the detecting comprising a first detection at a first time and a second detection at a second time, the second time following the first time, the second detection being unaffected by the first detection, and wherein the RNA molecule is not amplified.

[0005] Yet another preferred embodiment of the present invention provides a genomic sensing method comprising detecting the presence of a circular RNA molecule expressed by a biological tissue, the detecting comprising a first detection at a first time and a second detection at a second time, the second time following the first time, and the second detection being unaffected by the first detection.

[0006] Yet another preferred embodiment of the present invention provides a genomic sensing method comprising detecting the presence of a circular RNA molecule expressed by a biological tissue, the detecting comprising a first detection at a first time and a second detection at a second time, the second time following the first time, the second detection being unaffected by the first detection, and wherein the circular RNA molecule is not amplified.

[0007] Another preferred embodiment of the present invention provides a chip-on-chip device comprising an organoid chamber containing an organoid and a sensor operative to detect the presence of an RNA molecule expressed by the organoid, wherein the operation of the sensor does not affect the organoid.

[0008] Yet another preferred embodiment of the present invention provides a chip-on-chip device comprising an organoid chamber containing an organoid and a sensor operative to detect the presence of an RNA molecule expressed by the organoid, wherein operation of the sensor does not affect the organoid and the RNA molecule is not amplified.

[0009] Yet another preferred embodiment of the present invention provides a chip-on-chip device comprising: a chip including a plurality of humanoids, the chip including a plurality of organoid chambers containing a respective plurality of organoids; and a sensor operative to detect the presence of an RNA molecule expressed by one of the plurality of organoids, wherein operation of the sensor does not affect said one of the plurality of organoids.

[0010] Another preferred embodiment of the present invention provides an Organ Chip device comprising a first chamber and a second chamber, the first chamber and the second chamber being adjacent and separated by a membrane, a first tissue layer cultured in the first chamber on a first side of the membrane, and a second tissue layer cultured in the second chamber on a second side of the membrane, the membrane being in a vertical orientation.

[0011] Yet another preferred embodiment of the present invention provides a chip-on-chip system comprising: a first chip including a plurality of humanoids, each of the plurality of humanoids including a plurality of organoid chambers containing a respective plurality of organoids; a plurality of biochips operative to perform a plurality of actions on the first chip, the plurality of actions including at least one of diagnostic actions and pumping actions; and an actuator operative to interact with the microfluidic chip and the plurality of biochips.

[0012] Another preferred embodiment of the invention provides a system, wherein said pumping action comprises actively pumping and selectively routing a microfluidic circuit interconnecting two or more of said plurality of organoid chambers.

[0013] Yet another preferred embodiment of the present invention provides a system, wherein the diagnostic action includes measuring a level of a chemical in a subset of the plurality of organoid chambers by operating a respective plurality of nanosensors.

[0014] Yet another preferred embodiment of the present invention provides a chip-on-chip system, comprising a first chip and a second chip, the first chip comprises a plurality of humanoids, each of the plurality of humanoids comprises a plurality of organoid chambers containing a respective plurality of organoids. Each one of the plurality of organoid chambers comprises an independent microfluidic circulation tunnel. The second chip is operable to perform a microfluidic circulation process that interconnects the plurality of organoid chambers contained in the humanoid, and the circulation is actively pumped and selectively routed according to the command from a controller or processor.

[0015] Another preferred embodiment of the present invention provides a high throughput genomic human chip system, comprising: a first chip comprising a plurality of humanoids, each of the plurality of humanoids comprising a plurality of organoid chambers containing a respective plurality of organoids; and a biochip comprising a plurality of sensors operative to detect circular RNA molecules in each of the plurality of organoid chambers.

[0016] Yet another preferred embodiment of the present invention provides a human chip device comprising: a first chip comprising a plurality of humanoids, each of the plurality of humanoids comprising a plurality of organoid chambers containing a respective plurality of organoids, and a circulation interconnecting said plurality of organoid chambers; a second microfluidic chip comprising a plurality of sensors and operative to draw samples from said plurality of organoid chambers of said first chip and deliver said samples to said plurality of sensors; and an actuator operative to move said second chip relative to said first chip.

[0017] Another preferred embodiment of the present invention provides a human chip device comprising: a plurality of humanoids, each one of the plurality of humanoids comprising a plurality of organoid chambers containing a respective plurality of organoids; a first microfluidic chip comprising a circuit interconnecting the plurality of organoid chambers, the circuit being actively pumped, selectively routed and electronically controlled; a second microfluidic chip comprising a plurality of sensors and operative to draw samples from the plurality of organoid chambers of the first chip and deliver the samples to the plurality of sensors; and an actuator operative to move the second chip relative to the first chip.

[0018] Yet another preferred embodiment of the present invention provides a genomic human chip device comprising a first microfluidic chip comprising an organoid chamber containing organoids, and a second microfluidic chip operative to non-destructively extract single cells from the organoids.

[0019] Yet another preferred embodiment of the present invention provides a genome human chip system, comprising a first chip comprising an organoid chamber containing organoids, and a second chip operable to extract single cells from the organoids non-destructively. The second chip may extract one cell before applying a treatment to the organoids, and extract one or more other cells after applying the treatment. The system also comprises an analyzer operable to determine the correlation between the treatment and a phenotype identified from the extracted cells based at least in part on the extraction. The phenotype may be selected from the group comprising genomic data and epigenetic data.

[0020] Another preferred embodiment of the present invention provides an on-chip clinical trial system comprising a first chip comprising a plurality of humanoids reflecting a respective plurality of patients, each of the plurality of humanoids comprising a plurality of organoid chambers containing a respective plurality of organoids reflecting a respective plurality of tissues from one of the plurality of patients. The system also comprises a second chip comprising a plurality of sensors and operable to draw samples from the plurality of organoid chambers of the first chip. The second chip may also deliver the samples to a plurality of remote sensors not located on the second chip. The system also comprises an actuator operable to move the second chip relative to the first chip. The system also comprises an analyzer operable to predict the efficacy of a drug in a plurality of patients based at least in part on the first chip, the second chip, and the actuator.

[0021] Yet another preferred embodiment of the present invention provides a human chip automatic tagging method.The method includes obtaining the untreated phenotype extracted from organoid and the treated phenotype extracted from the same organoid after applying a first treatment to the organoid.The method also includes identifying the difference between the untreated phenotype and the treated phenotype.The method also includes tagging the treated phenotype based at least in part on the identified difference.The method also includes predicting the effectiveness of the second treatment to the organoid based at least in part on the tagging.

[0022] Yet another preferred embodiment of the present invention provides a human chip device comprising a plurality of humanoids, each of the plurality of humanoids comprising a plurality of organoid chambers, each of the plurality of organoid chambers containing a respective organoid. The device also comprises a circulation tunnel interconnecting the plurality of organoid chambers. The system comprises a controller or processor for controlling the pumping and routing process in the circulation tunnel. The device also comprises a sensing engine comprising a plurality of sensors and sensor sample circulation. The sensing engine operates to draw samples from the plurality of organoid chambers and deliver the samples to the plurality of sensors.

[0023] Another preferred embodiment of the present invention provides a human chip device comprising a plurality of humanoids, each of the plurality of humanoids comprises a plurality of organoid chambers, each of the plurality of organoid chambers contains a respective organoid. The device also comprises a circulation tunnel that interconnects the plurality of organoid chambers, and a controller for controlling the circulation, pumping and routing process that takes place in the circulation tunnel. The device also comprises a sensing engine that has a plurality of sensors and operates to draw samples from the plurality of organoid chambers and deliver the samples to each one of the plurality of sensors. In some cases, the plurality of humanoids is more than five humanoids. In some cases, the ratio of the volume of each of the plurality of organoid chambers to the respective organoids is less than 2, so that the organoids occupy more than half of the volume of the chamber. In some cases, the ratio of the volume of each one of the plurality of organoid chambers to the volume of the sample is greater than 1000, so that the sample occupies less than 0.1% of the volume of the chamber.

[0024] Yet another preferred embodiment of the present invention provides a human-chip system, comprising a human-chip device and a controller.The controller operates to make the first organoid and the second organoid better mimic the pharmacodynamics of their corresponding first tissue and second tissue in human body, and this includes controlling circulation, thereby delivering exposure to the first organoid to the circulation, delivering different exposure to the circulation to the second organoid, applying computational adjustment to the sensor readings of the sample from the first organoid, and applying different computational adjustment to the sensor readings of the sample from the second organoid.

[0025] Another preferred embodiment of the present invention provides a human chip system including a human chip device and a controller, the controller operative to perform an action of identifying defective organoids based at least in part on the sensing engine, and an action of controlling at least one of sensing, circulating, pumping, and routing based on said identification so as to avoid defective ones of the organoids.

[0026] Yet another preferred embodiment of the present invention provides a human chip system including a human chip device and a controller, the controller operative to perform an action of identifying a defective sensor of a plurality of sensors based at least in part on the sensing engine, and an action of controlling a sensor collection cycle based on said identification so as to avoid the defective sensor of said plurality of sensors.

[0027] Yet another preferred embodiment of the present invention provides a human chip device with a sensing engine that operates to deliver immune cells to one of a plurality of organoid chambers.This sensing engine also operates to extract a first sample of said immune cells from said one of said plurality of organoid chambers before processing, and extract a second sample of said immune cells from said plurality of organoid chambers after processing, thereby enabling the immune status of said first sample of said immune cells and said second sample of said immune cells to be evaluated.

[0028] Yet another preferred embodiment of the present invention provides a method for designing drug formulations, comprising: predicting the biological effect of each of a plurality of drug formulations; and using a human chip device to screen the biological effect of each of a subset of the plurality of drug formulations, the screening being based at least in part on the prediction. The human chip device includes a plurality of humanoids, each of the plurality of humanoids includes a plurality of organoid chambers, each of the plurality of organoid chambers contains a respective organoid. The device also includes a circulation tunnel that interconnects the plurality of organoid chambers. The device also includes a controller for controlling the circulation, pumping, and routing process. The device also includes a sensing engine that includes a plurality of sensors and operates to draw a sample from the plurality of organoid chambers and deliver the sample to each one of the plurality of sensors, and refines the prediction based at least in part on the screening. The method also includes formulating a drug based at least in part on said predicting, screening, and refining, wherein said predicting, screening, refining, and formulating steps are performed by at least one processor device.

[0029] Another object of the present subject matter is to disclose a method for evaluating a characteristic of a drug, comprising the steps of treating a tissue sample with the drug, extracting from the tissue sample at least one feature of the tissue sample as treated, providing the at least one feature to an engine, obtaining a prediction from the engine, and associating the prediction with the drug.

[0030] In some cases, the at least one feature comprises a plurality of features, each feature from the plurality of features being extracted from the tissue sample at a predetermined time after treating the tissue with the drug.

[0031] In some cases, the at least one characteristic is obtained by applying a chemical or biological test to the tissue.

[0032] In some cases, the at least one feature is obtained by capturing an image of the tissue and extracting features from the image using image analysis techniques.

[0033] In some cases, the engine is an artificial intelligence (AI) engine.

[0034] In some cases, the AI ​​engine is a convolutional neural network, a deep neural network, or a random forest engine.

[0035] In some cases, the prediction is related to a toxicity level of the drug.

[0036] Optionally, the method further comprises generating the engine, the generating step comprising: treating the plurality of tissue samples with a plurality of drugs, each drug from the plurality of drugs being associated with a property; extracting at least one feature from at least some of the plurality of tissue samples; providing the at least one feature from at least some of the plurality of tissue samples and a characteristic associated with a drug that treated the tissue samples to an AI training module; Get a trained AI engine Includes.

[0037] Optionally, the method further comprises generating the engine, the generating step comprising: treating the plurality of tissue samples with a plurality of drugs, each drug from the plurality of drugs being associated with a property; extracting at least one feature from at least some of the plurality of tissue samples; providing the at least one feature from at least some of the plurality of tissue samples and a characteristic associated with a drug that treated the tissue samples to an AI training module; Get a trained AI engine Includes.

[0038] Another object of the subject matter is to disclose a human chip device, comprising a first biological chip comprising a plurality of humanoids, each of the plurality of humanoids comprising a plurality of organoid chambers, each of the plurality of organoid chambers containing at least one organoid, and a second biological chip comprising a plurality of sensors that operate to collect measurements from the plurality of organoid chambers of the first chip.The second chip comprises a circulation channel that fluidically connects at least a portion of the plurality of organoid chambers, and an actuator that operates to move the second chip relative to the first chip.

[0039] In some cases, the device further includes a controller for controlling operation of the actuator and a memory including a set of rules for moving the second biological chip relative to the first biological chip.

[0040] In some cases, the device further comprises a pump operative to pump a circulating fluid to at least one organoid chamber of the plurality of organoid chambers.

[0041] Optionally, the device further comprises an administration port for delivering fluid to the plurality of chambers via the circulation channel.

[0042] In some cases, at least a portion of the organoid chamber further comprises a secondary circulation channel and a secondary pump.

[0043] In some cases, the device further comprises an exhaust chamber connected to the secondary circulation channel for storing fluid extracted from the corresponding organoid chamber.

[0044] In some cases, at least a portion of the organoid chamber is coupled to a chamber valve that allows or prevents the flow of contents from the circulation channel to the organoid chamber.

[0045] In some cases, at least one of the first and second biological chips is microfluidic.

[0046] In some cases, the second biological chip further includes an extraction device for extracting a single cell from an organoid located in one of the plurality of organoid chambers.

[0047] In some cases, the second biological chip further includes an analyzer operable to receive at least one cell before applying a treatment to the organoid, and at least one cell before applying a treatment to the organoid, the analyzer further operable to determine a correlation between a phenotype and the treatment, the phenotype being selected from the group including genomic data and epigenetic data.

[0048] In some cases, at least some of the plurality of sensors include nanotubes located at the bottom of the second biological chip, the nanotubes operable to extract fluid from or deliver fluid to the organoid chamber.

[0049] Optionally, the plurality of humanoids represent a respective plurality of patients; The plurality of organoid chambers contain a plurality of organoids representing a plurality of tissues of the plurality of patients; The device further includes an analyzer operative to predict efficacy of a drug in the plurality of patients.

[0050] In some cases, the device further includes an organoid lid located on top of the plurality of organoid chambers, the organoid lid having an open position through which the second biological chip can access the contents inside the organoid chambers and a closed position through which the second biological chip can access the contents inside the organoid chambers.

[0051] In some cases, at least one of the first and second biological chips is microfluidic.

[0052] In some cases, the device further includes an analyzer operative to receive at least one measurement of the cells before applying a treatment to the organoid, and at least one measurement of the cells before applying a treatment to the organoid, the analyzer further operative to determine a correlation between a phenotype and the treatment, the phenotype being selected from the group including genomic data and epigenetic data.

[0053] Optionally, the device further comprises a plurality of different second biological chips having different numbers of probes.

[0054] Optionally, the second biological chip further includes a transmitter for transmitting sensed data to a remote device.

[0055] In some cases, the second biological chip further includes a plurality of probes, each probe of the plurality of probes configured to be inserted into an organoid chamber of the plurality of organoid chambers.

[0056] In some cases, at least one probe of the plurality of probes has a nanosensor at a bottom of the probe. In some cases, at least one probe of the plurality of probes has a nanosensor at a top of the probe. [Brief description of the drawings]

[0057] The novel features believed characteristic of the invention are set forth in the appended claims, however the invention itself, together with its preferred mode of use, further objects and advantages, will best be understood by reference to the following detailed description of illustrative embodiments when read in conjunction with the accompanying drawings.

[0058] [Figure 1A] FIG. 1A presents generally a flow chart of a method for training an artificial intelligence (AI) engine adapted to predict drug signatures, according to a preferred embodiment of the present invention. [Figure 1B] FIG. 1B presents a schematic representation of an apparatus for predicting drug characteristics according to a preferred embodiment of the present invention. [Figure 1C] FIG. 1C illustrates a flow chart of one embodiment of a method for preparing classification of sensor measurements. [Figure 1D] FIG. 1D presents a schematic block diagram of a system for predicting drug signatures. [Figure 2A] FIG. 2A provides a schematic overview of a chip-on-chip system according to a preferred embodiment of the present invention. [Figure 2B]FIG. 2B presents a schematic illustration of a chip-on-chip system according to a preferred embodiment of the present invention. [Figure 2C] FIG. 2C shows a schematic diagram of an example of a diagnostic chip. [Figure 2D(1)] FIG. 2D illustrates a schematic of an example processing chip. [Figure 2D(2)] FIG. 2D illustrates a schematic of an example processing chip. [Figure 3A] 3A to 3G show schematic implementations of preferred embodiments of the present invention: Fig. 3A presents an example of a chip-on-chip implementation of the present invention (full system view). [Figure 3B] Figures 3A to 3G show schematic implementations of preferred embodiments of the present invention. Figure 3B presents an example of implementation of the present invention (chip-on-chip view). [Figure 3C] Figures 3A-3G show schematic implementations of preferred embodiments of the present invention. Figure 3C presents an example of implementation of the present invention (full chip view). [Figure 3D] Figures 3A-3G show schematic implementations of preferred embodiments of the present invention. Figure 3D presents an example of implementation of the present invention (humanoid figure). [Figure 3E] 3A to 3G are schematic diagrams showing implementations of preferred embodiments of the present invention. FIG. 3E shows an example of implementation of the present invention (organoid diagram). [Figure 3F] 3A-3G show schematic implementations of preferred embodiments of the present invention, and FIG 3F shows an example of implementation of the present invention (a chemical sensing chip). [Figure 3G] 3A to 3G are schematic diagrams showing implementations of preferred embodiments of the present invention. Fig. 3G shows an example of implementation of the present invention (circulation chip). [Figure 4] FIG. 4 shows a schematic diagram of an example of a cyclic architecture. [Diagram 5] FIG. 5 shows a schematic diagram of an example of cyclic pumping. [Figure 6] FIG. 6 shows a schematic of an example of a shunted flow. [Figure 7]FIG. 7 shows a schematic diagram of an example of on-chip scaling. [Figure 8A] Figures 8A, 8B, 8C, 8D and 8E show the architecture and preparation of organoids. Figure 8A shows the architecture and preparation process of spherical organoids. [Figure 8B] Figures 8A, 8B, 8C, 8D and 8E show the architecture and generation of organoids. Figure 8B shows the architecture of perfused organoids. [Figure 8C] Figures 8A, 8B, 8C, 8D and 8E show the architecture and generation of organoids. Figure 8C shows the stretched membrane of a perfused organoid. [Figure 8D] Figures 8A, 8B, 8C, 8D and 8E show the architecture and generation of organoids. Figure 8D shows the generation of perfused organoids. [Figure 8E] Figures 8A, 8B, 8C, 8D and 8E show the architecture and generation of organoids. Figure 8D shows an alternative generation of perfused organoids. [Figure 9A] 9A and 9B together show a schematic diagram of an example of a humanoid architecture, a circular architecture. [Figure 9B] 9A and 9B together show a schematic diagram of an example of a humanoid architecture, a circular architecture. [Figure 10A] FIG. 10A shows a schematic of an example of absorption and administration. [Figure 10B] FIG. 10B shows a schematic example of ejection. [Figure 10C] FIG. 10C shows a schematic diagram of an example of parallel ADME function. [Figure 11A] FIG. 11A shows a schematic of an example of chemical and genomic nanosensing. [Figure 11B] FIG. 11B shows a schematic of the real-time cell sensor architecture. [Figure 12] FIG. 12 shows a schematic example of a specific and non-specific sensor array. [Figure 13A] FIG. 13A shows a schematic diagram of an example of genomic nanosensing. [Figure 13B] Figure 13B shows a schematic diagram of an example of circ-RNA folding. [Figure 13C] Figure 13C shows a schematic diagram of an example of circ-RNA probe design. [Figure 13D] Figure 13D shows a schematic flow chart of the method for circ-RNA probe design. [Figure 14A] FIG. 14 shows a schematic diagram of an example of non-destructive genomic analysis. [Figure 14B] FIG. 14 shows a schematic diagram of an example of non-destructive genomic analysis. [Figure 15] FIG. 15 shows a schematic flow chart of an overview of the drug discovery process. [Figure 16A] 16A-16E show multiple organoid chambers with a circulation tunnel connecting the multiple organoid chambers, according to an exemplary embodiment of the subject matter. [Figure 16B] 16A-16E show multiple organoid chambers with a circulation tunnel connecting the multiple organoid chambers, according to an exemplary embodiment of the subject matter. [Figure 16C] 16A-16E show multiple organoid chambers with a circulation tunnel connecting the multiple organoid chambers, according to an exemplary embodiment of the subject matter. [Figure 16D] 16A-16E show multiple organoid chambers with a circulation tunnel connecting the multiple organoid chambers, according to an exemplary embodiment of the subject matter. [Figure 16E] 16A-16E show multiple organoid chambers with a circulation tunnel connecting the multiple organoid chambers, according to an exemplary embodiment of the subject matter. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0059] In the following detailed description of the preferred embodiments, reference is made to the accompanying drawings that form a part hereof, which are shown to illustrate specific embodiments in which the present invention may be practiced. It is understood that other embodiments may be utilized and structural changes may be made without departing from the scope of the present invention. The present invention may be practiced according to the claims without some or all of these specific details. For purposes of clarity, technical matters that are known in the technical field related to the present invention have not been described in detail so as not to unnecessarily obscure the present invention.

[0060] The difficulty and complexity of developing new drugs raises the need for systems and methods for evaluating the effects of drugs on humans in a shorter and easier manner, which may be evaluated using a variety of characteristics, such as, but not limited to, efficacy and safety, that predict the level of risk of developing drug-induced liver injury in humans.

[0061] One solution of the present disclosure includes obtaining a collection of classified drugs, where each drug classification is associated with one or more features. For example, the collection may include multiple drugs, each with a known toxicity level, i.e., DILIrank consists of 1,036 FDA-approved drugs divided into four classes according to their potential to cause drug-induced liver injury (DILI).

[0062] A drug may then be applied (administered) to the human tissue, as described in more detail below. The tissue may then be tested over time, for example, once per hour for a week. The measurements collected are associated with each class of specific drug applied to the body tissue being tested. The properties of the body tissue may be measured using laboratory techniques, image analysis of captured images of the tissue, etc. The body measurements are converted to digital data, and each measurement collected is associated with a specific drug and a specific body tissue. The measurements are then tagged according to the class of specific drug with which the measurement is associated.

[0063] The features extracted from the tissue measurements and the classification of the collected measurements (which matches the classification of the drug applied to the particular tissue associated with that measurement) may then be used to train an artificial intelligence (AI) engine.

[0064] The AI ​​engine may then be trained and a new drug may be applied to one or more tissue samples after the new drug is placed under development, and features of measurements of the tissue samples taken over time as the tissue is treated with the new drug may then be extracted and provided to the trained AI engine.

[0065] The AI ​​engine may then predict the classification of the new drug, for example its toxicity level.

[0066] This process has the potential to save multiple testing and analysis steps by providing an indication of a drug's effectiveness and risk at an early stage when the new drug is applied to a tissue sample rather than to a patient. This could significantly reduce the time, cost and complexity of testing new drugs.

[0067] Reference is now made to FIG. 1A, which presents a flowchart of a method for training an artificial intelligence (AI) engine adapted to predict drug signatures.

[0068] Various characteristics, e.g., drugs of toxicity level, e.g., drugs of toxicity level 1 (2), drugs of toxicity level 2 (2'), etc., up to drugs of toxicity level N (2''), may be obtained. Each drug is thus associated with a particular value, e.g., a value representing its toxicity level.

[0069] In step 8, a property such as a toxicity level may undergo biotransformation, ie, conversion into multiple characteristics.

[0070] A number of tissue samples 4 may be received for carrying out the biotransformation. In steps 12, 12' and 12", drugs 2, 2' and 2" may be applied to the tissue samples, respectively.

[0071] A plate with W wells may be used, each well containing a spheroid of human cells of an organ and a medium, such as liquid. The spheroid may consist of about C cells. Each well may have a predetermined purpose, such as control, vehicle or drug dose. The control wells may be used to normalize the feature values ​​obtained from other wells.

[0072] In the exemplary experiments, the following settings are used: D - Days 14 W - Number of wells for non-terminating data collection 96 Wt - Number of wells for end data collection 4 F - Number of features 5 S - number of samples taken per day - e.g. once a day 1 C - Number of cells in spheroids: approx. 1000 DILI - Drug-Induced Liver Injury

[0073] In steps 16, 16' and 16'', respectively, relevant features may be obtained from the processed samples.

[0074] The characteristics may include chemical or biological characteristics such as glucose, lactate, oxygen, urea, potassium, ammonia, pH, and the like.

[0075] For example, the features may be of the following types: 1. Time series - data collected on any tissue at regular intervals. The collection may be done by sensors or by non-terminating staining and automatic inspection. 2. Microscopic Images - Each tissue may be photographed using a microscope at regular intervals. 3. Endpoint – some tissues may be chosen to be "sacrificed" for data collection, which prevents the spheroids from living further.

[0076] It will be appreciated that additional or different features may also be provided based on the knowledge of the domain expert. For example, a "cell respiration type" vector may be added by modeling the relationship between glucose and lactate. If glucose consumption is reduced or stopped and lactate levels are increasing, it may be concluded that the respiration of the spheroid cells changes from aerobic to anaerobic. It will be appreciated that the relationship between different measurements may be determined within a predefined time frame, such as an hour, a day, etc., not necessarily measured at the same time. This predefined time frame may vary according to various parameters, such as the type of cell, the parameter or event being measured, etc. For example, given a window_size and a shift, the following algorithm may be used to determine the cell respiration type:

[0077] For each vector containing lactate and glucose measurements, a new vector may be created by averaging the corresponding values ​​over a rolling window of length window size. For one of them, say the lactate vector, the window may be shifted to the right by the amount of shift, i.e., taking the later measurement.

[0078] The first and second derivatives of the new vector may then be calculated.

[0079] If both second derivatives at the same index are greater than a predefined threshold and the first derivative at the same index is positive, indicating lower glucose consumption and higher lactose production, then return TRUE; otherwise return FALSE.

[0080] This algorithm may be run for all possible combinations of the following parameters (9 combinations in total): shift: {-1,0,1} and window size: {1,2,3}.

[0081] If any of the runs return TRUE, then a change in cellular respiration may be presumed to have occurred.

[0082] Biological knowledge may be added to the measured or otherwise assessed characteristics.

[0083] Features may be collected, for example, S times per day for D days, which involves collecting F features from W wells. The total number of vectors is then D * S * F * W may be also possible.

[0084] Additionally, after D days, one or more microscopic images may be collected for each of the W wells, and each image may be stored in grayscale 2048x2048 resolution.

[0085] In some embodiments, a computer vision neural network may be trained on the previously annotated images to detect the location of the spheroids in the wells in the microscopy image. The trained computer vision neural network may be used to determine the bounding box in which the spheroids are located, and the image may be cropped accordingly, e.g., resized to a resolution of 412x412.

[0086] The image may be normalized, for example, the value of each pixel after normalization may be set as follows: p[i,j]=(p[i,j]-0.5) / 0.5.

[0087] A model such as Scaling Vision Transformer may be trained to classify microscopy images into severity classes in a supervised fashion using the severity class labels. Each tissue feature is then appended with a log, which is used as input to the classification head.

[0088] In one example, the tissue sample 4 may be shear cells extracted from an unprocessed organoid, and the feature extraction may include genomic sequencing of one or more cells, such that the extracted features may include genomic expression or epigenetic profiles. In yet another example, the tissue sample 4 may be immune cells and serum, and the extracted features may include an immune profile of the cells.

[0089] It will be appreciated that each feature may be associated with a number of values ​​obtained at given times after application of the drug, thus constituting a vector of values.

[0090] Thus, the biotransformation process 8 outputs training data 20 comprising a number of data sets, classified with the same classification as the drugs associated with the samples from which the features forming said data sets were extracted. For example, a data set 24 consisting of a vector of values, classified as having a first value of the characteristic (because drug 2 applied to the associated sample was classified as having a first value), a data set 24' consisting of other vectors of values, classified as having a second value of the characteristic (because drug 2' applied to the associated sample was classified as having a second value), and so on up to a data set 24" consisting of vectors of values, classified as having the N-th value of the characteristic (because drug 2" applied to the associated sample was classified as having the N-th value). A feature may be a level of toxicity. Said vector thus has the same classification as the drugs associated with the particular body tissue analyzed during the biotransformation process.

[0091] In step 28, training data 20, including data sets 24, 24′, and 24″ with associated classifications, may be used to train an AI engine, which may be a random forest engine, a convolutional neural network (CNN), any other type of neural network, or any other type of AI engine.

[0092] The training produces a trained engine 28, such as a convolutional neural network, a deep neural network, any other type of network, or any other type of AI engine.

[0093] In the example experiments, each drug was classified using the Gradient Boosted Decision Trees (GBDT) library implemented by the XGBoost library v1.5.0 and available at https: / / xgboost.readthedocs.io / en / latest / . Brute force grid search hyperparameter tuning evaluated with 10-fold cross-validation was used with the following parameter grid: {"learning_rate":[0.05,0.10,0.15,0.20,0.25,0.30], "max_depth": [3,4,5,6,8,10,12,15], "min_child_weight (minimum number of data constituting a leaf)": [1, 3, 5, 7], "gamma": [0.0, 0.1, 0.2, 0.3, 0.4], "colsample_bytree (proportion of columns to subsample when building each tree)": [0.3, 0.4, 0.5, 0.7]} Therefore, 6 * 8 * 4 * 5 * 4=3840 models may be evaluated and the best model may be selected by the average score of the 10-fold cross-validation validation. Training lasted for 100 epochs, where the entire dataset was used for each epoch. When training GBDT using XGBoost, the num_round parameter was set to 100. The following training code was used: num_round=100 evallist=[(data_test,'eval'),(data_train,'train')] bst=xgb.train(param,data_train,num_round,evallist)

[0094] It will be appreciated that the above steps may be carried out in a preparatory phase, also called a training phase, and the steps disclosed below may be carried out in a testing phase when it is required to classify new drugs being developed.

[0095] The collected data may be split into a training set and a test set, for example in a ratio of 90% training set and 10% test set. In some embodiments, 10-fold cross-validation may be used on the training set for model selection, for example to find the best hyperparameters for each type of model.

[0096] During the testing phase, an unclassified drug 36 may be provided.

[0097] One or more tissue samples 4 may be used to perform a biotransformation process 8 as described above, including the processing or application 12 of a new drug 36 to the tissue sample, and a feature extraction process 16 from the processed tissue.

[0098] The trained engine 32 may then be applied to the extracted data, including the unlabeled data set 40 extracted from the application of a new drug to body tissue. (44 applicable engine) The trained engine 28 may then output a prediction 48, e.g., a classification for the unclassified drug 36. The prediction 48 may correspond to a provided label, e.g., toxicity level, for the drugs 2, 2', 2".

[0099] For an exemplary classification of a batch of N samples, where each sample is classified into one of M classes, the trained engine may output an N×M matrix, where each row is the class distribution of the samples and each (i,j) entry represents the probability that the i-th sample belongs to the j-th class.

[0100] The final classification may involve selecting, for each sample i, the class j such that the (i,j) entry of the matrix is ​​highest among all values ​​in the i-th row.

[0101] Thus, in step 49, a prediction is received for the sample, but may be associated with the drug applied, thereby providing a classification of the drug.

[0102] It will be understood that toxicity levels are merely an example and that the method may be applied to predict any other one or more properties of drug 36.

[0103] It will be understood that the training and testing phases may be performed by the same entity, e.g., a particular laboratory or institution. However, these phases may be performed separately by different institutions. For example, a first institution may collect data on known drugs, perform biotransformations, and train the engine. The first institution may then provide the trained engine to one or more other institutions for testing new drugs developed by the other institutions.

[0104] It will be appreciated that features may be extracted from samples prior to treatment. Additionally or alternatively, features may be extracted from a control group of tissue samples to which no treatment is applied. The features extracted from the treated samples may be compared to the features extracted from the control samples. If the features are significantly identical, e.g., the distance according to a predefined metric is less than a threshold, the feature may be skipped and not used to train the engine or to evaluate the drug classification, since it may be assumed that application of the drug does not affect the tissue.

[0105] Reference is now made to FIG. 1B, which illustrates a flowchart of one embodiment of a method for training a software model for classifying drugs, according to some embodiments of the present disclosure.

[0106] Categorized drug data may be obtained in step 50, indicating characteristics of each drug, such as the drug's toxicity level. This data may be obtained from a trusted source and may include data related to known tested drugs.

[0107] In step 52, the drug may be applied to tissue, such as, but not limited to, human tissue.

[0108] Sensor measurements may be obtained from the tissues to which the drug has been applied, at step 54. It will be appreciated that each tissue may be tested for multiple parameters at multiple time points and thus multiple data sets for each tissue may be obtained.

[0109] In step 55, additional data based on biological knowledge is extracted and may be used to enhance the measurements and to add, remove or modify some of the data.

[0110] In step 56, the dataset may be classified, i.e., associated with a classification according to the classification of the drug applied to the tissue. Thus, a vector representing measurements taken from a tissue to which a first drug has been applied is classified by the toxicity level of the first drug, etc. Classification of the dataset enables a software engine, such as an AI engine, to make decisions regarding other drugs, e.g., drugs in development, according to the classifications for known drugs applied to the body tissue. Classifications for known drugs may be obtained from public sources, such as the FDA, academic studies, etc.

[0111] A model, such as a software model, may be trained on the classified vector in step 58. The trained model may receive a vector representing a tissue to which a drug is applied and provide a classification for the drug.

[0112] Reference is now made to FIG. 1C, which illustrates a flowchart of one embodiment of a method for preparing classifications of sensor measurements, such as may be performed between steps 56 and 58 above, in accordance with some embodiments of the present disclosure.

[0113] In step 60, the categorized sensor measurements produced in step 56 may be obtained.

[0114] In step 62, the sorted measurements may be converted into ordered vectors. For example, each such vector may have a particular measurement in its first entry, another particular measurement in its second entry, etc.

[0115] In some embodiments, the vectors may be normalized.

[0116] In some embodiments, biological knowledge may be applied to the vectors in step 63 to enhance them. Thus, features may be modified, added, or removed according to domain expert knowledge. For example, one or more features may be added based on the relationships between other features, e.g., "cell respiration type" may be added by modeling the relationship between glucose and lactate. It may be determined that the respiration of spheroid cells changes from aerobic to anaerobic when glucose consumption is reduced or stopped and lactate levels are increasing.

[0117] In step 64, the distance between vectors classified into the same class due to the same property of the applied drug or biological state may be obtained. It will be appreciated that an ideal classification is one in which the distance between elements assigned to the same class is smaller than the distance between elements assigned to different groups.

[0118] Thus, irregular vectors may be filtered out in step 66, e.g., vectors whose distance from one or more other vectors assigned to the same class is greater than a threshold, or which are at least a predetermined threshold greater than the average of other distances between vectors assigned to the same class. It will be appreciated that the filtering should be performed such that too many vectors are not removed, e.g., not more than a predetermined percentage of vectors, in order to avoid creating an over-fitting situation where the model fits the training set very well but is not flexible enough to provide reliable results for other vectors during run-time.

[0119] The remaining vectors may then be used to train the modes in step 58 above.

[0120] Reference is now made to Fig. ID, which presents a schematic block diagram of a system for predicting a drug signature. It will be appreciated that the system shown in Fig. IB provides the computational aspects of the prediction. Mechanical operations such as setting up the sample, applying drugs, capturing images, applying biological or chemical tests, etc. may be controlled and performed by other components of the same system, by different systems, etc.

[0121] The system may include one or more computing platforms 80. In some embodiments, computing platform 80 may be located within a laboratory, institution, etc. However, computing platform 80 may also be located elsewhere and accessed through a network operatively connected to the laboratory or institution, such as a cloud computing network.

[0122] Computing platform 80 may include a processor 81, which may be one or more central processing units (CPUs), microprocessors, electronic circuits, integrated circuits (ICs), etc. Processor 81 may be configured to provide required functionality, for example, by loading into memory and activating modules stored on storage device 86, which are described in more detail below.

[0123] It will be understood that computing platform 80 may be implemented as one or more computing platforms, which may or may not be collocated, and which computing platforms may be in communication with one another. It will also be understood that processor 81 may be implemented as one or more processors, whether or not located on the same platform.

[0124] The computing platform 80 may also include input / output (I / O) devices 82, such as a display, pointing device, keyboard, touch screen, etc. The I / O devices 82 may be utilized to receive input from and provide output to an operator, such as, for example, setting parameters, displaying predictions such as toxicity levels of a developed drug, etc.

[0125] Computing platform 80 may include a communication device 84 for communicating with other computing platforms, such as other computing platforms in a server or cloud, via any communication channel, such as a cellular network, a wide area network, a local area network, an intranet, the Internet, etc.

[0126] The computing platform 80 may also include storage 86, such as a hard disk drive, flash disk, random access memory (RAM), memory chips, etc. In some exemplary embodiments, the storage 86 may hold program code that operates to cause the processor 81 to perform actions associated with any of the modules listed below or steps of the methods of Figures 1A, 1B, and 1C above. This program code may include one or more executable units, such as modules, functions, libraries, standalone programs, etc., adapted to execute instructions as detailed below.

[0127] The storage device 86 may include an image analysis module(s) 88 for analyzing one or more captured images of the processed sample. The images may be captured using any imaging device, such as a camera, a video camera, a hyperspectral camera, a multispectral camera, an IR camera, etc. It will be appreciated that the image analysis techniques applied may correspond to the type of image captured and the features extracted.

[0128] The storage device 86 may also include a feature extraction engine 90 for extracting one or more features from the image analysis results or from other sources such as chemical or biological tests.

[0129] The storage device 86 may include an engine training module 92 for obtaining features extracted from images, from chemical or biological tests, or from any other source, and a classification assigned to a drug associated with each feature vector, such as a toxicity level associated with a given time after application. The engine training module 92 may train one or more AI engines 98 to receive feature vectors representing the results of applying drugs under development to tissues and provide a prediction corresponding to a classification, e.g., toxicity level.

[0130] The storage device 86 may include an engine application module 94 for applying an AI engine 98 to the received feature vectors to output a prediction.

[0131] The storage device 86 may include a data and control flow module 96 for invoking relevant modules in the correct order with necessary inputs, such as obtaining feature vectors and corresponding classifications to invoke the engine training module 92, providing access to the feature vectors and an AI engine 98 to an engine application module 94 to obtain predictions, etc.

[0132] Reference is now made to FIG. 2A, which is a schematic diagram of a human chip-on-chip of the present invention used to carry out bioconversion step 8 of FIG. 1A.

[0133] Organ chips allow for the acceleration of drug development, especially when multiple organ chips are utilized. The present invention overcomes two major obstacles to dramatically accelerate drug development. First, there is currently no high-throughput and AI-integrated human-on-chip platform, i.e., a system that allows for rapid and inexpensive biological scanning of drug candidates on tens of thousands of human-on-chips. Second, there is currently no human-on-chip platform with integrated genomic expression profiling.

[0134] The present invention addresses both of these challenges. The chip-on-chip architecture of the present invention addresses the first challenge, providing a high-throughput, real-time sensing human chip platform. And nanowire-based non-amplified genomic sensing, particularly sensing of circular RNA, provides a solution to the second challenge.

[0135] Genomic-related challenges of current Organ-Chip systems include: Organ-Chip systems are very promising as they show predictive accuracy that sometimes exceeds that of animal studies. However, to date, animal studies have had one unique advantage. Animal studies have yielded a wealth of genomic data, but Organ-Chip systems, due to their miniaturization, simply cannot match that wealth of genomic data. Attempting to develop drugs without genomic data is akin to "guesswork."

[0136] Therefore, the present invention is unique in that it now provides non-destructive, real-time, non-amplified detection of circular RNA, thereby providing genetic profiles associated with most major diseases. The ability to accurately identify circular RNA in a high-throughput human on-chip platform has dramatic impact. As an example, about 330 circRNAs have been shown to be associated with 48 major diseases, including various types of cancer, Alzheimer's disease, cardiovascular disease, autism, and many others.

[0137] Current methodologies for drug development are highly inefficient and prohibitively expensive, with the cost of developing a drug currently exceeding $2 billion. The present invention provides devices and methods that dramatically improve this process.

[0138] Drug development is based on two basic tools that have not changed much for over 100 years: flat cell culture dishes and mouse experiments. Flat cell culture dishes are convenient and are thought to give researchers an early indication of the effects a candidate drug is expected to have in various tissues. Mouse experiments are then used to verify the efficacy of the drug in a living mammal that is not likely to be very different from humans.

[0139] The efficacy of this approach has been dismal: over 92% of all drugs that successfully pass both flat (two-dimensional) cell culture tests and mouse experiments fail clinical trials. And, astonishingly, of the 8% of drugs that pass clinical trials and are approved for use, roughly one-third end up being withdrawn or repurposed again due to serious side effects that none of the trials uncovered.

[0140] To understand the gravity of these numbers, consider this analogy: Imagine a world where the business of building skyscrapers had a similar success rate. You build 10 skyscrapers. And you have to calmly accept the reality that 9 out of every 10 buildings you build will collapse. Only one will survive, but you have no way of knowing which one. This is the current state of drug development.

[0141] Therefore, what the pharmaceutical industry needs is a robust, high-throughput platform for evaluating and testing drug formulations with dramatically higher predictive accuracy than the current 92% failure rate of flat tissue culture dish and mouse experiments.

[0142] The present invention provides a platform that addresses this challenge by leveraging advances in two emerging fields, Organ-Chips and Deep Learning, to overcome the fundamental scientific, technical, and architectural shortcomings of current such systems, thereby providing a high-throughput solution.

[0143] Organ Chips imply the use of miniaturized three-dimensional tissues encased in a microfluidic device where the tissue is exposed to a "blood-like" flow. Although no robust, high-throughput Organ Chip systems currently exist, they overcome the fundamental shortcomings of tissue culture and mice.

[0144] In short, the scientific problem with flat tissue culture dishes is that they are not really representative of the natural tissues from which they are derived. This is tissue that has been "immortalized," which is convenient for researchers, essentially becoming cancerous, but changing its genomic characteristics. Tissue culture dishes are flat, immobile replicas of a single cancer cell. They do not have the diversity, differentiation, morphology, and physiology of living tissues within the human body, as is now well established. Also, obviously, they completely lack the vital inter-organ interactions that occur in the body. What is the point of testing a potential drug in a "heart" tissue dish if such testing ignores how that drug is absorbed through the intestine, metabolized by the liver, filtered by the kidneys, etc.

[0145] Mouse experiments seem to overcome many of these limitations. They give us the opportunity to observe how drugs act in a living mammalian body, which is likely not highly different from mammals, and how the drug responds to all the important organ-organ interactions. But ultimately, we will have to work to develop an "aspirin" for a "headache" in humans, not mice. The results are clear: 92% of tests in mice fail to predict success in the human body. Advances over the past decade have highlighted the importance of epigenetics and non-coding regions of the genome, with human-mouse homology dropping from 85% to less than 50%. This brings us to the fact that mouse experiments are neither high-throughput nor cheap (not to mention cruel).

[0146] Therefore, Organ Chips represent the next cutting edge and are very promising. In particular, linking multiple Organ Chips together and the next generation of multi-organ chips, or "human on a chip", have recently emerged. Importantly, recent early studies have now established that linking multiple Organ Chips together actually predicts pharmacodynamics and drug toxicity in the human body in a manner that dramatically surpasses flat tissue culture and, importantly, mouse experiments.

[0147] However, currently these are delicate and complex research tools, devices that are not high throughput. Although Organ Chips may seem simple, in reality, rather than sophisticated "lab-on-a-chip", they are still very much "a-chip-in-a-big-messy-lab" in nature. Therefore, existing Organ Chip products and approaches are currently not very scalable.

[0148] The present invention provides a high throughput human chip, which can compress years of drug discovery into weeks. The device disclosed in the present subject matter is fully autonomous, unlike current solutions that attempt to connect multiple organ chips, sensing devices, etc., manually or by other robotic devices.

[0149] The objective of the present subject matter is to disclose a fully integrated multi-organ chip system with pumping circulation and full administration and excretion (ADME) capabilities that exhibits seven key unique aspects: (i) high throughput, fully automated; (ii) automated real-time "cell" nanosensor array metabolite sensing; (iii) flexible circulation routing control enabling on-chip integrated tissue scaling; (iv) automated organoid generation, validation and rerouting; (v) integrated non-destructive genomic and epigenetic analysis; (vi) real-time immune profiling; and (vii) automated self-learning system integration. These and other inventive and unique aspects are further described below.

[0150] The chip-on-chip system 100 is a device constructed according to a preferred embodiment of the present invention and includes three core elements: a multi-human chip array 110, one or more automated chips 120, and an actuator 150. These three elements and their interactions are key aspects of the uniqueness of the present invention. The term "chip" refers to a vessel capable of carrying body tissue. The chip may be of any size and shape desired by a person skilled in the art. The chip may have an opening that allows sensing of the body tissue.

[0151] For clarity, these three elements are briefly described first. The multi-human chip array 110 includes multiple "humans on chips", each of which includes multiple miniaturized "organ chips". One or more automation chips 120 may be a set of dedicated lab-on-chip devices that perform actions on the multi-human chip array 110, both diagnostic actions, such as measuring chemicals, and processing actions, such as administering nutrients or drugs. The actuator 150 is a device that moves at least one of the multi-human chip array 110 and the automation chip 120. This movement results in an interaction between the multi-human chip array 110 and the automation chip 120, for example, inserting one of the automation chips 120 into each organoid chamber of the multi-human chip array 110 to collect measurements of the organoids contained therein.

[0152] The device with the actuator 150, the multi-human chip array 110 and the automation chip 120 offers the following advantages: 1. Reduces complexity since sensors are not built around each organoid 113, but rather are "shared" by many organoids 113. This is crucial for chips that may contain preferably hundreds of humanoids 111 and thousands of organoids 113. 2. Allows easy and fast replacement of sensors since the sensors are separate from the multi-human chip array 110. 3. Minimizes sample size by minimizing the length of the microfluidic channels required to transport the samples from the organoids 113 to the sensors in the automation chip 120.

[0153] The multi-human chip 110 includes a plurality of humanoids 111, each of which includes a plurality of organoid chambers 112, each of which includes an organoid 113. Each organoid 113 is a miniaturized three-dimensional biological tissue arranged in the multi-human chip 110. The biological tissue may represent the function of an organ in the human body. In a preferred embodiment of the present invention, the plurality of humanoids 111 may include at least 300, at least 200, at least 100, at least 90, at least 80, at least 70, at least 60, at least 50, or at least 40 humanoids 111. The chip-on-chip system 100 may deliver a microfluidic flow onto the organoids 113 in each organoid chamber 112 in the humanoids 111. The chip-on-chip system 100 includes three different types of microfluidic flows, one type is delivered by the multi-human chip 110 and the other two types are delivered by the automated chip 120 as follows:

[0154] The first type of microfluidic flow is the flow within the organoid chamber 112. This is the "intra-organoid" flow that provides the flow-induced shear forces required for the vitality of the organoids 113. This flow is independent of the connections between the organoids 113 and is implemented by the multi-human chip array 110.

[0155] The other two types of microfluidic flows are provided by the automation chip 120 and are described herein below. Briefly, the second type of microfluidic flow is a flow that connects multiple organoid chambers 112 in the humanoid 111, and is therefore an "inter-organoid" flow that mimics the cardiovascular flow that connects different organs in the body. This flow is provided by one of the automation chips 120 described herein below and is intermittent. The third type of microfluidic flow is an inflow flow of nutrients and an outflow flow of waste removal, mimicking the administration and excretion (ADME) body function. This second and third flow are provided by one of the automation chips 120. Both types of flows are described herein below with respect to the automation chips 120 that execute them.

[0156] It is understood that the present invention leverages chip-on-chip architecture as a means for extreme scalability and cost reduction when performing these three different types of microfluidic flows onto the organoids 113 within the organoid chamber 112.

[0157] Another aspect of the circulation capabilities of the chip-on-chip system 100 with an emphasis on intra-organoid flow is that it may be utilized for on-chip scaling, i.e., providing different levels of nutrient and drug exposure to different tissue-organoids in order to "scale" the tissue-organoids to accurately represent the corresponding tissue in the body.

[0158] In some exemplary embodiments of the present subject matter, the automation chip 120 includes two types of chips: a set of diagnostic chips 130 and a set of processing chips 140. In some other cases, the automation chip 120 includes a single type of chip. The diagnostic chip 130 is used to collect measurements of parameters of the organoids 113. The diagnostic chip 130 includes at least one of a chemical sensing chip 131, a genomic chip 132, and a cell extraction chip 133.

[0159] The chemical sensing chip 131 operates to automatically measure chemicals in the fluid within the organoid chamber 112, thereby providing monitoring of the organoids 113 within the organoid chamber 112.

[0160] The genomic chip 132 operates to automatically measure genomic expression of RNA sequences and the presence of DNA sequences within the fluid within the organoid chamber 112, thereby providing genomic analysis of the organoids 113 within the organoid chamber 112.

[0161] The cell extraction chip 133 operates to automatically extract single cells from the organoids 113 using flow shear forces and extract these single cells from the fluid in the organoid chamber 112 that contains the organoids 113. The cell extraction chip 133 therefore provides genomic analysis of the organoids 113 in the organoid chamber 112. This allows for obtaining genomic and epigenetic profiles of the organoids 113 before and after a drug 168 is administered onto the tissue, thereby allowing for analysis of the effect of genomic expression of the drug 168 on the tissue represented by the organoids 113. The cell extraction chip 133 further allows for extraction of immune cells from the fluid in the organoid chamber 112, thereby allowing for immune profiling of the organoids 113.

[0162] The processing chip 140 is used to perform actions on the organoids 113 in the organoid chamber 112. The processing chip 140 includes at least one of an initiation chip 141, a circulation chip 142, an ADME chip 143, and a meds chip 144.

[0163] The initiation chip 141 operates to receive tissues 164 provided by a user of the system and to use these tissues to automatically create organoids 113 that mimic the organs of these tissues and place them in their respective organoid chambers 112. Identifiers of the respective organoid chambers 112 associated with each tissue are stored in memory addresses accessed by a controller that manages the processes performed by the chip-on-chip system 100. The tissues 164 include, by way of example only, lung, intestine, liver, heart, kidney, capillary epithelium, blood-brain barrier, brain, and various other tissues.

[0164] The circulation chip 142 is operable to actuate "inter-organoid" microfluidic flows, i.e., flows connecting different organoids 113 within the humanoid 111, that mimic the cardiovascular flows connecting different organs within the body. As mentioned above, the circulation flows provided by the initiation chip 141 are separate from and in addition to the "intra-organoid" flows that deliver separate flows into each organoid chamber 112.

[0165] The circulation chip 142 also operates to selectively connect and direct the flow between certain organoids 113 (technically, this flow is to the organoid chamber 112 that each "contains" these organoids 113) while bypassing other organoids 113 (by bypassing the respective organoid chambers 112 of the other organoids 113) by delivering a shunted flow. As an example, the circulation chip 142 may be used to mimic the flow passing only through the intestinal and liver organoids 113 while bypassing all other organoids 113 in the humanoid 111, thereby mimicking the first-pass blood circulation in the body.

[0166] The ADME chip 143 is operable to perform functions that mimic absorption (or administration) and excretion in the human body. The acronym ADME stands for Absorption / Administration, Distribution, Metabolism, and Excretion. Of these four functions, the ADME chip 143 is operable to actuate absorption and excretion, since the other two functions, distribution and metabolism, are provided by other elements of the chip-on-chip system 100.

[0167] The ADME chip 143 performs an absorption / administration function by directing nutrients and other reagents into specific organoid chambers 112. The nutrients may then be circulated through the circulation of the humanoid 111 (done by the circulation chip 142).

[0168] The ADME chip 143 also performs an excretory function by flushing waste products, particularly perfusion-type waste products, from the organoid chamber 112. As an example, it removes the "urine" produced by the nephron-perfused organoids 113.

[0169] The actuator 150 may be a robotic device that operates to automatically perform interactions between the multi-human chip 110 and one or more automated chips 120. In a preferred embodiment of the present invention, the actuator 150 preferably includes at least one of the following components: a controller 151, a precision robot 154, a circulation pump 155, an incubator 156, a set of reagent chambers 157, and a set of drug chambers 158.

[0170] The actuator 150 preferably includes a controller 151, preferably a printed circuit board or a computer, which controls the automated operation of the actuator 150. The controller 151 may be located remotely from the actuator 150 and communicate with the actuator 150 via wired or wireless communication. The controller 151 may receive as input a test plan 160, which is stored in a memory coupled to the controller 151 as a stored test plan 152. The controller 151 may provide an output of stored results 162 in a memory accessible to the controller 151 as stored results 153. The controller 151 commands the actuator 150 to automatically execute the instructions of the stored test plan 152. The controller 151 may receive input from the diagnostic chip 130 and store the input in the stored results memory 153.

[0171] The precision robot 154 may be a robotic device that performs all operations of the chip-on-chip system 100 by automatically and cooperatively operating the multi-human chip 110 and the automation chip 120 according to the instructions of the controller 151. The precision robot 154 is preferably constructed to operate with sub-millimeter precision, as an example, with ten microns (10 μm) precision, or with fifty microns (50 μm) precision, or with one hundred microns (100 μm) precision. This precision is required to insert a capillary or probe of the automation chip 120, which may preferably be fifty microns (50 μm) or one hundred microns (100 μm) in diameter, into the organoid chamber 112 of the multi-human chip 110. The organoid chamber 112 may have a size of 1 mm in diameter or less. In a preferred embodiment of the present invention, multiple sensor probes may be inserted simultaneously into the same organoid chamber 112, which requires even higher precision. In a preferred embodiment of the present invention, the precision robot 154 inserts such probe(s) into the organoid chamber 112 of the multi-human chip 110. Multi-human chip Cameras and image processing techniques may be utilized to achieve such micron-level precision movement when inserted into the organoid chamber 112.

[0172] In a preferred embodiment of the present invention, the precision robot 154 also automates the workflow of an experiment in which multiple multi-human chips 110 are used. The number of multi-human chips 110 that one actuator 150 can handle depends on at least one of the following characteristics: (i) the frequency of the operations (such as sensing and cycling), (ii) the number of humanoids 111 and organoids 113 in the multi-human chip 110, (iii) the number of parallel operations that the automated chip 120 can perform (for example, how many sensor probes are inserted into adjacent organoid chambers 112 at the same time), (iv) optimization that allows reducing the number of operations based on data analysis and smart sample testing, etc. As an example, in an exemplary embodiment of the present invention, a single precision robot 154 may preferably handle a large number of multi-human chips 110, the number of which may be selected from at least 9, at least 10, at least 15, and at least 20 multi-human chips 110.

[0173] In other preferred embodiments of the present invention, the precision robot 154 may operate optical analysis instruments, microscopic optical analysis, and other devices useful for analyzing the multi-human chip 110. For example, the precision robot 154 may move such instruments so that they are appropriately positioned relative to the organoid chambers 112 in the multi-human chip 110 to collect measurements of the organoids 113 in the multi-human chip 110.

[0174] The circulation pump 155 is a pump device that drives the pumping of microfluidic flows inside at least some of the organoid chambers 112 on the multi-human chip 110.

[0175] The incubator 156 is operable to maintain environmental conditions within the chip-on-chip system 100 at desired levels to ensure desired conditions for the biological and chemical operation of the organoids 113 within the multi-human chip 110, nanosensors, and other components of the chip-on-chip system 100. Such conditions include, but are not limited to, desired levels of temperature, moisture, and the presence of gases such as oxygen and CO2.

[0176] The reagent chamber 157 and the drug chamber 158 are two sets of chambers that store a plurality of reagents 166 and drugs 168, respectively, to be provided and uploaded to the chip-on-chip system 100. The reagent chamber 157 and the drug chamber 158 may be located within the actuator 150. The reagents 166 include, by way of example, a blood replacement common circulating medium, a rinsing reagent, and various other reagents. The drugs 168 include various drug formulations to be tested and may include drug formulations of different doses or concentrations.

[0177] The reagents 166 stored in the reagent chambers 157 are used by the chip-on-chip system 100 for the functioning of each humanoid 111 and each organoid 113. As an example, a circulating medium may be pumped from its corresponding one of the reagent chambers 157 to the cardiovascular-like circulation of each humanoid 111. In another preferred embodiment of the present invention, the reagents 166 may preferably include, but are not limited to, inhalation reagents that mimic air inhaled by the lungs, food ingested by the digestive system, contaminants, infectious agents, or other substances that the user wishes to affect the humanoid 111. In another preferred embodiment of the present invention, when the chip-on-chip system 100 takes (draws) a sample from the organoid chamber to be analyzed, a "rinse" reagent may be used to rinse the sample tubing, etc. In another preferred embodiment of the present invention, one or more of the additional reagent chambers 157 and additional drug chambers 158 may also be included in the multi-human chip 110 itself, such that the multi-human chip 110 may draw reagents and drugs from its own reservoirs.

[0178] The drugs 168 stored in the drug chambers 158 are used by the chip-on-chip system 100 according to a test protocol described later in this specification, whereby a desired one of the drugs 168 is pumped from a corresponding one of the drug chambers 158 at a desired dose and delivered to the organoids 113 of the appropriate humanoid 111 on the multi-human chip 110. The chip-on-chip system 100 enables inexpensive, automated, high-throughput testing of multiple drugs at different doses on multiple humanoids 111. Thus, in a preferred embodiment of the present invention, the actuator 150, together with the meds chip 144, operates to automatically administer each of the multiple drugs 168 to the respective organoid chambers 112 with the humanoids 111 on the multi-human chip 110.

[0179] In another preferred embodiment of the present invention, one of the multiple drugs 168 may be delivered to multiple humanoids 111 such that each humanoid 111 receives a different dose of this same drug 145. In another preferred embodiment of the present invention, the meds chip 144 and the actuator 150 may administer different doses of one of the multiple drugs 168 to different organoids 113 in the humanoid 111, so that different organoids 113 in the same humanoid 111 receive different doses of the drug 168, thereby more accurately mimicking the conditions in the human body. At this time, the drug 168 may be known to exist in different concentrations in different body organs. The meds chip 144 uses microfluidics to automatically deliver precise titrations (i.e., different doses) of the drug 168 to a series of different organoid chambers 112, for example, to different humanoids 111.

[0180] The deep-matching engine 170 is preferably an artificial intelligence deep learning software used to analyze the results 162 and iteratively adjust the test plan 160 accordingly, which adjustments may then be preferably fed back to the stored test plan 152 to drive the next experiment modified on the chip-on-chip system 100.

[0181] It will be appreciated that the chip-on-chip system 100 of the present invention offers several unique advantages, which are briefly outlined below.

[0182] The first advantage of the present invention is that it provides for the first time a fully automated high-throughput human-on-a-chip system. Currently, there is no single device that is both an integrated human-on-a-chip device, i.e., a single device that includes the integration of multiple miniaturized interconnected organs and mimics human absorption, circulation, metabolism and excretion, and is also a fully automated high-throughput testing platform, i.e., a testing platform that performs tests in parallel on hundreds of humanoids 111 containing thousands of organoids 113 in a fully automated manner from start to finish, while continuously monitoring and analyzing the metabolic effects on each organoid 113.

[0183] A second unique advantage of the present invention is the integrated ADME, including absorption, distribution, metabolism and excretion. Currently, there are no integrated human chips with ADME capabilities that are also automated and high-throughput.

[0184] The third unique advantage of the present invention is the production and validation of organoids. In current organ-chip systems, the process of producing organoids or miniaturized tissues is manual and time-consuming, and in the case of membrane-like perfused organoids113, the validation of tissue-membrane integrity is also manual. This limitation is manageable in single-organ chips, but becomes prohibitive when trying to scale up to fully automated high throughput, and even more so in the high throughput of multi-organ chips.

[0185] A fourth unique advantage of the present invention is on-chip integrated tissue scaling. A major limitation of existing attempts to produce effective human chips is that simply connecting multiple organ chips together does not effectively represent the human body, even if each such organ chip is effective alone. This is because each "organ chip" must be scaled differently to reflect the different mass and circulatory criteria of the different tissues. As an example, myocardial organoids 113 and liver tissue organoids 113 having similar sizes, e.g., 1000 cells, must then each be scaled very differently to adequately represent the human body, as an example, to reflect the larger mass and much greater "circulatory exposure" of liver tissue compared to myocardial tissue. The multi-human chip 110 is unique in that it provides integrated scaling through a combination of physical on-chip compensation along with computerized extrapolation correction.

[0186] A fifth unique advantage of the present invention is real-time cellular metabolic sensing. Currently, there is no high-throughput human chip that provides automated, repetitive, real-time, unbiased cellular metabolic sensing. Access to frequent repetitive metabolic measurements is important in assessing the function of tissues and their mitochondria as part of the evaluation of the pharmacodynamics and pharmacokinetics of the tested drug, as is well known in the art. Such measurements may preferably include, but are not limited to, glucose, lactate, urea, potassium, ammonia, and pH. An important unique feature of the present invention is that it incorporates an automated chip 120 (or in another preferred embodiment, on-chip peripheral sensing 155) that provides unique real-time cellular metabolic sensing.

[0187] The present invention provides the unique advantage of providing genomic expression and epigenetic analysis of each organoid 113 in each humanoid 111 in response to treatment with drug 145. Non-destructive herein means the ability to carry out genomic expression of organoid 113 without destroying organoid 113, so that, as an example, genomic expression of pre-treated organoid 113 can be compared with genomic expression of the same organoid 113 after treatment with one of drugs 168.

[0188] A seventh unique advantage of the present invention is real-time immune profiling, which provides automated testing and analysis of the immune response of each organoid 113 to one administration of a drug 168.

[0189] An eighth unique novelty of the present invention is autonomous self-learning. It is understood that current Organ Chip systems lack this capability because they lack the following features that are unique to the present invention: fully automated high throughput capabilities with automated, self-training artificial intelligence analysis with real-time cellular sensing, and an architecture that integrates all of this into a self-learning system.

[0190] Reference is now made to FIG. 2B, which provides a schematic illustration of a human chip-on-a-chip system 100 according to a preferred embodiment of the present invention.

[0191] In a preferred embodiment of the present invention, the chip-on-chip system 100 includes three main components: a multi-human chip 110 , an automation chip 120 , and an actuator 150 .

[0192] The multi-human chip 110 comprises multiple humanoids 111, each of which preferably comprises multiple organoid chambers 112 containing corresponding multiple organoids 113 of different tissue types, e.g., lung, liver, brain, etc. Figure 2BIn the example shown in Figure 1, the humanoid 111 includes a "row" of six adjacent organoid chambers 112, and the humanoid 111 includes only six humanoids (six such rows). It will be understood that the number of chambers is shown by way of example only and is not intended to be limiting. In preferred embodiments of the present invention, the multi-human chip 110 may preferably include more than 100 such humanoids 111, and each humanoid 111 may include more than six organoids 113.

[0193] Each of the organoid chambers 112 may have an associated organoid flow tube 114. The organoid chambers 112 may also include an organoid tube sac 115 located at the bottom of the organoid flow tube 114. When the organoid tube sac 115 is pressed by a peristaltic pump (not shown) operated by the actuator 150 directly below the humanoid 111, this pressing causes the organoid tube sac 115 to contract in volume, thereby pumping the fluid and causing it to flow into the organoid flow tube 114 and the organoid chamber 112. This flow causes a shear flow effect on the organoids 113, which is required for the vitality of the organoids 113.

[0194] The automated chip 120 includes a set of chips designed to perform a wide variety of actions on the humanoids 111 and organoids 113 of the multi-human chip 110. Figure 2Bshows an exemplary chip that is a "circulation chip" that pumps fluid between the organoid chambers 112 of a humanoid 111, thereby mimicking the cardiovascular circulation of the body. As shown in this figure, this automated chip 120 includes a set of capillaries, each such capillary tube extending into two adjacent organoid chambers 112, pumping fluid from one organoid chamber 112 to its neighboring organoid chamber 112 (and one longer capillary transfers fluid from the right-most organoid chamber 112 to the left-most organoid chamber, thereby closing the cycle). This circulation chip also operates to guide the routing of the circulation, for example, directing the flow to flow only through the "gut" and "liver", thus mimicking the first-pass circulation of the liver.

[0195] The actuator 150 performs the operations of the chip-on-chip system 100. The actuator 150 includes a controller 151 that commands the operation of the actuator 150, for example, based on a predetermined schedule, according to user input, etc. The actuator 150 also includes a precision robot 154 that operates the multi-human chip 110 and the automated chip 120. As shown in this figure, the precision robot 154 includes a robot arm that operates to pick up and move the appropriate one of the multiple automated chips 120 so that its capillary tube or probe (depending on the type of automated chip 120) is positioned within the set of organoid chambers 112 to perform their actions. Figure 2B As shown, in a preferred embodiment of the present invention, actuator 150 may also include a "box enclosure" that encloses all of the elements of chip-on-chip system 100 to enable actuator 150 to provide an incubator function (not shown) and maintain desired levels of properties such as temperature, humidity, etc. Figure 2BIt should be understood that although only a single multi-human chip 110 is shown within the enclosure of the actuator 150, the present subject matter also discloses embodiments in which a single actuator 150 may "house" multiple multi-human chips 110, preferably more than nine such multi-human chips 110, and is operable to handle such multi-human chips 110 simultaneously.

[0196] FIG. 2B also shows three additional chips, a chemical sensing chip 131, a genome chip 132, and a meds chip 144. The three additional chips may be part of the set of automated chips 120 and are shown by way of example only. It is understood that these chips are placed in an enclosure of an actuator 150 so that a precision robot 154 may lift the appropriate chip for the desired operation to be performed. The chemical sensing chip 131 includes a set of nanosensors and operates to perform real-time measurements of biologically relevant chemicals in multiple organoid chambers 112, preferably in parallel, in the organoid chambers 112 of a humanoid 111 as an example. The genome chip 132 operates to perform sensitive, non-amplified measurements of the expression of cell-free nucleic acids, such as RNA or DNA, in the fluid of the organoid chambers 112. The genome chip 132 may also be operable to detect the expression of circulating RNA (circRNA), a type of short RNA that is highly stable outside cells and associated with many major diseases, including cancer, heart disease, and autism. The meds chip 144 is an automated chip 120 that operates to administer drugs to the organoid chamber 112.

[0197] It is understood that the key uniqueness of the present invention is that the organoids 113 are exposed to two complementary modes of actively pumped and controlled microfluidic flow. The first mode of microfluidic flow is "intra-organoid" circulation, i.e., circulation within individual organoid chambers 112, for example through organoid flow tubes 114, causing constant shear force stimulation of the organoids 113. The second mode of microfluidic flow is "inter-organoid" circulation, which mimics the cardiovascular circulation in the body and is actuated by an automated chip 120 of "circulation" type. Furthermore, it is understood that the inter-organoid circulation is dynamically flexible and may bypass certain organoids 113 as needed.

[0198] Reference is now made to FIG. 2C, which illustrates a schematic diagram of an example of a diagnostic chip constructed in accordance with the present invention.

[0199] Figure 2C provides schematic diagrams of two examples of diagnostic chips. The first example is depicted in the figure designated "Chemical and Genomic Diagnostic Chip". The chemical sensing chip 131 is a diagnostic chip constructed and operative to perform nanosensor testing of chemicals present in the fluid of the organoid chamber 112.

[0200] The chemical sensing chip 131 includes six probes 200 that are constructed and operative to be lowered (e.g., by a precision robot 154, not shown) into the fluid of each of the six organoid chambers 112 contained within the humanoid 111. This design allows the chemical sensing chip 131 to perform simultaneous chemical testing of the six organoid chambers 112 of the humanoid 111.

[0201] A chemical nanosensor 201 is mounted on the probe 200 and operates to perform measurements of the levels of chemical compounds found in the fluid of the organoid chamber 112 and thus correlated with the organoids 113.

[0202] Each probe 200 may preferably include several chemical nanosensors 201, allowing each probe 200 to perform several chemical tests simultaneously. Figure 2A depicts three chemical nanosensors 201 attached to each of the probes 200. The number of chemical nanosensors 201 in a probe may be selected by one of skill in the art.

[0203] At least some of the organoid chambers have associated organoid flow tubes 114 and organoid tube sacs 115, both of which allow for intra-organoid flow.

[0204] This diagram is also relevant to genomic sensing, with it being understood that at least some of the chemical nanosensors 201 are nanosensors constructed and operative to identify the hybridization of specific nucleic acid molecules.

[0205] A second example is depicted in the diagram designated "Cell Extraction Chip." The Cell Extraction Chip 133 is a diagnostic chip constructed and operative to extract single cells from the fluid of the organoid chamber 112 so that these single cells may be analyzed, including by single-cell genome sequencing and epigenetic analysis.

[0206] The cell extraction chip 133 includes six cell extraction nanotubes 202 that are constructed and operative to be lowered (e.g., by a precision robot 154, not shown) into the fluid of each of the six organoid chambers 112 contained within the humanoid 111. This design allows the cell extraction chip 133 to perform simultaneous extraction of single cells from the organoids 113 within each of the six organoid chambers 112 of the humanoid 111.

[0207] Reference is now made to FIG. 2D, which illustrates a schematic diagram of an example of a processing chip constructed in accordance with the present invention.

[0208] FIG. 2D includes four diagrams showing four examples of processing chips 140, designated "start chip," "circulation chip," "ADME chip," and "meds chip," respectively.

[0209] The starting chip 141 is constructed and operative to automatically create and place organoids in the organoid chambers 112. The starting chip 141 includes a plurality of placement nozzles 203 corresponding to a plurality of organoid chambers 112 of the humanoid 111, and the starting chip 141 can place organoids 113 in each organoid chamber 112. While FIG. 2B depicts the creation of spheroid organoids, this is not intended to be limiting, and it is understood that the starting chip 141 is also operative to create perfused organoids.

[0210] The circulation chip 142 is constructed and operative to perform circulation of fluids between organoids 113 disposed in organoid chambers 112 of the humanoid 111, mimicking the cardiovascular circulation in the body. The circulation chip 142 provides circulation to the humanoid 111 having multiple organoid chambers 112. The circulation chip 142 includes multiple circulation tubes 204 and one long circulation tube 205. The multiple circulation tubes 204 transfer fluids from one organoid chamber 112 to an adjacent organoid chamber 112, and the long circulation tube 205 transfers fluids from the rightmost organoid chamber 112 of the humanoid 111 to the leftmost organoid chamber 112. It is understood that the circulation provided by the circulation chip is an inter-organoid flow, i.e., a flow of fluids between different organoid chambers 112 of the humanoid 111.

[0211] ADME Chip 143 is a humanoid 207 The ADME chip 143 is constructed and operative to perform the functions of administration / absorption and excretion on the organoid chamber 112 of the humanoid. 207The ADME tubes 206 correspond to the organoid chambers 112 of the humanoid. 207 The organoids 113 are lowered (e.g., by a precision robot 154, not shown) into the fluid of each of the organoid chambers 112 and operate to (i) administer nutrients and (ii) remove waste products from the organoids 113.

[0212] meds chip 144 is a humanoid 210 The meds chip 144 is constructed and operative to administer a drug 208 to the organoid chamber 112 of the humanoid. 210 The organoid chamber 112 includes a plurality of meds tubes 209 corresponding to the plurality of organoid chambers 112. The meds tubes 209 include a humanoid 210 The meds chip 144 is operated to be lowered (e.g., by a precision robot 154, not shown) into the fluid of each organoid chamber 112 to administer the drug 208 to the organoids 113. It is also understood that the meds chip 144 is also operated to automatically deliver different doses of the drug 208 to the different organoids 113.

[0213] Reference is now made to FIG. 3A, which shows an example of a schematic implementation of the chip-on-chip of the present invention. FIG. 3A shows a prototype in production of one embodiment of the chip-on-chip system 100 of the present invention, revealing its main components, the multi-human chip 110, a set of automated chips 120, and an actuator 150, and their close interaction. The chip-on-chip system 100 includes a body configured to be placed on a surface such as a floor, a table, etc. The body may include a frame carrying an actuator 150, for example a robotic arm that moves the automated chip 120 along the line of organoid chambers included in the multi-human chip 110.

[0214] Reference is now made to FIG. 3B, which illustrates an example of an implementation of the present invention, showing its chip-on-chip embodiment and its main components: a multi-human chip 110, a set of automation chips 120, a chemical sensing chip 131, a circulation chip 142, an actuator 150, and a calibration and rinsing well 300. It is shown that the actuator 150 may carry the automation chip 120. The automation chip 120 may be coupled to the actuator 150, for example, using a magnetic field that attracts the frame of the automation chip 120 to the body of the actuator 150, or using another connector that can be controlled by the controller of the chip-on-chip system 100.

[0215] It is understood that the calibration and rinsing well 300 may be utilized to rinse the automated chip 120 after performing an action or measurement on one set of organoid chambers 112 and before performing a next action or measurement on the same or another set of organoid chambers 112, thereby avoiding contamination of other sets of organoid chambers 112 with fluid from the previous set of organoid chambers 112. It is further understood that the calibration and rinsing well 300 is also important to enable automatic calibration of the diagnostic chip 130.

[0216] In a preferred embodiment of the present invention, the calibration and rinsing wells 300 are located in close proximity to each multi-human chip 110 to minimize the distance of movement of the precision robot 154 between tests or actions performed by the automated chip 120 on the multi-human chip 110 and to allow frequent rinsing and calibration actions of the automated chip 120. In some exemplary cases, the calibration and rinsing wells 300 may move along the multi-human chip 110 along with the automated chip 120.

[0217] Reference is now made to Fig. 3C, which illustrates an example of a chip-on-chip implementation of the present invention, providing an overview of an implementation of a multi-human chip 110 and a chemical sensing chip 131, one of a set of automated chips 120 constructed and operative in accordance with a preferred embodiment of the present invention, and the interaction between these two chips.

[0218] As can be seen in Figure 3C, an exemplary implementation of the multi-human chip 110 includes 108 humanoids 111 that collectively contain 648 organoids 113. The organoids may have dimensions of 3mm x 60mm x 70mm.

[0219] As an example, one humanoid is highlighted and surrounded by a dashed line. Each humanoid contains a "row" of six interconnected "wells," each of which is understood to be an organoid chamber 112 containing an organoid 113 (not shown). As an example, one such organoid chamber 112 is also highlighted and surrounded by a dashed line.

[0220] In another preferred embodiment of the present invention, inter-organoid circulation is accomplished not by a circulation chip, but by a circulation pump and circulation channel 285, both located within the multi-human chip 110. In this embodiment, each humanoid 111 also includes a humanoid circulation pump, which pumps circulation between the organoid chambers of that humanoid, analogous to the "heart" for blood circulation in the human body. It is understood that circulation between organoid chambers 112 is not passive, but rather actively pumped, selectively routed, and electronically controlled.

[0221] Also shown is a chemical sensing chip 131 that interacts with the multi-human chip 110, one of a set of automated chips 120. It is understood that the chemical sensing chip 131 interacts with multiple organoid chambers 112 simultaneously, drawing samples from the chambers 112 and flowing them simultaneously to multiple corresponding nanosensor arrays on the chemical sensing chip 131. Samples drawn from the organoid chambers 112 are distributed to multiple arrays of nanosensors that measure multiple chemicals, for example, to indicate tissue viability, and specific desired assays. In some exemplary embodiments, each such sensor array includes seven sensors, and the seven sensors may preferably measure chemicals that indicate cellular and mitochondrial vitality. The sensors may measure oxygen, glucose, lactate, urea, potassium, ammonia, pH, and the like.

[0222] Moving the chemical sensing chip 131 relative to the multi-human chip 110 (i.e., moving it over the row of organoid chambers 112 and lowering it so that its multiple capillaries are "immersed" into the multiple organoid chambers 112 of each of the next humanoid 111 to be tested) may be performed by a precision robot 154 of the actuator 150 (both not shown), as would be known to one of skill in the art.

[0223] Reference is now made to Fig. 3D, which shows another view of an example of a chip-on-chip implementation of the present invention, providing a close-up view of an implementation of a multi-human chip 110 and a chemical sensing chip 131 (one of a set of automated chips 120) and the interaction between these two chips, constructed and operative in accordance with a preferred embodiment of the present invention.

[0224] As can be seen in this figure, Figure 3D "zooms in" to show a more detailed view of six humanoids 111, each containing six organoid chambers 112 (each containing a corresponding organoid 113 (not shown)), and shows a chemical sensing chip 131 interacting simultaneously with the six organoid chambers 112.

[0225] FIG. 3D includes three figures, which together provide insight into key features of the present invention. The first figure, designated "1. Six Tissues of a Humanoid," illustrates how the chip-on-chip system 100 of the present invention operates to perform high-throughput screening, in this example, testing six organoids 113 contained within a single humanoid 111 simultaneously, in a fully automated manner. As described above, each row of six "wells" is a humanoid 111. Thus, as this first figure shows, when the chemical sensing chip 131 tests a row of "wells" simultaneously, the chemical sensing chip 131 is testing six organoids 113 of the humanoid 111.

[0226] The second figure, designated "2. Same Tissue, Six Humanoids," illustrates how the inventive chip-on-chip system 100 operates to perform high-throughput screening. In this example, the chemical sensing chip 131 tests the same tissue of interest (e.g., "liver") from six different humanoids 111 simultaneously in a fully automated manner.

[0227] It will be appreciated that the multi-human chip 110 may also enable one to focus on "humanoid" testing (i.e., testing of all organoids 113 within a humanoid 111) or to focus on "tissue" testing (i.e., testing a particular tissue (e.g., the liver) simultaneously) within multiple humanoids 111.

[0228] "3.6 Humanoids" ChamberThe third drawing, designated "Circulation" depicts taking a sample from within the circulation itself. In this embodiment, sampling of the "circulation" is done through a circulation well located at the end of the row of "wells" that each humanoid contains. The circulation well is an empty well designated the "whole body chamber". While the other six wells all contain organoids 113, the "whole body chamber" well is an empty "well" that does not contain organoids, and its function is to allow for taking a sample from within the circulation.

[0229] It is understood that this functionality is similar to the utility of testing samples from a patient's peripheral blood sample before proceeding to take a "biopsy" from the patient's organ. In one embodiment of the present invention, such samples from the "whole body chamber" may be used as a means of "screening." The screening may be used to determine whether or not to draw (take) a sample from a particular organoid 113. Doing so may dramatically reduce the amount of testing performed, thus further supporting non-linear scaling and cost reduction.

[0230] As a first example for utilizing the "screening" capability, if a sample from the "whole body chamber," reflecting a combined sample from all organoid chambers 112 of a humanoid 111, shows elevated lactate, this may prompt taking samples from all organoid chambers 112 of that humanoid 111 to determine which samples have elevated lactate. As another example, if a test from the "whole body chamber" produces a result that may indicate a possible elevated level of one or more liver enzymes, similar to a whole body blood test in a living person, this may prompt taking a sample from the liver organoid chamber 112 to confirm, verify, and quantify this abnormality.

[0231] This screening capability may be directly linked to and dependent on the deep learning and automatic tagging capabilities of the present invention. While it may be impossible to manually determine the correlation between the "whole body chamber" test and the test of individual organoids 113, automatic tagging and deep learning can actually discover, learn and implement such correlations automatically. With every experimental cycle, the system observes and automatically improves the correlation between the findings from the test applied to the "whole body chamber" and the test performed on the individual organoids 113, optimizing the use of the "whole body chamber" test.

[0232] Reference is now made to Figure 3E, which shows yet another diagram of an example of a chip-on-chip implementation of the present invention, further "zooming in" to show a more detailed view of one organoid 113 and its interaction with one of the automated chips 120, the chemical sensing chip 131.

[0233] The first figure, designated "1. Organoid Diagram," depicts a "cross-section-like" view of a single organoid chamber 112 unit (which is part of the multi-human chip 110) interacting with a single sensor array unit (which is part of the chemical sensing chip 131, which is one of the automated chips 120).

[0234] The figure depicts an organoid chamber 112 and organoids 113 disposed at the bottom of the organoid chamber 112. Also depicted is an alternative circulation channel 285 that flows through the bottom of the organoid chamber 112, thus applying shear forces to the organoids 113, to enhance the vitality of the organoids 113. The edges of the organoid chamber 112 may be tapered to aid in the safe insertion of the capillary 1112 into the organoid chamber 112. It is understood that the inter-organoid circulation that flows through and connects the multiple organoids 113 of the humanoid 111 is not provided by the circulation chip 142, but by the circulation pump and circulation channel 285, both of which are present as fixed parts of the multi-human chip 110.

[0235] The illustrated sensor array unit is part of a chemical sensing chip 131 (one of the automation chips 120). The capillary 1112 is a capillary that protrudes from the automation chip 120 and is placed in the organoid chamber 112. The action of moving the automation chip 120 to its position on the multi-human chip 110 is performed by a precision robot 154 of an actuator 150, both of which are not shown.

[0236] In another preferred embodiment of the present invention, the organoid chamber 112 may be sealed at its top, and instead of inserting the capillary 1112 directly into the organoid chamber 112, a second well is created adjacent to the organoid chamber 112, and fluid is transferred between the organoid chamber 112 and the second well, for example, through a labyrinth-like mechanism, or other similar mechanisms as known in the art. This allows the organoid chamber 112 to remain sealed, while eliminating the extra step of opening and closing a "lid" for each test cycle.

[0237] In the embodiment depicted here, the capillary 1112 opens into seven sensor chambers each containing one or more nanosensors. The nanosensors may preferably include, but are not limited to, the following sensors: oxygen, glucose, lactate, urea, potassium, ammonia, and pH. The sensors may be electrochemical, with or without suitable membrane or enzyme coatings, as required to identify their desired target, as is well known in the art, or they may be fluorescent-based, or based on other nanosensor technologies known in the art.

[0238] A ventilation hole is provided at the distal end of the capillary 1112. The ventilation hole may allow flow through the capillary 1112. In another embodiment of the invention, an actively pumped flushing mechanism may be connected to the ventilation hole so that the capillary 1112 and sensor chamber may be flushed and rinsed after each sample.

[0239] It is understood that the spherical organoids 113 are typically one hundred to three hundred microns (100-300 μm) in diameter. The organoid chamber 112 in this embodiment is 1 mm in diameter and eight hundred microns (800 μm) deep. Preferably, the organoid chamber 112 is four hundred microns (400 μm) or more in diameter to comfortably accommodate the organoids 113 and to account for the need for sufficient nutrient fluid surrounding the organoids 113. However, other important considerations are taken into account, including but not limited to the following: The width of the "wall" of the capillary 1112 protruding from the edge of the automated chip 120 should be sufficient to have sufficient structural strength and to be safely inserted into the organoids 113. Walls of fifty to one hundred microns (50-100 μm) in thickness may be preferably used as they are strong enough when fabricating chips using etching in glass materials, as is known in the art. The desired inner diameter of the sample extraction nanotube 297 must be feasible to manufacture using commercially available, cost-effective manufacturing methods and must accommodate efficient and unimpeded flow of tissue-supporting fluids. An inner diameter of fifty to one hundred microns (50-100 μm) may be preferably and safely used. The structure of the tip of the capillary 1112 is important to provide it with sufficient robustness, especially considering that it will be repeatedly inserted into the organoid 113. The cross section of the wall of the tip of the capillary 1112 is preferably more robust at its base and thinner at its tip, but is preferably fifty to one hundred microns (50-100 μm) or larger.

[0240] This embodiment preferably uses organoids 113 with a diameter of 100 to 300 microns (100 to 300 μm). The organoid chamber 112 may preferably have an inner diameter of 700 microns (700 μm) to 1 mm (1 mm in the embodiment depicted herein). The inner diameter of the circulation channel 285 may be 50 to 100 microns (50 to 100 μm) in diameter (100 microns (100 μm) in the embodiment depicted herein). The "walls" of the structure are preferably 100 microns (100 μm) or larger, and the tip of the capillary 1112 may preferably be 100 microns (100 μm) or larger at its narrow distal end and preferably 300 to 500 microns (300 to 500 μm) at its thicker proximal end. The circulation channel 285 enters the organoid chamber 112 at its bottom. In other preferred embodiments, the circulation channel 285 may enter the organoid chamber 112 at other locations. The nanosensor wells in the automated chip 120 are 100 microns (100 μm) in the depicted embodiment.

[0241] The pump used to drive the flow in the circulation channel 285 may preferably be a piezoelectric pump. A significantly smaller pump may be used, so the pump may have dimensions of 2 mm x 2 mm. The fluid sample entering the sampling nanotube 297 may preferably be driven by capillary forces without a pump, as may the capillary and the output of the nanosensor well. Alternatively, a pump similar to that described above may be used, and may be used to flush the nanosensor well between test samples.

[0242] Using the above architecture, a chip containing 108 humanoids and 648 organoids may be constructed, which will have dimensions of approximately 60 mm x 70 mm x 3 mm.

[0243] The second drawing, designated "2. Chemical Sensing Chip," provides a diagram of a chemical sensing chip 131 which, in this embodiment, includes six nanosensor arrays to simultaneously sample all six organoids 113 of the humanoid 111.

[0244] Reference is now made to FIG. 3F, which shows an example of an implementation of the chemical sensing chip 131. In this preferred embodiment, the chemical sensing chip 131 includes six chemical sensors that operate to measure chemicals in each of the six organoid chambers 112. In this preferred embodiment, each sensor includes two probes with a length of 1.1 mm and a total width of 0.6 mm. The probes fit into the organoid chambers 112 and protrude into the fluid within the organoid chambers 112.

[0245] In a preferred embodiment, each such sensor includes three electrodes, as depicted in Figure 3F. On the first probe of the sensor there is preferably a first electrode, which may preferably be made of gold. On the second probe there may preferably be two other electrodes, a second electrode, preferably made of silver or silver chloride, and a third electrode, preferably made of platinum. The body of the sensor and its two probes may preferably be made of silicon or silicon dioxide.

[0246] The gold electrode preferably functions as a "working" electrode that operates to interact with a target chemical. The platinum electrode may preferably function as a "counter" electrode. And the silver electrode may preferably function as a "reference" electrode.

[0247] It is understood that the counter electrode (platinum) and the reference electrode (silver) may be common to several "working" (gold) electrodes. In another preferred embodiment of the invention (not shown), the first probe includes several separate smaller gold working electrodes, each of which is operative to detect a different target chemical, and a single counter electrode and a single reference electrode are common to all of these smaller gold working electrodes. In such an embodiment, a single two-probe sensor unit is operative to identify multiple target chemicals, such as two, or three, or four, or five, or six, or seven, or eight, or nine, or ten different target chemicals. In such an embodiment, the sensor unit includes multiple smaller gold working electrodes, rather than one large gold working electrode, each of which is operative to detect a respective target chemical, and all of the smaller gold working electrodes operate in conjunction with a single platinum counter electrode and a single silver counter electrode.

[0248] In a preferred embodiment of the invention, two sides, a front and a back sensor, may be utilized such that each of the two sides may include two or three or four or five gold working electrodes, a single platinum counter electrode, and a single silver counter electrode, such that the sensor device may operate to identify four, six, eight or ten chemicals.

[0249] It will be appreciated that the calibration and rinsing well 300 may be utilized to rinse the chemical sensing chip 131 after a measurement is made and before the next measurement is made, so as to avoid the next organoid chamber 112 to be tested being contaminated with fluid from the previous set of organoid chambers 112 that have been tested.

[0250] Reference is now made to FIG. 3G, which illustrates an example of an implementation of the present invention, showing an implementation of a circulation chip 142. In this preferred embodiment, the circulation chip 142 operates to pump fluid between six organoid chambers 112 of a humanoid 111. Six organoids 113 are seen at the bottom of each of the six organoid chambers 112. In this preferred embodiment, the fluid flows from left to right, and the image shown in FIG. 3G captures the point at which the flow reaches the first four chambers from the left. The solid arrows indicate the inflow and outflow of fluid into the first four organoid chambers 112 from the left, and the dashed arrows indicate the path of fluid flow from the fourth organoid chamber 112 onward.

[0251] Reference is now made to FIG. 4, which illustrates a schematic of the cyclic architecture of the present invention.

[0252] The chip-on-chip system 100 includes two complementary forms of microfluidic flow to organoids 113 in respective organoid chambers 112 within a humanoid 111.

[0253] The first type of microfluidic flow is a continuous independent actively pumped flow within the organoid chamber 112. This is an "intra-organoid" flow that provides the necessary flow-induced shear force required for the vitality of the organoids 113. This is independent of the connections between the organoids 113 and is implemented by the multi-human chip 110 as follows. Each of the organoid chambers 112 has an associated organoid flow tube 114 and an organoid tube sac 115. The organoid flow tube 114 and the organoid tube sac 115 may be implemented in different ways. In one embodiment of the invention depicted in FIG. 4, the organoid flow tube 114 is a circular tube that operates to flow fluid from and back to the organoid chamber 112. In this embodiment, the organoid tube sac 115 is a widened portion of the organoid flow tube 114 and is depicted in FIG. 4 as being below the organoid chamber 112. When the organoid tube sac 115 is compressed, preferably by peristalsis (described herein below), fluid flows circularly through the organoid flow tube 114, through and across the organoid chamber 112, exerting shear forces on the organoids 113.

[0254] In another preferred embodiment of the present invention (not shown), the organoid tube sac 115 may be a chamber that is preferably adjacent to the organoid chamber 112 and connected to the organoid chamber 112 by the organoid flow tube 114, which in this embodiment is a linear (non-circular) tube rather than a circular tube that operates to flow fluid back and forth, i.e., from the organoid chamber 112 to the organoid tube sac 115 and from the organoid tube sac 115 back to the organoid chamber 112. In this embodiment, the bottom of this chamber (the organoid tube sac 115) is flexible, so that when compressed (described below) and released, fluid flows between the organoid chamber 112 and the adjacent chamber (which in this embodiment is the organoid tube sac 115) in a back-and-forth manner rather than a circular one. This flow also does not cross the organoid chamber 112, but still applies sufficient shear flow stimulation to the organoids 113.

[0255] In a preferred embodiment of the present invention, the multiple organoid tube sacs 115 may be pulsed compressed by a compression wheel (compressor impeller) 404, preferably driven by a compression shaft 402 and an engine pump 400. The compression wheel 404, compression shaft 402, and engine pump 400 are all part of the actuator 150 and are firmly positioned at the base of the actuator 150 directly below the multi-human chip 110, so that when the compression wheel 404 protrudes upward, it compresses the multiple organoid tube sacs 115, thereby driving fluid through the corresponding organoid flow tube 114 and organoid chamber 112. In the above alternative embodiment (not shown), the compression wheel 404 compresses the flexible bottom of the adjacent compartment (which in this embodiment is the organoid tube sac 115), thereby similarly generating intra-organoid flow.

[0256] In another preferred embodiment of the invention, compression of the organoid tube sac 115 may be actuated by other means, rather than by a compression shaft 402 turning a compression wheel 404, such as by a rotating band having protrusions, the rotation of which causes repeated compression of the organoid tube sac 115 as each of the protrusions aligns with the organoid tube sac 115. In another preferred embodiment of the invention, the compression wheel 404 or sliding protrusions may not directly push the organoid tube sac 115, but rather push a rod constrained in the vertical direction, which pushes the organoid tube sac 115.

[0257] The second type of microfluidic flow is the "inter-organoid" flow, which connects multiple organoid chambers 112 in the humanoid 111, and thus mimics the cardiovascular flow connecting different organs in the body. This flow is provided by the circulation chip 142. The circulation chip 142 is constructed and operates to mimic the cardiovascular circulation in the body, which circulates fluids between the organoids 113 placed in the organoid chambers 112 of the humanoid 111. The circulation chip 142 depicted in FIG. 4 includes five circulation tubes 204 and one long circulation tube 205. Each one of the circulation tubes 204 flows fluid from one organoid chamber 112 to an adjacent organoid chamber 112, and the long circulation tube 205 flows fluid from the right-most organoid chamber 112 to the left-most organoid chamber 112 of the humanoid 111. The example depicted in FIG. 4, in which the circulation chip 142 provides circulation to a humanoid 111 having six organoid chambers 112, is merely an example and is not meant to be limiting, and it is understood that the humanoid 111 may include a different number of organoid chambers 112.

[0258] Reference is now made to FIG. 5 which illustrates diagrammatically the operation of the circulatory pumping described in FIG. 4 herein above.

[0259] FIG. 5 includes two views designated "Side View" and "3D View" that collectively explain the operation of circular pumping. The illustration designated "Side View" includes three drawings depicting three time phases. It is understood that the compression wheels 404 are concentric, and therefore, as the compression wheels 404 rotate, they compress the organoid tube sac 115 and create flow in the organoid flow tube 114. The view designated "3D View" shows how a row of compression wheels 404 may be rotated together by a common compression shaft 402 driven by an engine pump 400.

[0260] Reference is now made to FIG. 6, which illustrates a schematic example of a shunted flow in a humanoid 111.

[0261] The circulation chip 142 is designed and operative to provide complete control of the routing of fluid flow between the organoid chambers 112, allowing the user to "short-circuit" or bypass any one of the organoid chambers 112 as desired. This short-circuiting is accomplished by bypass tubing 600 and a number of valves designated 602-644 that control flow within the circulation tubing, circulation tube 204, and long circulation tube 205.

[0262] 6 provides an example of a shunt function. In this example, the chip-on-chip system 100 redirects the flow through the humanoid 111 so that fluid flows only through the organoid chambers 112 of the liver organoids 646 and the heart organoids 648, but bypasses all other organoid chambers 112 in this humanoid 111. To do so, the following valves are opened, valves 602, 608, 614, 616, 622, 626, 628, 634, 636, and 642, and all other valves are closed. As a result, a unique redirected flow path is activated, which allows fluid to flow into the chamber of the liver organoids 646 and from there into the chamber of the heart organoids 648, while bypassing all other organoid chambers 112 in the humanoid 111.

[0263] It is understood that this example shows the importance of the two separate circulation modes of the present invention. The short circuit shown in FIG. 6 is performed by the circulation chip 142, and has changed the circulation connecting the organoid chambers 112. However, at the same time, the separate circulation of each of these organoid chambers 112 through the organoid flow tube 114 and the organoid tube sac 115 of each of these organoid chambers 112 remains unchanged. This is important because it means that the inter-organoid flow can be short circuited while still maintaining the unchanged shear flow in each organoid chamber 112.

[0264] It is understood that the above is intended as merely an example and is not intended to be limiting.In a preferred embodiment of the present invention, bypassing may be achieved by intermittent shorting, rather than by complete shorting as detailed above, so that the shorted organoid chamber is allowed to have a minimal flow, but the flow to the bypassed organoids is kept to a minimum, so as not to deprive the shorted organoid chamber of shear forces and nutrients.In another preferred embodiment, the valves may be opened to various degrees, so that the valves to the "bypassed" organoid chambers remain slightly open, allowing continuous limited flow to the "bypassed" organoid chambers, while the valves to the non-bypassed organoid chambers remain fully open, giving full flow to these organoids.In yet another preferred embodiment of the present invention, there are multiple parallel channels, so that the flow rate that the organoid chamber receives is controlled by the number of open channels, not by the degree to which the valves are open.

[0265] In yet another preferred embodiment of the present invention, the degree to which the flow tapers to the "bypassed" organoid chamber is adjusted autonomously by monitoring the vitality of the organoid 113 by one of the diagnostic chips 130, such as a chemical sensing chip. In another preferred embodiment of the present invention, the degree to which the flow tapers to the "bypassed" organoid chamber depends on the tissue type of the corresponding organoid 113, since different tissues may preferably require different flows. In another preferred embodiment of the present invention, this mechanism of shunted flow is used to adjust the appropriate flow for each tissue type of the organoid 113, rather than temporarily shunting the organ. In yet another preferred embodiment of the present invention, the shunted flow is used to adjust the flow for each instance of the organoid 113, rather than for tissue types in general, e.g., to adjust the flow rate for each instance of liver organoids on the multi-human chip 110, rather than for "liver" organoids in general. It may be that one instance of liver organoid may preferably require a different flow than a different instance of liver organoids.

[0266] In this respect, it can be seen that the key unique novelty of the present invention is its ability of autonomous self-learning.Whether it is evaluating and comparing different architectures from the above preferred embodiment, and whether it is adapting the preferred embodiment to different tissue types, and whether it is fine-tuning to different cases of organoids of given tissue during the "runtime" operation of the system, in all these cases, diagnostic chip 130 can be used in combination with the powerful artificial intelligence of deep matching engine 170, and preferably autonomously evaluate, fine-tune and optimize the flow that works best for tissue organoid 113.

[0267] Reference is now made to FIG. 7 which illustrates a schematic of one example of on-chip integrated tissue scaling for a humanoid 111.

[0268] Those skilled in the art will understand that simply connecting multiple organ chips together will not effectively represent the human body, even if each such organ chip is effective on its own. This is because each "organ chip" must be scaled differently to reflect the different mass and circulatory criteria of different tissues. As an example, myocardial organoid 700 and liver tissue organoid 702 with similar size, e.g., 1000 cells, must then be scaled very differently to adequately represent the human body, as an example, to reflect the larger mass and much larger "circulatory exposure" of liver tissue compared to myocardial tissue.

[0269] Current Organ Chip systems have made early attempts to address this challenge by relying on post-experiment computational correction. Simply put, these systems simply perform biological experiments on the Organ Chip, knowing that they are highly biased, and then attempt to correct some of the bias computationally. Such attempts, while useful, are seriously flawed. Experience has shown that biological systems are complex and do not fit theoretical models well.

[0270] The present invention utilizes organoid individual circulation and shunting capabilities to provide on-chip physical scaling integrated with computational fine-tuning correction.

[0271] FIG. 7 provides an example where on-chip scaling requires that liver organoids 702 receive twice as much "circulation exposure" as cardiac organoids 700. The organoid individual circulation feature is the key to address this. Because each organoid 113 is connected to all other organoids 113 in humanoid 111, and at the same time, each organoid 113 is also completely isolated from other organoids 113 and is continuously subjected to organoid individual circulation, but with full control of its connection with other organoids 113 and full control of its nutrients, clearance and drug exposure. Therefore, the amount of nutrients 704, clearance 706 and meds (drugs) 708 that cardiac organoids 700 receive is preferably set to be appropriately different from the nutrients 710, clearance 712 and meds 714 that liver organoids 702 receive. Short circuit capability may be used for this purpose as well.

[0272] It is well known in the art to determine how much compensation is generally needed for Organ Chips representing different organs in order to scale up appropriately. In a preferred embodiment of the present invention, such compensation is utilized here to generate an on-chip physical correction of the differential "blood" flow through the different Organ Chips, rather than simply applying post-experimental computational adjustments. In a preferred embodiment of the present invention, the amount of adjustment may be further fine-tuned by matching the PK / PD profiles of known drugs in the human body to their patterns on the Human Chip.

[0273] It will be understood that the above are intended as examples only and are not intended to be limiting. As discussed above, differential flow may be achieved by a variety of means and strategies.

[0274] Reference is now made to FIG. 8A, which illustrates, in a simplified manner, an example of the architecture of a spherical organoid 113 constructed and operative in accordance with a preferred embodiment of the present invention, as well as a method for producing such a spherical organoid 113 in accordance with a preferred embodiment of the present invention.

[0275] The organoid 113 is a type of organoid 113. The spherical organoid 800 is placed or formed in the organoid chamber 112. In a preferred embodiment of the present invention, the organoid chamber is shaped and sized to closely fit the spherical organoid 800. As an example, the spherical organoid may be a sphere preferably having a diameter of 100 to 300 microns (100 to 300 μm), and therefore the organoid chamber may be preferably shaped to approximate a round inclusion of 400 to 500 microns (400 to 500 μm), or other dimensions to approximate the size and shape of the spherical organoid 800.

[0276] The common circulating medium flows continuously through the organoid chamber 112 by the organoid flow tube 114 and is pumped by the organoid tube sac 115 that is pulsed and compressed by the pumping mechanism. This circulating medium is also intermittently shared with other organoids 113 in the humanoid 111 by the circulation chip 142. The common circulating medium serves both to nourish the spherical organoids 800 and to apply shear force to the spherical organoids 800, thereby activating the spherical organoids.

[0277] It is understood that the selection of the optimal dimensions of the tubes that carry out the flow of fluid in the organoid chamber 112 is greatly facilitated by the real-time measurement from the diagnostic chip 130. The width, pressure and flow rate of both the organoid flow tube 114 and the tubes of the circulation chip 142 (circulation tube 204 and long circulation tube 205) determine the amount of shear force applied to the spherical organoids 800 and the amount of nutrients to which the spherical organoids 800 are exposed. If too much or too little shear is applied, or if not enough nutrients are provided by the flow, the vitality of the organization of the spherical organoids 800 is hindered, as reflected by the real-time measurement of the diagnostic chip 130.

[0278] 8A shows a preferred process of placing organoid 800 in organoid chamber 112, which is preferably performed by precision robot 154 of actuator 150. In a preferred embodiment of the present invention, precision robot 154 may first form spherical organoid 800 (not shown here) as known in the art. Various methods of making spherical organoid 800 are known in the art. In another preferred embodiment of the present invention, precision robot 154 may make spherical organoid 800 by 3D printing, including by single cell resolution 3D printing, or may receive spherical organoid 800 made externally by such process or by other commercially available process.

[0279] In a preferred embodiment of the present invention, the precision robot 154 opens the chamber lid 804 (this action is not shown here), thereby opening and exposing the organoid chamber as an "open well" for the process of inserting the spherical organoids 800 into the organoid chamber 112.

[0280] The image on the right, entitled "Before Insertion," depicts a spherical organoid 800 held by a placement nozzle 802 of the precision robot 154 above the "open well" organoid chamber 112, ready to be placed within the organoid chamber 112.

[0281] The central figure, entitled "After Insertion," depicts a spherical organoid 800 placed in the organoid chamber 112 after being placed therein by the placement nozzle 802 of the precision robot 154.

[0282] The left diagram, entitled "Lid Closure", depicts the chamber lid 804 positioned to seal the organoid chamber 112 containing the spherical organoids 800. After inserting the spherical organoids 800 into the organoid chamber 112, the precision robot 154 closes the chamber lid 804.

[0283] In a preferred embodiment of the present invention, the precision robot 154 may close the chamber lid 804 of each organoid chamber 112. In another preferred embodiment of the present invention, the chip-on-chip system 100 operates in an "open well" without the use of a chamber lid 804. In yet another preferred embodiment of the present invention, the design of the organoid chamber 112 is such that the organoid chamber 112 preferably has a chamber lid 804 that is closed, except that the organoid chamber 112 is opened for placement of the organoid chamber 112, and then ongoing maintenance of the organoid chamber 112 (e.g., use of the circulation chip 142 and the chemical sensing chip 131) is performed through the adjacent connected well. In this implementation, each organoid chamber 112 has an associated chamber to which the organoid chamber 112 is connected by a connecting tube, the organoid chamber 112 has its chamber lid 804 closed throughout the duration of the experiment on the multi-human chip 110 after the organoids 113 are created and inserted into the organoid chamber 112, and all interactions between the automated chip 120 and the organoid chamber 112 are performed by the automated chip 120 interacting with the adjacent connected chamber, which then exchange fluids driven by gravity flow.

[0284] In a preferred embodiment of the present invention, before inserting the spherical organoids 800 into the organoid chamber 112, the multi-human chip 110 preferably continuously flows a common circulating medium through the organoid flow tube 114 associated with the organoid chamber 112, so as to fill the organoid chamber 112 and expel any air bubbles from the organoid flow tube 114. In a preferred embodiment of the present invention, this process may preferably be repeated after the spherical organoids 800 are inserted into the organoid chamber 112 and the chamber lid 804 is closed. This second pumping of the common circulating medium completely fills the organoid chamber 112 after the spherical organoids 800 are inserted, and again expels any air bubbles from the organoid flow tube 114.

[0285] It is understood that the above description of the spherical organoids 800 inserted by the precision robot 154 is intended as an example only and is not intended to be limiting. In another preferred embodiment of the present invention, the spherical organoids may be formed in the organoid chamber 112 itself, rather than being first created and then inserted entirely therein. In such an embodiment, a suitable mixture of cells of a desired tissue in a desired ratio of different cell types is poured into the spherical organoid chamber 112, and 3D spheroids are allowed to form within the organoid chamber 112 using methods well known in the art. As an example, the construction material of the organoid chamber 112 or its coating may preferably be such that the cells are prevented from adhering to the surface of the organoid chamber 112, thereby aggregating to form spheroids. Such a mixture of cells may be flowed through the placement nozzle 802 of the starting chip 141, which is operated by the precision robot 154 and the incubator 156.

[0286] Reference is now made to Figure 8B, which illustrates a schematic example of a perfusion type architecture for organoids 113.

[0287] In the preferred embodiment of the present invention, perfused organoid can be preferably in a membrane shape, and often includes two adjacent tissue layers, thereby mimicking the perfusion site in human body.The examples of perfused organoid include but are not limited to lung, intestine, kidney, and blood-brain barrier (BBB).

[0288] Specifically, the lung preferably includes an alveolar epithelial tissue layer adjacent to a capillary epithelial tissue layer, which may mimic the pulmonary blood-gas perfusion interface; the intestine preferably includes an intestinal epithelial tissue layer adjacent to a capillary epithelial tissue layer, which may mimic the gastrointestinal nutrient absorption perfusion interface; the kidney preferably includes a capillary epithelial tissue layer adjacent to an adjacent nephron glomerular tissue layer, which may mimic renal secretory function; and the blood-brain barrier preferably includes a capillary epithelial tissue layer adjacent to brain barrier tissue, which may mimic the blood-brain barrier perfusion interface.

[0289] 8B illustrates a perfused organoid unit 830 constructed and operative in accordance with a preferred embodiment of the present invention. For clarity, both a 3D view and top and side views are provided.

[0290] The perfused organoid unit 830 is a type of organoid chamber 112 having two chambers, here designated chamber-1 806 and chamber-2 808, which contain two perfused organoids 832, designated organoid-1 810 and organoid-2 812, respectively. In a preferred embodiment of the present invention, the perfused organoids 832 are tissues formed in a flattened membrane-like morphology, allowing perfusion therethrough.

[0291] A membrane 816 separates the two adjacent chambers mentioned above. The membrane 816 is a porous elastic membrane. In a preferred embodiment of the present invention, this membrane may preferably serve as a scaffold on which two adjacent perfused organoids 832, organoid-1 810 and organoid-2 812, are formed. In a preferred embodiment of the present invention, the membrane 816 is constructed and operates simply to allow uninterrupted perfusion through the membrane, such that after these adjacent perfused organoids 832 are formed on the membrane, the membrane 816 then regulates the perfusion, so that it is the perfused organoids 832. In a preferred embodiment of the present invention, the membrane 816 may be a synthetic biologically inert membrane with sufficiently large pores, as is well known in the art. In another preferred embodiment of the present invention, the membrane 816 may be a biodegradable membrane that preferably disintegrates after the perfused organoids 832 are formed on its sides.

[0292] Chamber-1 806 is associated with the organoid flow tube 114, which provides continuous circulation within this chamber, i.e., it is not connected to other organoids as any "normal" spherical organoids would be. Chamber-2 808 includes a secondary circulation 814, which is pumped by a dedicated secondary pump, not shown. It will be appreciated that the secondary circulation 814 is similar in structure to the organoid flow tube 114, and may preferably be pumped by a similar mechanism or the same mechanism that pumps the organoid flow tube 114 and described above with reference to Figures 4 and 5.

[0293] Below are some examples of perfused organoids 832. In the "lung" type perfused organoids 832, chamber-1 806 may represent the air side of the alveoli, and organoid-1 810 may be alveolar tissue. Chamber-2 808 may represent blood vessels, and organoid-2 812 may be epithelial tissue, and oxygen passes through the two adjacent layers of the perfused organoids 832 and is absorbed into the "blood" component of chamber-2 808.

[0294] In the "gut" type perfused organoid 832, chamber-1 806 may represent the intestinal lumen and organoid-1 810 may be intestinal wall tissue. Chamber-2 808 may represent blood vessels and organoid-2 812 may be epithelial tissue, and digested nutrients traverse the two adjacent layers of the perfused organoid 832 and are absorbed into the "blood" component of chamber-2 808.

[0295] In the "kidney" type perfused organoid 832, chamber-1 806 may represent blood vessels and organoid-1 810 may be epithelial tissue. Chamber-2 808 may represent the renal lumen and organoid-2 812 may be glomerular tissue, with urinary filtrate traversing the two adjacent layers of the perfused organoid 832 and being excreted in the "urine" component of chamber-2 808.

[0296] The ADME chip 143 is depicted with its ADME tube 206 lowered into its home position within chamber-2 808. As an example, FIG. 8B depicts a "kidney" perfused organoid unit 830 scenario, where chamber-1 806 is the "blood" component and chamber-2 808 is the "urine" component. In this example, the ADME tube 206 of the ADME chip 143 is lowered into chamber-2 808 and aspirates and removes the "urine" that has accumulated within this chamber, mimicking an action such as the removal of urine from the kidney to the bladder. Fluids removed by the ADME tube 206 may be collected, tested, or discarded.

[0297] It is understood that an important inventive aspect of the present invention is that the membrane 816 and the perfused organoids-organoid-1 810 and organoid-2 812- are top-down (vertical). This has not been done before. In all previous devices, the membrane of the perfused organoids was placed horizontally. This is because a horizontal arrangement was thought to be essential for culturing tissues using the "sedimentation" of cells on the membrane due to gravity, and because a horizontal arrangement was thought to be more convenient when handling manually, especially when the main monitoring method is microscopy. In a preferred embodiment of the present invention, the membrane 816 and the perfused organoids-organoid-1 810 and organoid-2 812- are top-down. This aspect of the present invention provides a unique advantage that allows the high throughput capabilities of the present invention. As an example, it allows automated high throughput access to both adjacent perfusion chambers from above by a precision robot 154.

[0298] Reference is now made to Figure 8C, which illustrates a schematic diagram of the extension of a perfused organoid unit 830.

[0299] One or more PZT contractors 818 are built into the walls of the perfused organoid unit 830 such that the membrane 816 and the perfused organoids 832, designated organoid-1 810 and organoid-2 812, are stretched between the one or more PZT contractors 818. The PZT contractors 818 are small piezoelectric elements that operate to contract or deform, thereby stretching and relaxing the membrane 816 and the perfused organoids 832, designated organoid-1 810 and organoid-2 812. In FIG. 8B, the left diagram, entitled "Native State," shows the two elements of the PZT contractors 818 in a relaxed state, such that the membrane 816 and the perfused organoids 832, designated organoid-1 810 and organoid-2 812, are in a native state and are not stretched.

[0300] In the image on the right, entitled "Stretched Membrane," two elements of the PZT contractor 818 contract, thereby stretching the membrane 816 and the perfused organoids 832, designated organoid-1 810 and organoid-2 812.

[0301] It is understood that the functionality of PZT constrictor 818 may be implemented in different modes: In another preferred embodiment of the present invention, a PZT constrictor 818 may be woven into membrane 816, thereby deforming the membrane, for example forming a dome shape, thereby stretching the membrane, as known in the art.

[0302] It is understood that the controller 151 may actuate multiple PZT contractors 818 within multiple perfused organoids 832 on the multi-human chip 110 so that the membranes and perfused organoids 832 rhythmically stretch and relax, mimicking peristalsis in gastrointestinal tissue, respiration in pulmonary tissue, etc., which are important to effectively mimic the permeability of such perfused organoids 832 and their true three-dimensional tissue activity.

[0303] It is understood that this implantation of PZT-driven stretching and relaxation of perfused organoids 832 is uniquely novel to the present invention. Current organ chips use membrane tissue stretching, but are so driven by pneumatic channels, which require an external pneumatic device to drive such peristalsis. The present invention is unique in that it uses a miniaturized, autonomous on-chip mechanism for membrane stretching. It is understood that this element is the key required to enable membrane stretching on a fully autonomous, all-on-chip, high-throughput, truly scalable, human-on-chip system.

[0304] Reference is now made to FIG. 8D, which illustrates a schematic representation of the mode of perfused organoid production by both the starting chip 141 and the precision robot 154.

[0305] Figure 8D includes four diagrams which together show how perfused organoids are generated and validated according to a preferred embodiment of the present invention. The four diagrams depict the sequence of four respective steps:

[0306] Reference is now made to the first diagram in FIG. 8D, entitled "1. Installing tissue-1," which simply shows the initial state of two adjacent empty perfused organoid chambers before any perfused organoids 832 are installed in the perfused organoid chambers. It is important to note that this dual chamber includes a first chamber designated chamber-1 806 and a second chamber designated chamber-2 808, and in this first diagram, the first chamber is shown at the top and the second chamber is shown at the bottom. The chambers are separated by a membrane 816. A circulating inflow 810 and a circulating outflow 815 are connected to chamber-1 806.

[0307] It will be appreciated that there are two organoid chamber lids shown here: organoid chamber lid-1, designated 822, sealing the top of chamber-1 806, and organoid chamber lid-2, designated 824, sealing the bottom of chamber-2 808. These organoid chamber lids function as part of the mechanism for creating perfused organoids 832, and therefore, for clarity of understanding, As shown .

[0308] Reference is now made to the first diagram entitled "1. Placing Tissue-1" in FIG. 8D. In this first step, the organoid chamber lid-1 822 (not shown) is removed by the precision robot 154, leaving chamber-1 806 as an "open well." The precision robot 154 positions the placement nozzle 802 directly over the open well of chamber-1 806 and pours tissue-1-cell fluid, here designated 820, into chamber-1 806 in direct contact with the membrane 816 on the right side.

[0309] The membrane 816 comprises a suitable porous surface for cells to attach to. The tissue-1-cell fluid 820 is then incubated for a suitable duration of several days (e.g., 4 days) to allow the cells therein to organize into a thin membrane-like tissue that adheres to the membrane 816 and acquires a morphology and physiology similar to the native human tissue from which these cells originate. All aspects of forming a membrane-like organ from cells incubated on top of a membrane or other porous surface, some of which are described herein above with reference to this figure, are well known to those skilled in the art and will therefore only be described briefly.

[0310] For simplicity, the two PZT constrictors 818 are not shown, as their role has been discussed above with reference to FIG. 8C and is not relevant to understanding the process of creating perfused organoids. In this first stage, it is understood that both lids are in place, closing the upper and lower parts of the two chambers, respectively, as shown in the diagram entitled "1. Stand Up."

[0311] Reference is now made to the second diagram entitled "2. Close Lid-1" in FIG. 8D. In this step, a solid layer of tissue designated organoid-1 810 is formed and adheres to the membrane 816, and excess fluid and cell supernatant is gently washed away as is well known in the art. In a preferred embodiment of the present invention, washing of excess fluid and cells may be performed through a mounting nozzle 802 of the starting chip 141, preferably operated by a precision robot 154, which is operable to aspirate and wash excess fluid from the chamber. The organoid chamber lid-1 822 is reattached to close the chamber-1 806. Finally, the multi-human chip 110 preferably continuously flows a suitable fluid medium through the circulation and channels to fill the chamber-1 806 and to expel any air bubbles from the channels and secondary circulation. The end result of this process is that perfused organoids 830, designated organoid-1 810, are formed within chamber-1 806, attached to membrane 816, and surrounded by appropriate circulating fluid.

[0312] Reference is now made to the third diagram of FIG. 8D, entitled "3. Placing Tissue-2." In this third step, the organoid chamber lid-2 824 is removed by the precision robot 154, leaving chamber-2 808 as an "open well." The precision robot 154 places the placement nozzle 802 directly over the open well of chamber-2 808 and pours the tissue-2-cell fluid, here designated 826, into chamber-2 808 in direct contact with the membrane 816 on the left side. The tissue-2-cell fluid 826 is incubated for a suitable duration of several days (e.g., 4 days) to allow the cells therein to organize into a thin membrane-like tissue that adheres to the membrane 816 and acquires a morphology and physiology similar to the native human tissue from which these cells are derived.

[0313] Reference is now made to the fourth diagram, entitled "4. Close Lid-2" in FIG. 8D. In this step, a solid layer of tissue designated organoid-2 812 is formed and adheres to the membrane 816, and excess fluid and cell supernatant is gently washed away as is well known in the art. In a preferred embodiment of the present invention, washing of excess fluid and cells may be performed through a mounting nozzle 802 of the starting chip 141, preferably operated by a precision robot 154, which is operable to aspirate and wash excess fluid from the chamber. The organoid chamber lid-2 824 is reattached to close chamber-2 808. Finally, the multi-human chip 110 preferably continuously flows a suitable fluid medium through the circulation and channels to fill chamber-2 808 and to expel any air bubbles from the secondary circulation and channels 860 and 865. The end result of this process is that two perfused organoids 832, designated organoid-1 810 and organoid-2 812, are formed within chamber-1 806 and chamber-2 808, respectively, attached to either side of membrane 816 and surrounded by appropriate circulating fluid.

[0314] It is understood that the figure entitled "4. Close Lid-2" also shows a unique method for automatic verification of the integrity of the perfused organoids formed by the above process, and how the verification process is carried out as follows. As this figure shows, the two chambers also include two verification sensors, chamber-1-membrane verification sensor designated as 892 and chamber-2-membrane verification sensor designated as 894 (not shown in the above five figures and in FIG. 8B for clarity). These two sensors are preferably electrodes constructed and operative to detect the potential drop across the membrane. As is well known in the art, if the biological membranes (biomembranes), such as organoid-2 812 and organoid-1 810 formed in the above method, are defective and / or not fully formed, this may be detected by measuring the potential drop in this manner. Therefore, the present invention automatically detects and analyzes the input from these sensors 826 and 828 for each perfused organoid 113 on the multi-human chip 110. Importantly, the deep matching engine 170 analyzes the input from these sensors during the creation of the organoids, and if this evaluation determines that the perfused organoids 830 are defective, the deep matching engine 170 redesigns the experiment to ignore the perfused organoids that are determined to be defective.

[0315] Similarly, for both the sphere organoids and the perfused organoids 832, real-time cell metabolism sensing continuously measures organoid metabolites that indicate tissue health, including mitochondrial function. These measurements are also evaluated by the deep matching engine 170 during the production of both the organoids, the sphere organoids and the perfused organoids 832, to evaluate the "health" of each organoid 113 on the multi-human chip 110. Thus, organoids 113 that are evaluated as not performing well are automatically identified by the deep matching engine 170, which then redesigns the experiment and ignores these organoids.

[0316] It is understood that this aspect of the present invention of automatic organoid validation is very important and unique. A multi-human chip 110 may preferably contain hundreds of organoids 113, and it is well known in the art that a significant amount of organoids (both perfused and spheres) are often found to be defective at the beginning of an experiment, and furthermore, some organoids become defective during an experiment, for example, before administration of a drug. Current organ chips do not have an automatic means to analyze each organoid, determine its health, automatically redesign the experiment, ignore the defective ones, and, if relevant, bypass them in the fluid interconnection with other organoids. Addressing this manually is possible, albeit difficult and time-consuming, for small-scale experiments with a manual standalone single organ chip. However, it is understood that without such automatic validation and automatic experiment redesign, high throughput is not possible.

[0317] It will be understood that all aspects relating to the formation of membrane-like organelles from cells incubated on top of a membrane or other porous surface, their incubation, and subsequent rinsing, etc. (some of which have been mentioned herein above with reference to the six diagrams of FIG. 8E) are well known to those skilled in the art and will therefore only be mentioned briefly.

[0318] Reference is now made to Figure 8E, which illustrates a schematic of the generation and validation of perfused organoids 113.

[0319] Figure 8E includes six diagrams that collectively show how the perfused organoids of Figure 8B are generated and validated, according to a preferred embodiment of the present invention. The six diagrams depict the sequence of six respective steps:

[0320] Reference is now made to the first diagram of FIG. 8E, entitled "1. Standing Up," which simply shows the initial state of the two adjacent empty perfused organoid chambers of FIG. 8B before any perfused organoids 832 are placed in the perfused organoid chambers. As in FIG. 8B, it is important to note that this dual chamber includes a first chamber designated chamber-1 806 and a second chamber designated chamber-2 808, and in this first diagram, the first chamber is shown at the top and the second chamber is shown at the bottom. The chambers are separated by a membrane 816. A circulating inlet 810 and a circulating outlet 815 are connected to chamber-1 806.

[0321] It is understood that two organoid chamber lids are shown here: organoid chamber lid-1, designated 822, sealing the top of chamber-1 806, and organoid chamber lid-2, designated 824, sealing the bottom of chamber-2 808. These organoid chamber lids function as part of the mechanism for creating perfused organoids 832, and therefore are for clarity of understanding.

[0322] For simplicity, the two PZT constrictors 818 are not shown, as their role has been discussed above with reference to FIG. 8C and is not relevant to understanding the process of creating perfused organoids. In this first stage, it is understood that both lids are in place, closing the upper and lower parts of the two chambers, respectively, as shown in the diagram entitled "1. Stand Up."

[0323] Reference is now made to the second diagram of FIG. 8E, entitled "2. Placing Tissue-1." In this second step, the organoid chamber lid-1 822 of the previous diagram is removed by the precision robot 154, which causes the chamber-1 806 to become an "open well." The precision robot 154 places the placement nozzle 802 directly over the open well of the chamber-1 806 and pours the tissue-1-cell fluid, here designated 820, into the chamber-1 806 on top of the membrane 816. The membrane 816 includes a suitable porous surface that is optimal for the cells to attach. The tissue-1-cell fluid 820 is then incubated for a suitable duration of several days (e.g., 4 days), which allows the cells therein to organize into a thin membrane-like tissue that adheres to the membrane 816 and acquires a morphology and physiology similar to the native human tissue from which these cells are derived. All aspects relating to the formation of membrane-like organelles from cells incubated on top of a membrane or other porous surface, some of which are described above with reference to this figure, are well known to those skilled in the art and will therefore only be described briefly.

[0324] Reference is now made to the third diagram entitled "3. Close the lid" in FIG. 8E. In this step, a solid layer of tissue designated organoid-1 810 is formed and adheres to the membrane 816, and excess fluid and cell supernatant is gently washed away as is well known in the art. In a preferred embodiment of the present invention, washing of excess fluid and cells may preferably be performed through the placement nozzle 802 of the precision robot 154, and in another preferred embodiment of the present invention, washing of excess fluid and cells may preferably be performed through the circulation channel 300. The organoid chamber lid-1 822 is reattached to close the chamber-1 806. And finally, the multi-human chip 110 preferably continuously flows the common circulation medium through the circulation channel 300 of FIG. 3 through the circulation inlet 810 and circulation outlet 815 to fill the chamber-1 806 and to expel any air bubbles from the circulation channel 300. The end result of this process is that perfused organoids 830, designated organoid-1 810, are formed within chamber-1 806, attached to membrane 816, and surrounded by appropriate circulating fluid.

[0325] Reference is now made to the fourth diagram, entitled "3. Inversion," of Figure 8E. In this step, the multi-human chip 110 is flipped by the precision robot 154 to place chamber-2 808 on top.

[0326] Reference is now made to the fifth diagram, entitled "5. Placing Tissue-2" in FIG. 8E. In this fifth step, the organoid chamber lid-2 824 of the previous diagram is removed by the precision robot 154, which causes chamber-2 808 to become an "open well". The precision robot 154 places the placement nozzle 802 directly over the open well of chamber-2 808 and pours tissue-2 cell fluid, here designated 826, into chamber-2 808 on top of the membrane 816. The membrane 816 contains a suitable porous surface that is optimal for cells to attach to. The tissue-2-cell fluid 826 is then incubated for a suitable duration of several days (e.g., 4 days) to allow the cells therein to organize into a thin membrane-like tissue that adheres to the membrane 816 and acquires a morphology and physiology similar to the native human tissue from which these cells are derived.

[0327] Reference is now made to the sixth diagram, entitled "6 Close Lid-2" in FIG. 8E. In this step, a solid layer of tissue designated organoid-2 812 is formed and adheres to the membrane 816, and excess fluid and cell supernatant is gently washed away as is well known in the art. In a preferred embodiment of the present invention, washing of excess fluid and cells may preferably be performed through the placement nozzle 802 of the precision robot 154, and in another preferred embodiment of the present invention, washing of excess fluid and cells may preferably be performed through the secondary circulation and channels 860 and 865. The organoid chamber lid-2 824 is reattached to close the chamber-2 808. And finally, the multi-human chip 110 preferably continuously flows a suitable fluid medium through the secondary circulation and channels 860 and 865 to fill the chamber-2 808 and to expel any air bubbles from the secondary circulation and channels 860 and 865. The end result of this process is that two perfused organoids 832, designated organoid-1 810 and organoid-2 812, are formed in chamber-1 806 and chamber-2 808, respectively, attached to either side of membrane 816 and surrounded by appropriate circulating fluid.

[0328] It is understood that the figure entitled "6. Close Lid-2" also shows a unique method for automated verification of the integrity of perfused organoids formed by the above process, and how the verification process is carried out as follows. As this figure shows, the two chambers also include two verification sensors, chamber-1-membrane verification sensor designated as 892 and chamber-2-membrane verification sensor designated as 894 (not shown in the above five figures and in FIG. 8B for clarity). These two sensors are preferably electrodes constructed and operative to detect the potential drop across the membrane. As is well known in the art, if the biological membranes such as organoid-2 812 and organoid-1 810 formed in the above method are defective and / or not fully formed, this may be detected by measuring the potential drop in this manner. Therefore, the present invention automatically detects and analyzes the input from these sensors 826 and 828 for each perfused organoid 113 on the multi-human chip 110. Importantly, the deep matching engine 170 analyzes the input from these sensors during the creation of the organoids, and if this evaluation determines that the perfused organoids 830 are defective, the deep matching engine 170 redesigns the experiment to ignore the perfused organoids that are determined to be defective.

[0329] Similarly, for both the spherical organoids and the perfused organoids 832, real-time cell metabolism sensing continuously measures organoid metabolites that indicate tissue health, including mitochondrial function. These measurements are also evaluated by the deep matching engine 170 during the production of both the organoids, the spherical organoids and the perfused organoids 832, to evaluate the "health" of each organoid 113 on the multi-human chip 110. Thus, organoids 113 that are evaluated as not performing well are automatically identified by the deep matching engine 170, which then ignores these organoids and redesigns the experiment.

[0330] It is understood that this aspect of the present invention of automatic organoid validation is very important and unique. A multi-human chip 110 may preferably contain hundreds of organoids 113, and it is well known in the art that a significant amount of organoids, both perfusion and spheres, are often found to be defective at the beginning of an experiment, and furthermore, some organoids become defective during an experiment, for example, before administration of a drug. Current organ chips do not have an automatic means to analyze each organoid, determine its health, automatically redesign the experiment, ignore the defective ones, and, if relevant, bypass them in the fluid interconnection with other organoids. Addressing this manually is possible, even if difficult and time-consuming, for small-scale experiments with manual standalone single organ chips. However, it is understood that without such automatic validation and automatic experiment redesign, high throughput is not possible.

[0331] It will be understood that all aspects relating to the formation of membrane-like organelles from cells incubated on top of a membrane or other porous surface, their incubation, and subsequent rinsing, etc. (some of which have been mentioned herein above with reference to the six diagrams of FIG. 8E) are well known to those skilled in the art and will therefore only be mentioned briefly.

[0332] According to another preferred embodiment of the present invention, a fluid containing the cell admixture is placed in chamber-1 806 and chamber-2 808 and their subsequent rinsing (rinsing out) is not performed by opening the respective lids 822 and 824 and injecting and then draining the fluid by the precision robot 154 and its installed nozzles, but rather this may preferably be performed by flowing the fluid into these chambers through a microfluidic channel and rinsing them through the same channel after sufficient incubation.

[0333] Reference is now also made to Figures 9A and 9B, which illustrate generally one example of the architecture of a humanoid 111 constructed and operative in accordance with the present invention.

[0334] The humanoid 111 of the present invention may be flexibly constructed according to the user's preferences and to suit the desired experimental objectives. Below is an example of a typical possible architecture of the humanoid 111, with a description of its different components and considerations.

[0335] The humanoid 111 is comprised of a set of organoid chambers 112 containing respective sets of organoids 113, which may be of the "spherical" or "perfused" type. The organoids 113 and their respective organoid chambers 112 are modular, and thus many different humanoids 111 may be implemented in accordance with the present invention by selecting the tissue types of organoids 113 to be included in the humanoid 111 and their arrangement.

[0336] FIG. 9A shows an example of a humanoid 111 on a multi-human chip 110 constructed and operative in accordance with the present invention, comprising nine elements: (i) a "lung-type" perfused organoid unit 830 designated "Lung"; (ii) a "intestine-type" perfused organoid unit 830 designated "GI"; (iii) a "liver-type" organoid chamber 112 designated "Liver"; (iv) a "heart-type" organoid chamber 112 designated "Liver"; (v) a "kidney-type" perfused organoid unit 830 designated "kidney"; (vi) a "bbb-type" perfused organoid unit 830 designated "bbb"; (vii) a "brain-type" organoid chamber 112 designated brain; and a "whole body chamber".

[0337] It is understood that the above-mentioned perfused organoid unit 830 mimics the body's absorption and excretion and is important to support brain delivery, as follows: The "Lung" perfused organoid unit 830 mimics inhalation absorption, inhaled contents enter its left compartment and are absorbed into its right compartment, which mimics blood flow. The "GI" perfused organoid unit 830 mimics gastrointestinal absorption, ingested contents enter its left compartment and are absorbed into its left compartment, which mimics blood flow. The "Kidney" perfused organoid unit 830 mimics glomerular drainage, whose left compartment mimics the body's bloodstream contents, which are then filtered by glomerular tissue and drained into its left compartment. And the "bbb" perfused organoid unit 830 mimics the blood-brain barrier, whose left compartment mimics the body's bloodstream contents, which are then selectively filtered into its right compartment, which reflects cerebral circulation.

[0338] The "Liver" organoid chamber 112 is also important because it metabolizes the circulating contents passing through it, as it does in the body. "Brain" may contain specific brain tissues and / or loci that are of interest to the particular study at hand. Tissue-X is another tissue that the user may choose to include.

[0339] The whole body chamber is a chamber that does not contain any organoids therein, mimicking a "peripheral blood draw" in the body, and may be used as a means of "screening" on the basis of which further decisions are made to take samples from a particular organoid 113 or not. While the chip-on-chip system 100 operates to perform large scale repeat testing at an unprecedented scale, it is a dramatic advantage to take a "pre-screening" sample from the "whole body chamber" and determine based on these results whether further testing is warranted. Doing so may dramatically reduce the amount of testing performed, thus further supporting non-linear scaling and cost reduction. As a first example for utilizing this capability, if a sample from the "whole body chamber" reflecting a combined sample from all of the organoid chambers 112 of a humanoid 111 shows elevated lactate, this may prompt the taking of samples from all of the organoid chambers 112 of that humanoid 111 to determine which of them have elevated lactate. As another example, if testing from the "whole body chamber" produces results that may indicate potentially elevated levels of one or more liver enzymes, this may prompt taking a dedicated sample from the liver organoid chamber 112 to confirm, verify, and quantify this abnormality, similar to a whole body blood test in a living person.

[0340] Therefore, an important advantage of the inventive chip-on-chip system 100 is the flexibility in defining the configuration of the humanoid 111 according to the experimental design and requirements. The example provided in Fig. 9A encompasses the main modes of intake and excretion, but this comes at the expense of the number of other involved tissues. As an example, a particular experiment was precise using only the intestine alone (e.g., without lungs, kidneys, bbb) coupled with liver metabolism, which allows leaving room for the inclusion of many additional target tissues.

[0341] Reference is now made to FIG. 9B, which illustrates generally one example of the architecture of a humanoid 111 constructed and operative in accordance with the present invention.

[0342] FIG. 9B shows the humanoid 111 on a multi-human chip 110 described by FIG. 9A, but in this figure adds its interaction with the circulatory chip 142.

[0343] Figure 9B shows how eight circulation tubes 204 and one long circulation tube 205 are used by the circulation chip 142 and the precision robot 154 to connect all the organoids 113 of the humanoid 111 to one circulation that mimics the cardiovascular circulation in the body. As an example, Figure 9B illustrates how the humanoid 111 mimics the process of a drug 208 that is ingested orally, absorbed into the bloodstream by the intestine, reaches the liver in a manner that mimics the liver "first pass", is metabolized by the liver, and then filtered and excreted by the kidneys, and finally reaches several important target tissues.

[0344] Figure 9B also shows how each one of the organoid chambers 112 and perfused organoid units 830 is subject to an "intra-organoid" circulation, i.e., an internal independent organoid-specific flow through the organoid flow tube 114 (in the case of the organoid chamber 112) and through the organoid flow tube 114 and secondary circulation 814 (in the case of the perfused organoid unit 830).

[0345] It is understood that the architecture of the humanoid 111 is flexible and may include only the organoid chamber 112, or only the perfused organoid unit 830, or any combination thereof.

[0346] Reference is now made to FIG. 10A, which illustrates a schematic example of an integrated absorption and delivery chip-on-chip system 100.

[0347] It is understood that the chip-on-chip system 100 is designed and operates to simulate and mimic human absorption / administration, distribution, metabolism and excretion (ADME). Figure 10A shows three examples of absorption and administration.

[0348] The first use case, designated systemic administration 1010, involves administering nutrients or drugs 208 to the systemic chamber 1012 through the ADME tubes 206 of the ADME chip 143. This mimics systemic arterial and venous administration of the drug 208, as the drug 208 is distributed to all other organoids 113 of the humanoid 111 via the circulation chip 142, without the need for absorption via the intestine, etc.

[0349] The second use case, designated oral administration 1020, illustrates mimicking oral administration. Drug 208 is administered through the ADME tube 206 of the ADME chip 143 into the "intestinal-type" perfused organoid unit 830, designated "GI", into its left compartment, designated 1022, which mimics the intestinal lumen. Drug 208 is then absorbed both across the gastrointestinal membrane and adjacent capillary epithelium of the GI organoid unit 830 into the right compartment, designated 1024, which mimics the intestinal vasculature.

[0350] It is understood that drugs 208 administered orally into the human body are first absorbed into the bloodstream through the intestine, then circulate to the liver, where they are metabolized, a process known as "first-pass" metabolism. Similarly, in the chip-on-chip system 100, the drug 208 is absorbed by the perfused GI perfusion organoid unit 830, and then flows through the circulation provided by the circulation chip 142 to the sphere-shaped liver organoid 113 designated as "liver", where the drug 208 is actually metabolized by this three-dimensional liver tissue, and then continues to circulate to other "tissues" of the humanoid 111.

[0351] The third use case, designated as inhalation administration 1030, illustrates mimicking inhalation administration. Drug 208 is administered through the ADME tube 206 of the ADME chip 143 into the "lung-type" perfused organoid unit 830, designated as "lung", into its left compartment, designated as 1032, which mimics the ventilated alveolar space. Drug 208 then crosses both the alveolar membrane and adjacent capillary epithelium of the lung organoid unit 830, and is absorbed into the right compartment, designated as 1034, which mimics the alveolar vasculature. Drug 208 then flows through the circulation provided by the circulation chip 142 to the other "organs" of the humanoid 111.

[0352] Reference is now made to FIG. 10B, which illustrates a schematic example of integrated evacuation of the chip-on-chip system 100.

[0353] The first use case, designated "urinary excretion" 1040, shows mimicking urinary excretion. The drug 208 and its drug metabolites flow through the circulation provided by the circulation chip 142 and thus enter the "kidney-type" perfused organoid unit 830, designated "kidney", and enter its left compartment, designated 1042. The drug 208 and its metabolites then diffuse together across the adjacent layers of capillary epithelium and nephron glomeruli of the kidney perfused organoid unit 830 to its right compartment, designated 1044, which mimics the renal lumen, and then exit through the ADME tube 206 of the ADME chip 143 and may be collected, discarded, or analyzed as relevant.

[0354] The second use case, designated "GI Emptying" 1050, illustrates mimicking gastrointestinal emptying. The drug 208 and its metabolites flow through the circulation provided by the circulation chip 142, thus entering the "GI-type" perfused organoid unit 830, designated "GI", and entering its right compartment, designated 1052. The drug 208 and its metabolites then diffuse together across the capillary epithelium and adjacent gastrointestinal membrane of the GI perfused organoid unit 830 to its left compartment, designated 1054, which mimics the gastrointestinal lumen, and then exit through the ADME tube 206 of the ADME chip 143, where they may be collected, discarded, or analyzed as relevant.

[0355] The third use case, designated "Pulmonary Ejection" 1060, illustrates mimicking pulmonary ejection. The drug 208 and its drug metabolites flow through the circulation provided by the circulation chip 142, thus entering the "lung-type" perfused organoid unit 830, designated "Pulmonary", and entering its right compartment, designated 1062. The drug 208 and its metabolites then diffuse together across the capillary epithelium and adjacent alveolar membrane of the lung perfused organoid unit 830 to its left compartment, designated 1064, which mimics the ventilated alveolar space, and then exit through the ADME tube 206 of the ADME chip 143, where they may be collected, discarded, or analyzed as relevant.

[0356] Reference is now made to FIG. 10C, which illustrates a schematic of the parallel ADME capabilities of the chip-on-chip system 100.

[0357] Figure 10C shows how each of the ADME actions of the ADME chip 143 of the present invention may be preferably performed simultaneously on multiple perfused organoid units 830 of the same tissue type from different humanoids, designated as 1070.

[0358] 10C depicts six ADME tubes 206 of the ADME chip 143 being lowered into and simultaneously extracting fluid from the compartments of six respective perfused organoid units 830, designated 1071-1076. As an example, if the perfused organoid units 830, designated 1071-1076, are of the "kidney" type, FIG. 10C illustrates how the ADME chip 143 simultaneously removes urine from these six units.

[0359] It is understood that the multi-human chip 110, the automated chip 120 and the actuator 150 are all designed such that the automated chip 120 may perform actions simultaneously on multiple humanoids (other than chips such as the circulatory chip 142 that are specifically designed to perform actions on multiple organoid chambers 112 within the humanoid 111).

[0360] Reference is now made to FIG. 11A, which illustrates generally several preferred embodiments of a chemical sensing chip 131 constructed and operative in accordance with the present invention.

[0361] FIG. 11A shows a different embodiment of the invention relating to the manner in which the invention performs nanosensor-based chemical analysis of the fluid within the organoid chamber 112 or the organoids 113 therein.

[0362] The figure designated "Preferred embodiment" provides a schematic diagram of a preferred embodiment of the chemical sensing chip 131. In this preferred embodiment, a probe 1100 is inserted into the organoid chamber 112 such that a portion of the probe is immersed in the supernatant fluid. The probe 1100 is solid and not a capillary. Preferably, a segment of the probe 1100 is coated with one or more nanosensors 1102, and the probe is positioned such that the one or more nanosensors 1102 are immersed in the supernatant fluid. In a preferred embodiment of the present invention, multiple nanosensors 1102, such as electrochemical nanosensors attached (imparted) by nanocoating on a surface or probe, may be used, preferably in the same compartment, and each one of the nanosensors 1102 may indicate the level of a different target chemical. Each such electrochemical sensor may operate to detect the target chemical either directly through its charge, or in conjunction with a membrane coated on an electrode that makes the sensor specific to the desired chemical, or a membrane coated with an enzyme that makes the electrode specific to the detection of the desired target chemical, both as is well known in the art.

[0363] It will be appreciated that a plurality of such different electrochemical sensors may utilize a single "second electrode" that is electrically coupled to each of the sensors and thus "common" to these different sensors (electrically coupled means that this common electrode is coupled to each of the different electrochemical nanosensors such that it closes an electrochemical circuit with each of the different electrochemical nanosensors, respectively).

[0364] In another preferred embodiment of the present invention, a single electrochemical nanosensor may be preferably used to identify multiple different chemicals, where the range of levels of the different chemicals is generally known. As an example, a single electrochemical nanosensor may operate to distinguish between two chemicals even if their electrochemical charges are similar, by taking into account the physiological range of each of the two chemicals in a biological fluid. As an example, if the expected range of concentrations that Na+ is expected to be found in a given biological medium is many times higher than the range of concentrations that K+ is expected to be found in the given biological medium, it is possible to automatically "infer" from the sensor readings to "tentatively" distinguish between the two chemicals, as is well known in the art.

[0365] FIG. 11A further illustrates a further configuration, designated "other embodiment," of a chemical sensing chip 131 that may be used according to another preferred embodiment of the present invention.

[0366] The drawing designated "#1" provides a schematic diagram of a first preferred embodiment of a sample test probe. A capillary 1112 connected to an analyzer nanosensor chip 1100 is inserted into the organoid chamber 112 such that its tip is immersed in the supernatant fluid. Capillary forces draw fluid into the capillary 1112 and drive it to the nanosensors and / or sensor array contained within the analyzer nanosensor chip 1100. Washing and cleaning of the capillary after a sample has been tested and before testing a subsequent sample may be performed, preferably by contacting the open tip of the capillary with a porous surface (sponge, paper towel, etc.) that draws fluid out of the capillary and analyzer nanosensor chip 1100, also by exploiting surface tension. The capillary 1112 may preferably be a single capillary (not shown), or the capillary 1112 may preferably branch into multiple sub-capillaries, thus further using capillary forces to drive fluid to multiple sensors contained within the analyzer nanosensor chip 1100. The branching sub-capillaries may preferably be thinner than the main capillary.

[0367] The drawing designated "#2" provides a schematic diagram of a second preferred embodiment of the sample test probe. In this preferred embodiment, one or more fluorescence-based nanosensors 1120 are fixed to the bottom inner surface of the organoid chamber 112 or are free floating in the fluid within the organoid chamber 112. The fluorescence-based nanosensors 1120 may preferably comprise fluorescent beads or fluorescent coatings, or other fluorescence-based nanosensors, and may operate to detect levels of oxygen or other biologically relevant chemicals. A photodetector 1122 is used to detect the amount of light emitted from the fluorescence-based nanosensors 1120 and indicative of the chemicals they operate to detect, as is well known in the art. This photodetector is automatically actuated by an actuator 150. It is understood that if the nanosensors 1102 are fixed to the bottom inner surface of the organoid chamber 112, gravity will likely bring the organoids 113 into direct contact with them or in close proximity to them, giving a more accurate reading.

[0368] The drawing designated "#3" provides a schematic diagram of a third preferred embodiment of a sample test probe. In this preferred embodiment, a capillary 1134 is used, but rather than connecting to the analyzer nanosensor chip 1100, the capillary 1134 extends at its distal end into a nanosensor cavity 1130 that contains a plurality of nanosensors 1132. As an example, the nanosensors 1132 may be nanocoated on the inner surface of the nanosensor cavity 1130. It is understood that the advantage of this third embodiment is that the nanosensors are not inserted into the supernatant fluid. As is well known in the art, some nanosensors may release ions into the fluid, and thus may harm and / or otherwise affect the biological fluid they are testing. Therefore, this third embodiment uses this simple probe design, but the fluid is drawn into the nanosensor cavity 1130, where it interacts with the nanosensors 1132 outside of the organoid chamber 112 and the fluid.

[0369] It will be further understood that while this embodiment is not logically very different from the first embodiment designated "#1," there are important differences in practical construction. This third embodiment is the simplest, being merely a "cavity" that spans the end of the capillary, and one or more nanosensors 1132 may preferably be coated or otherwise embedded within the cavity.

[0370] The drawing designated "#4" provides a schematic diagram of a fourth preferred embodiment of a sample test probe. In this preferred embodiment, a flexible probe 1140 is used. One or more nanosensors 1142 are coated on corresponding segments of the flexible probe 1140. According to a preferred embodiment of the present invention, the flexible probe 1140 is designed such that it preferably "flaps" to achieve contact with the organoid 113, thereby bringing the nanosensors 1142 coated thereon into direct contact or very close proximity to the cells of the organoid 113. A photodetector 1144 may be used in conjunction with a fluorescent nanosensor, such as for the detection of oxygen and other biologically relevant chemicals.

[0371] The above embodiments are provided by way of example only, and it is understood that combinations of these may be used preferably. As an example, in a preferred embodiment of the present invention, the flexible probe 1140 of embodiment "#4" may be preferably used with nanosensors 1142 that are fluorescent and / or do not significantly release ions and therefore may be safely immersed in a test sample without significantly affecting the test sample, with the "slack" of the flexible probe 1140 achieving physical contact or proximity of the fluorescent sensor to the organoid 113 itself. This may preferably be used in conjunction with the aspect of embodiment "#3", where electrochemical sensors may be used in the nanosensor cavity 1130 at the distal end of the capillary 1134, so that these electrochemical sensors are not immersed in and do not affect the supernatant fluid surrounding the organoid 113 in the organoid chamber 112.

[0372] In all of the above embodiments of the present invention, it is understood that the nanosensor may also include one or more fluorescence-based sensors (similar to those in embodiment "#2") for detecting levels of oxygen and other biologically relevant chemicals, as is well known in the art. A photodetector 1212 is used to detect the amount of light emitted from the fluorescence-based sensor, which may preferably not be part of the probe 1100 and is automatically operated by the actuator 150.

[0373] Reference is now made to FIG. 11B, which illustrates a schematic of the real-time cell sensing architecture.

[0374] The chemical sensing chip 131 enables real-time cellular metabolic sensing, which is a unique advantage of the present invention. Figure 11B illustrates a unique functional and architectural concept that enables real-time cellular metabolic sensing.

[0375] 11B provides an abstracted illustration of the interplay between some important characteristics of the multi-human chip 110 and chemical sensing chip 131 of the present invention.For clarity, the example provided is a sphere-shaped organoid 113, but it is understood that the same applies to perfused organoids or other types of organoids.

[0376] The volume of the organoid chamber 1150 is equal to the volume in the organoid chamber 112 minus the volume of the organoids 113. The tighter the fit, the smaller the volume of the organoid chamber 1150. As an example, the organoids 113 are typically very small, e.g., 100-300 microns (100-300 μm) in diameter. Thus, the organoid chamber 112 is designed and fabricated to be as small as practicable while still being sufficient to accommodate the organoids 113, while taking into account further factors described herein below. As an example, if the organoid chamber 112 has a diameter of less than 1 millimeter, assuming the organoids have a diameter of 100-300 microns (100-300 μm), the volume of the organoid chamber 1150 will be very small, as opposed to when the organoid chamber 112 is much larger, e.g., 4 millimeters.

[0377] As an example, in a preferred embodiment of the present invention, the organoids may typically be 100-300 microns (100-300 μm) in diameter, and the organoid chamber may be 400 microns (400 μm) or 500 microns (500 μm) or 1000 microns (1000 μm). Thus, in such an embodiment, the ratio of the volume of the organoid chamber to the volume of the organoid is 2.8 (500 micron chamber to 300 micron organoids), 25 (500 micron chamber to 100 micron organoids), 11 (1000 micron chamber to 300 micron organoids), or 100 (1000 micron chamber to 100 micron organoids).

[0378] The volume of the sample tube 1158 is the volume of the entire tube from the organoid chamber 112 to the nanosensor chamber 1154. The longer and wider the sample tube 1152, the larger the volume of the sample tube 1158. Nanosensor Chamber Volume 1159is simply the internal volume of the nanosensor chamber 1154 minus the volume of the nanosensor 1156. In reality, the nanosensor 1156 will typically have a tiny volume and the nanosensor chamber 1154 will be much larger, simply depending on the fabrication used. As an example, current nanosensor arrays often have nanosensor chambers 1154 that are a few microns in size, while the nanosensors themselves may be a thousandth of that size, i.e., measured in nanometers.

[0379] The sample volume 1160 is the volume that must be removed from the organoid chamber 112, flow through the sample tube 1152, and placed into the nanosensor chamber 1154 so that the nanosensor 1156 can make its measurement. The sample volume 1160 is therefore the sum of the volume of the sample tube 1158 and the nanosensor chamber volume. 1159 1156. Briefly, the longer and wider the sample tube 1152 and the larger the nanosensor chamber 1154, the larger the sample volume 1160. In practice, the sample volume may be several times larger than described above, but it is understood that this is because the above description describes taking a sample only to a single nanosensor 1156, whereas in practice the sample taken from the organoid chamber 112 may need to be sufficient to contain multiple samples delivered to each of the multiple nanosensors 1156 included in the nanosensor array.

[0380] It will be appreciated that the volume of the organoid chamber 1150 strongly influences the ability to obtain sensitive and accurate measurements of the metabolites released by the organoids 113 in the supernatant surrounding the organoids 113 in the organoid chamber 112. The smaller the volume of the organoid chamber 1150, the more accurately the sample taken from the organoid chamber 112 will reflect the metabolites released by the cells of the organoids 113. This may be illustrated by the following example: If an ice cube is placed in a very small glass of whiskey, the ice cube will dilute the whiskey much more than if the same ice cube were placed in a large barrel of whiskey. many .

[0381] It is also understood that the sample volume 1160 strongly influences the frequency of measurements that may be taken from the organoid chamber 112. In simple terms, in the extreme case, if the sample volume 1160 is equal to the volume of the organoid chamber 1150, taking one sample will completely "empty" the organoid chamber 112. Conversely, if the sample volume 1160 is 1 / 1000th the volume of the organoid chamber 1150, 100 measurements may be taken without significantly depleting the organoid chamber and thus significantly disturbing the organoids 113. This may be illustrated by the following example. If you were to hire a group of expert wine tasters and have each taste a glass of wine, repeatedly every 15 minutes over the course of several days, to measure how the wine changes as it sits in the glass and aeration occurs, it would be reasonable for each wine taster to take a small, sippy sip, whereas if the wine tasters took large gulps, the tasting would quickly empty the glass completely.

[0382] It is further understood, and is well known in the art, that placing nanosensors or other measurement devices or compounds within or directly adjacent to the organoid chamber 112 itself can be disadvantageous in that nanosensors may not be harmless and typically release ions into the immediate environment, which can affect the tissue or fluid the nanosensor is testing.

[0383] Therefore, FIG. 11B shows the central dual uniqueness of the present invention, firstly, the ability to obtain accurate "cellular" metabolic sensing, and secondly, the ability to do so in real time, i.e., to provide such measurements frequently and repeatedly. "Cellular" metabolic sensing or highly sensitive metabolic sensing of metabolites released by organoids 113 into organoid chamber 112 is achieved by the unprecedented miniaturized design and fabrication of organoid chamber 112, which keeps the size and shape of organoid chamber 112 close to that of organoids, so that the volume of organoid chamber 1150 is small. Real-time frequent and repeated testing is also made possible by the unique architecture and sub-micron fabrication, which minimizes the length and width of sample tube 1152 and the size of nanosensor chamber 1154, thereby minimizing sample volume, thereby achieving a high ratio between the volume of organoid chamber 1150 and sample volume 1160.

[0384] It is understood that the chip-on-chip architecture reduces the length and width of the sample tube 1152 to an absolute minimum by bringing the sensor closer to the organoid chamber 112 via the actuator 150. Thus, the chip-on-chip architecture allows two microfluidic chips to "act as one" with the advantages of (i) shortening the length of the sample tube 1152, (ii) eliminating the complexity of multiple valves / channels / pumps, thereby (iii) reducing cost, and (iv) improving robustness.

[0385] As an example, in a preferred embodiment of the present invention, the organoids 113 may preferably be one hundred to three hundred microns (100-300 μm) in diameter, the organoid chamber 112 may preferably be four hundred microns (400 μm) or smaller in diameter, and the sample tube 1152 may be sub-micron in diameter and several millimeters in length.

[0386] As a practical example, taking a sample into a single sensor once every 15 minutes means 1344 samples over two weeks, and / or 13440 samples if 10 sensors are used. Therefore, assuming at least one sensor, and preferably more than 10 sensors, repeated real-time measurements will require a ratio between the volume of the organoid chamber 1150 and the sample volume 1160 of at least 1,000, preferably 10,000.

[0387] In preferred embodiments of the present invention, the ratio between the volume of the organoid chamber 1150 and the sample volume 1160 is preferably greater than 100,000, greater than 50,000, greater than 10,000, greater than 1000, greater than 500, greater than 100, greater than 50, or greater than 10.

[0388] In preferred embodiments of the invention, the ratio of the volume of the organoid chamber to the volume of the organoids is less than 1, less than 2, less than 3, less than 5, less than 10, or less than 30, less than 100, or less than 500.

[0389] In a preferred embodiment of the present invention, (i) the size of the organoid chamber 112 is kept to a practical minimum so that the fluid reflects the organoid 113 as closely as possible, and (ii) a nanosensor is used that can be safely inserted into the fluid in the organoid chamber 112 without affecting the media fluid, so that measurements are made directly in the chamber without the need to take a sample even if the sample is a very small volume and through a short tube. It is also understood that when there are practical limitations of size and volume, for example the volume of the organoid chamber 112 or nanosensor chamber is significantly smaller than 1 mm, evaporation becomes a limiting issue. Therefore, again, a nanosensor that is inserted into the organoid chamber 112 rather than extracting a sample is preferred.

[0390] Reference is now made to FIG. 12 which illustrates a schematic diagram of the use of specific and non-specific sensor arrays.

[0391] The present invention utilizes two types of nanosensor arrays: specific sensor arrays and non-specific sensor arrays. As illustrated by Figure 11A, the chemical sensing chip 131 preferably includes a probe 1100 having a plurality of nanosensors 1102 coated or attached thereto.

[0392] The first figure, designated "Probe with Specific Sensor Array", shows the probe 1100 with an attached group of nanosensors, designated specific sensor array 1210. The specific sensor array 1210 is an array of nanosensors, where each nanosensor is fabricated and trained to identify a specific chemical, as is well known in the art. In a preferred embodiment of the present invention, the specific sensor array 1210 may include, but is not limited to, sensors for oxygen, glucose, lactate, urea, potassium, ammonia, and pH. As depicted in FIG. 12, the specific sensor array 1210 indicates different levels of each of these metabolites for a sample to which the array is exposed. As is well known in the art, these measurements indicate the vitality of cells in the tissue, and the vitality of mitochondria in those cells. These measurements are therefore useful to indicate the response of cells in that tissue to an administered drug, including the toxicity of the drug to the tissue.

[0393] The second diagram, designated "Probe with Non-Specific Sensor Array", shows the probe 1100 with an attached group of nanosensors, designated specific sensor array 1210. The non-specific sensor array 1220 is an array of nanosensors operatively constructed such that the sensor arrays work together as a group and may "learn" to collectively identify a very wide range of chemical compounds with the aid of artificial intelligence, as is well known in the art. The non-specific sensor array 1220 "learns" to identify patterns typical of the chemical compounds. The non-specific sensor array 1220 "learns" to identify the particular patterns depicted here of the levels of signals registered by each of the seven sensors shown in this diagram, and thus "learns" that such patterns are typical of particular chemical compounds. All this as is well known in the art.

[0394] It is understood that the above are intended as examples only and are not intended to be limiting. In other preferred embodiments of the present invention, a combination of specific and non-specific sensors may be used, and multiple probes 1100 may be used simultaneously in the same organoid chamber 112 to use more nanosensors at one time. The third figure, designated "Analyzer Nanosensor Chip," shows an analyzer nanosensor chip 1110, in which the sensors of the analyzer nanosensor chip 1110 may include a specific sensor array 1230 and a non-specific sensor array 1240. A capillary 1112 is used to draw a minute sample from the fluid in the organoid chamber 112 and deliver it to each of the sensors in the sensor array. These specific and non-specific sensor arrays are similar to those described above, but rather than being submerged in the fluid of the organoid chamber 112 and measuring chemicals there directly, they are placed on the analyzer nanosensor chip 1110 and therefore analyze the drawn sample.

[0395] Optionally, an ITP concentrator engine 1250 The ITP Concentrator Engine is used to concentrate chemical compounds within the withdrawn sample. 1250 is a microfluidic device that uses isotachophoresis, as known in the art, to concentrate chemical compounds within a sample fluid drawn through a sample channel 290.

[0396] Reference is now made to FIG. 13A, which illustrates a schematic diagram of one example of genome nanosensing performed by genome chip 132.

[0397] Organ chips open up powerful opportunities to dramatically accelerate drug development, especially when multiple such organ chips are utilized. The present invention overcomes two major obstacles that exist to date and allow the present invention to actually dramatically accelerate drug development. First, there is currently no high-throughput and AI-integrated human-on-chip platform, i.e., a system that allows for rapid and inexpensive biological scanning of drug candidates on tens of thousands of human-on-chips. Second, there is currently no human-on-chip platform with integrated genomic expression profiling.

[0398] The present invention addresses both of these challenges. Our chip-on-chip architecture addresses the first challenge, providing the first truly high-throughput real-time sensing human-on-a-chip platform. And our nanowire-based non-amplified genomic sensing, especially sensing of circular RNA, provides a revolutionary solution to the second challenge.

[0399] Genomic-related challenges of current Organ-Chip systems include: Organ-Chip systems are very promising as they show predictive accuracy that sometimes exceeds that of animal studies. However, to date, animal studies have had one unique advantage. Animal studies have yielded a wealth of genomic data, but Organ-Chip systems, due to their miniaturization, simply cannot match that wealth of genomic data. Attempting to develop drugs without genomic data is akin to "guesswork."

[0400] Therefore, the present invention is unique in that it now provides non-destructive, real-time, non-amplified detection of circular RNAs, resulting in unparalleled and unprecedented genetic profiles associated with most major diseases. The ability to accurately identify circular RNAs within a high-throughput human-on-chip platform will have dramatic impact. As an example, approximately 330 circRNAs have been shown to be associated with 48 major diseases, including various types of cancer, neurodegenerative disorders, cardiovascular diseases, autism, and many other major diseases.

[0401] FIG. 13 shows an ultrasensitive nanosensor and system that operates to accurately identify and quantify, without amplification, minute amounts of circular RNA molecules secreted by organoids 113 into the supernatant fluid in the organoid chamber 112 in a high-throughput multi-human chip 110.

[0402] Circular RNAs (circRNAs) are novel short RNA genes that are circular and have been linked to multiple major diseases, including cancer, heart disease, neurodegenerative diseases, autism, and many other entities, and are highly expressed in the brain and associated with cell differentiation. Surprisingly, unlike "normal" mRNAs, thousands of circRNA molecules have been found to be "cell-free," meaning outside the cell, in these various major disease states. The circular structure of circRNA molecules makes them resistant to ribonucleases, unlike normal "linear" mRNAs, which collapse when outside the cell.

[0403] The circ-RNA genome nanosensor 1300 is a nanosensor device that is an example of the genome chip 132. The circ-RNA genome nanosensor 1300 includes a probe 1100, on which multiple genome sensors are attached or coated.

[0404] The circ-RNA genomic nanosensor 1300 is lowered into the fluid in the organoid chamber 112. The organoid chamber 112 contains organoids 113 and is subjected to a controlled, actively pumped microfluidic flow through the organoid flow tube 114, which is important for the activation of the organoids 113. The organoids 113 naturally secrete cell-free circRNA molecules into the fluid of the organoid chamber 112, and this circRNA is different from regular longer mRNA in that it frequently survives in a cell-free form in the fluid of the organoid chamber 112 without being degraded, and thus may be identified by the circ-RNA genomic nanosensor 1300.

[0405] The circ-RNA genomic nanosensor 1300 operates to detect specific circRNA molecules of interest, e.g., circRNAs that have been shown to correlate with specific disease states.

[0406] In the example provided by Figure 12, the circ-RNA genomic nanosensor 1300 includes three genomic nanosensors, each of which can bind to and detect a specific circular RNA molecule. A first sensor, designated circ sensor 1302, identifies a circular RNA molecule designated circRNA 1312. A second sensor, designated circ sensor 1304, identifies a circular RNA molecule designated circRNA 1314. A third sensor, designated circ sensor 1306, identifies a circular RNA molecule designated circRNA 1316.

[0407] Each of these three sensors includes a nucleic acid molecule designed with an antisense sequence to the nucleic acid sequence of its target circRNA, which hybridizes and binds to its target but does not bind to other molecules, including other circRNA molecules. Once a sensor, preferably a nanowire-type nanosensor, binds to its target circRNA, this binding changes the electric field properties of the sensor, providing an electronic indication of the binding.

[0408] As an example, Figure 13 shows three circRNAs in the fluid of the organoid chamber 112, designated as 1322, 1324, and 1326, respectively; these circRNAs do not match the nucleic acid sequences that the sensor is designed to bind, and therefore they are not bound and are not recognized by the sensor.

[0409] In a preferred embodiment of the present invention, the sensor may be designed with an antisense sequence that specifically binds to the targeted circRNA, as known in the art.

[0410] One aspect of the present invention is that the circ-RNA genomic nanosensor 1300 may preferably be designed such that its binding to its target is enhanced using the unique physical properties of circRNA due to its circular structure, as opposed to the physical properties of linear, non-circular RNA.

[0411] As an example, a "tail-to-head probe" may preferably be designed that binds to a circRNA sequence only if the circRNA sequence is indeed in a circular form, and does not bind to the circRNA if the RNA is in a linear form.

[0412] As another example, the sensors may be designed such that their binding nucleic acids are themselves in a circular form, thereby not only matching the nucleic acid sequence of their target, but also such that their circular form significantly enhances their binding to the target.

[0413] In yet another example, the sensors may be designed to exploit other physical properties of their targets that stem from the circular physical properties of circRNAs, including, but not limited to, the absence of polyA tails, the absence of free 5' and 3' ends, and the circular ribose backbone. Such exploitation means that not only do sensors bind to their targets because their antisense probes specifically bind to complementary sequences within the target circRNAs, but this antisense binding is further enhanced by these circular physical properties.

[0414] It is understood that the key part of the present invention is to exploit the unique structure of circRNA to achieve high sensitivity and accurate detection by nanosensors.The main challenge of highly sensitive nanosensors, especially nanowire-based sensors, is that they are very sensitive and the type and size of their target greatly affect the electronic signal.Target of different size requires different calibration, which makes the design, calibration, manufacture and operation of such sensors very difficult and expensive to such an extent that it is impractical to pursue the rapid development of hundreds or thousands of sensors to detect different targets.As an example, when binding to an RNA target, a longer targeting RNA molecule will give a very different signal than a shorter targeting RNA molecule.

[0415] Therefore, the present invention provides two inventive steps that make this sensor feasible. First, by focusing on targeting a group of circRNAs. Since they are all, as a group, relatively small molecules, and importantly, relatively consistent in size, mostly around 300-360bp, this dramatically reduces the within-group variance of the electronic signal of circRNA binding by the nanowire sensor compared to protein binding (very large variance), or regular mRNA binding (since mRNAs vary in length from tens to thousands of nucleotides). Second, by using the unique physical properties of circRNAs that arise from their circular structure and distinguish circRNAs from those of linear non-circular RNAs.

[0416] Reference is now made to Figure 13B, which shows a schematic example of circ-RNA folding. Figure 13B shows the two-dimensional structures of two exemplary circ-RNAs. As Figure 13B shows, circRNAs fold on themselves to form two-dimensional structures that are highly unique to each circRNA, typically forming a central circle with one or more double-stranded (DS) "hairpin-like" segments protruding from this circle.

[0417] The first example in Figure 13B, designated "Circular RNA Example 1," shows the folding of a circRNA named circCAMSAP1, which is 425 nucleotides long and has three double-stranded hairpins. The second example in Figure 13B, designated "Circular RNA Example 2," shows the folding of a circRNA named circFCHO2, which is 268 nucleotides long and has two double-stranded hairpins.

[0418] Understanding the secondary structure folding of each individual circular RNA is important when attempting to detect circRNAs by nanosensors, because designing an antisense probe that binds to a sequence of the circRNA that is within the double-stranded RNA hairpin can significantly reduce its ability to bind to the circRNA.

[0419] FIG. 13B further shows the back splice junction (BSJ) of the two circRNAs shown. Each circRNA has a back splice junction, which is a point in the circular RNA where the "head" and "tail" of the linear sequence of the circRNA join to form a circle, as is well known in the art. It is understood that the BSJ may be located within the circular single-stranded portion of the circRNA or within the double-stranded hairpin portion. In the example provided in FIG. 13B, it is understood that in "Example 1" the BSJ is within the circular portion of the circRNA, whereas in "Example 2" the BSJ is within the double-stranded hairpin section.

[0420] In a preferred embodiment of the present invention, this aspect is analyzed and taken into account when designing suitable probes for identifying circRNAs.

[0421] Reference is now made to FIG. 13C, which illustrates a schematic example of circ-RNA probe design.

[0422] The figure designated "Linear circ-RNA" represents the structure of circ-RNA in linear form. Linear circ-RNA 1330 is a circular RNA molecule, which is a type of RNA molecule characterized by spontaneously forming a circular shape, which is more resistant to degradation by ribonuclease enzymes, as is well known in the art. As this figure shows, linear circ-RNA 1330 comprises a head 1332 and a tail 1334. Head 1332 is the leading nucleic acid sequence, meaning the 5' end, of the nucleic acid sequence of linear circ-RNA 1330. Tail 1334 is the trailing nucleic acid sequence, meaning the 3' end, of the nucleic acid sequence of linear circ-RNA 1330.

[0423] The figure designated "looped circ-RNA" represents the structure of circ-RNA in its "looped" or circular form. It is understood that looped circ-RNA1340 is the same RNA molecule as linear circ-RNA1330 and contains the same nucleic acid sequence. The only difference between looped circ-RNA1340 and linear circ-RNA1330 is that looped circ-RNA1340 forms a circular shape, with tail1334 connected to head1332.

[0424] In a preferred embodiment of the present invention, the tail-to-head probe 1342 is preferably designed to have a nucleic acid sequence that includes an antisense to the nucleic acid sequence of the tail 1334 followed by an antisense to the nucleic acid sequence of the head 1332. Therefore, the tail-to-head probe 1342 selectively and strongly binds to looped circ-RNA 1340 in which the nucleic acid sequence of the tail 1334 is adjacent to the nucleic acid sequence of the head 1332. However, the tail-to-head probe 1342 binds only weakly to linear circ-RNA 1330 in which the nucleic acid sequences of the head 1332 and tail 1334 are not adjacent.

[0425] As an example, the tail-to-head probe 1342 may preferably be 20 nucleotides long, which is an effective length for antisense probes as known in the art, and the tail 1334 and head 1332 are each 10 nucleotides long. In such a scenario, it is understood that when binding to looped circ-RNA 1340, the entire 20 nucleotides of the tail-to-head probe 1342 bind to the tail 1334 sequence and the adjacent head 1332 sequence. On the other hand, when attempting to bind to linear circ-RNA 1330, only 10 of the 20 nucleotides of the tail-to-head probe 1342 sequence may bind to the linear circ-RNA 1330, binding to either the head 1332 or the tail 1334, but not both simultaneously. It is understood that the sensing may be calibrated to only distinguish strong binding of the entire 20 nucleotides and not partial binding of only 10 nucleotides, as known in the art.

[0426] The point in the sequence of looped circ-RNA 1340 where tail 1334 joins head 1332 is known in the art as a back splice junction, or BSJ, a term for how circ-RNA is formed. The figure designated "looped circ-RNA" shows two possible probe designs. The first probe design is a tail-to-head probe 1342 designed to target the BSJ region of looped circ-RNA 1340 by targeting the adjacent tail 1334 and head 1332 sequences.

[0427] The second probe design addresses any sequence within looped circ-RNA 1340 that does not contain a BSJ, where such a sequence of looped circ-RNA 1340 is designated non-BSJ 1344 and its antisense probe is designated non-BSJ probe 1346.

[0428] It will be appreciated that the tail-to-head probe 1342 is far more advantageous than the non-BSJ probe 1346 in several important respects. First, the tail-to-head probe 1342 binds to the circular RNA molecule only in its looped circ-RNA 1340 form and not in its linear circ-RNA 1330 form. Second, the tail-to-head probe 1342 has much higher specificity for identifying circular RNAs since circ-RNAs are often putative and thus the tail-to-head probe 1342 ensures the detection of real circular RNAs as opposed to hypothetical ones. Third, the tail-to-head probe 1342 has much higher specificity and a lower risk of off-target binding, meaning unintended binding to other genomic sequences. Specifically, since circRNAs are often part of the mRNA of a gene, the non-BSJ 1344 sequence is therefore also part of one or more mRNAs in which the circRNA is included. Thus, the non-BSJ probe 1346 may unintentionally bind to one or more of these mRNAs, including circRNAs, while the tail-to-head probe 1342 only binds to the looped circRNA 1340. Fourth, because the nucleic acid sequence of the tail-to-head probe 1342 is not antisense to a naturally occurring nucleic acid sequence, but is antisense to an artificial combination of two nucleic acid sequences, head 1332 and tail 1334, the tail-to-head probe 1342 has much higher specificity because it is less likely to unintentionally bind to "off-target" sequences elsewhere in the genome. As is well known in the art, due to the evolutionary process of duplication and conservation, naturally occurring genomic sequences are more likely to have genomic sequences found elsewhere in the genome that are similar to the naturally occurring genomic sequence than to find artificial sequences elsewhere in the genome.

[0429] It is understood that if the circRNA of interest is one in which the BSJ is found within a double-stranded hairpin, as depicted in "Example 2" in Figure 13B, the tail-to-head probe 1342 may not bind optimally to the looped circRNA 1340. This is because the tail 1334 and head 1332 sequences are found within a double-stranded secondary fold of the RNA molecule, and are therefore bound to the complementary RNA strand, making them less likely to bind to the probe. In such cases, the use of a non-BSJ probe 1346 is more effective since the BSJ is within the double-stranded region of the circRNA, and therefore it is preferable to select a non-BSJ 1344 nucleic acid sequence that is within the loop region of the circRNA rather than within the double-stranded region of the circRNA, to which the non-BSJ probe 1346 may easily bind without being hindered by double-stranded RNA binding.

[0430] Reference is now made to FIG. 13D, which illustrates a schematic flow chart of the method for circ-RNA probe design.

[0431] In a preferred embodiment of the present invention, the following steps are performed for each of multiple circ-RNA sequences desired to be measured, e.g., multiple circRNAs known to be associated with a disease of interest, as known in the art.

[0432] First, in step 1350, a circ-RNA sequence 1350 is received or otherwise stored in the memory of the system disclosed herein. This circ-RNA sequence may be one of a plurality of circ-RNA sequences known to be associated with a disease or biological process of interest.

[0433] Next, step 1352 discloses calculating the RNA folding of the RNA molecule of the circ-RNA sequence 1350 using various algorithms and software well known in the art. This step typically reveals one circular region and the major circular region of the circ-RNA sequence 1350, and typically one or more double-stranded hairpin structures.

[0434] Next, step 1356 determines whether the back-splice junction of the circ-RNA sequence 1350 is found within the double-stranded hairpin region or within a non-double-stranded region. Methods for finding back-splice junctions of circRNAs are well known in the art.

[0435] If the BSJ is found in a non-double stranded region of the circ-RNA sequence, the process proceeds to step 1358 to design a unique antisense tail-to-head probe. This may be done by designing the tail-to-head probe 1342 such that the probe 1342 contains an antisense sequence to the BSJ region of the circ-RNA sequence 1350.

[0436] If a BSJ is found in a double-stranded region of the circ-RNA sequence 1350, the process proceeds to design a unique antisense probe. This may be done by designing a non-BSJ probe 1346 such that the probe 1346 contains an antisense sequence to a non-BSJ 1344 sequence that is not within the double-stranded region of the circ-RNA sequence 1350.

[0437] In both of the above cases, it is understood that the probes are designed to target nucleic acid sequences within the circ-RNA sequence 1350 that are unique within the genome, have the lowest similarity to sequences within the genome, and follow other criteria for designing antisense probes, as is well known in the art, to achieve optimal probe-target binding and minimal unintended off-target binding.

[0438] The method then includes a step 1362 of generating an antisense probe by synthesizing a probe according to the nucleic acid sequence determined by the previous steps, designated 1358 and 1360, respectively. The synthesis of antisense probes, including chemical modifications that render such probes stable and effective, is well known in the art.

[0439] Next, the method includes attaching the antisense probes generated in the previous step to a nanosensor, such as a nanotip or nanowire sensor, as is well known in the art, step 1364. Many nanosensors are known in the art, and many methods are known in the art for attaching nucleic acid probes to nanosensors.

[0440] Next, the method includes a step 1366 of detecting the nanochip detecting cell-free circ-RNA 1366. This step may utilize a nanochip having immobilized nucleic acid probes to detect the cell-free circular RNA.

[0441] It is understood that the above description is intended as merely an example and is not intended to be limiting. In other preferred embodiments, the RNA to be detected may not be circular RNA or other types of short RNA or other types of RNA. In other preferred embodiments of the present invention, different nanosensors may be used, different probe design methods may be applied, and methods of amplification may be used, as known in the art.

[0442] Reference is now made to FIG. 14A, which illustrates a schematic diagram of one example of non-destructive genomic analysis.

[0443] FIG. 14A shows a unique and innovative embodiment of the present invention that allows for the genomic expression profile of a particular organoid 113 to be obtained both before and after the organoid is exposed to a drug.

[0444] In this respect, it is understood that there are two important aspects of uniqueness. First, currently, there is no organ chip that provides automated high-throughput genomic expression analysis of organoids in organ chips. Second, currently, there is no organ chip method or device that performs genomic analysis of organoids in a non-destructive manner. Performing genomic expression analysis typically requires dissolution of tissue, and therefore its destruction. Therefore, it is currently not possible to perform genomic analysis of the same organoid tissue in organ chips before and after drug intervention. It is understood that comparing genomic analysis such as expression profiling or epigenetic profiling of two different organoids, one treated and the other untreated, is not ideal because the two organoids are not identical.

[0445] The present invention achieves non-destructive genome sequencing analysis by separating a small number of single cells from spherical organoid before and after applying drugs to organoid, and sending these cells for single-cell genome analysis.This process is described as follows:

[0446] Pump 1410 pumps circulating fluid into the organoid chamber 112 containing the organoids 113.

[0447] As depicted by Figure 14A, the pump 1410 applies a shear force on the organoid 113, thereby detaching one or more single cells 1400. The system may include a cell counter 1420 operable to count cells flowing through the microfluidic channel. The one or more single cells 1400 flow out of the organoid chamber 112 and are counted by the cell counter 1420.

[0448] It is understood that in order to shear a small number of single cells 1400, the cell counter 1420 may preferably operate in conjunction with the pump 1410. As an example, the flow and / or pressure generated by the pump 1410 may start low and then be gradually increased until the desired small number of single cells 1400 are detached from the organoids 113 and detected by the cell counter.

[0449] It is also understood that when designing the exact implementation of the above mechanism, including but not limited to the location, angular aperture and pressure of the channel, and its location relative to the organoid chamber 112, a similar methodology can be adopted, that is, using cell counter 1420 as feedback to enable determining the best setting to separate desired single cells 1400 without damaging the spherical organoid, not too few and not too many.In this way, it is further understood that different settings can be used for different tissue types, and the above method is not only relevant to spherical organoids, but also to perfused organoids and other types of organoids.

[0450] Optional storage array 1430 may optionally store single cells 1400 for later passage to an external device for external testing, as described further herein below. The need for optional storage array 1430 is because the mechanism described herein receives a small number of single cells 1400 from each of a plurality of very large organoids 150, and therefore it may be advantageous to store a small number of single cells 1400 from each organoid 113 in a separate compartment within optional storage array 1430 so that all single cells 1400 may be transferred to an external device for external processing, such as genome sequencing. It is further understood that optional storage array 1430 may be a storage location on multi-human chip 150 or may be a separate device from multi-human chip 150.

[0451] Thus, the single cell 1400 may be captured, optionally stored, and then transferred to an external device for external analysis. Such further analysis of the single cell 1400 may be performed using an external single cell genome sequencer 1440 and an external single cell epigenetic analyzer. 1450 Including, but not limited to:

[0452] Reference is now made to FIG. 14B, which illustrates a schematic of three embodiments of chip-on-chip cell shearing as part of the non-destructive genomic analysis aspect of the present invention.

[0453] In an embodiment of the "chip-on-chip" architecture of the present invention, as shown generally in FIG. 1A hereinabove, cells from organoids 113 are shear detached by cell extraction nanotubes 202 of a cell extraction chip 133. These extracted single cells may then be analyzed by single cell genomic or epigenetic analysis, or other desired forms of analysis. Some embodiments are operative to actuate such shear detachment of cells from organoids 113.

[0454] A first embodiment of chip-on-chip cell shear detachment is shown diagrammatically by the drawing designated "#1". In this embodiment, the cell extraction nanotube 202 is a single tube with a shear flow pump 1460. In this embodiment, shear detachment of single cells from the organoid 113 is achieved by two actions diagrammatically by the two drawings designated "Pulse-1" and "Pulse-2", respectively. In the first action, shown by the drawing designated "Pulse-1", the shear flow pump 1460 pumps a pulse of fluid into the organoid chamber 112 to pull one or more single cells 1400 away from the organoid 113. In a second action, illustrated by the drawing designated "Pulse-2", the shear flow pump 1460 reverses its direction and pumps a pulse of fluid out of the organoid chamber 112 to aspirate one or more single cells 1400 that are free floating in the fluid in the organoid chamber 112 after being detached from the organoid 113 by the previous action of "Pulse-1". In a preferred embodiment of the invention, the cell extraction nanotube 202 connects to a fluid reservoir (not shown) from which fluid is pumped in the drawing designated "Pulse 1" and into which fluid flows in the drawing designated "Pulse 2". In another preferred embodiment of the invention, separate reservoirs are used, with fluid flowing from one of the reservoirs in "Pulse 1" and fluid flowing from the other of the reservoirs in "Pulse 2", and the two reservoirs may be separate or connected.

[0455] A second embodiment of chip-on-chip cell shear detachment is shown generally by the drawing designated "#2". In this embodiment, the cell extraction nanotube 202 includes an inner tube 1470 contained within an outer tube 1475. One or more shear flow pumps 1460 operate to pump fluids into the organoid chamber 112 through the outer tube 1475 and out of the organoid chamber 112 through the inner tube 1470 as shown by drawing "#2". It is understood that this is intended as an example only. One or more shear flow pumps 1460 may similarly pump fluids into the organoid chamber 112 through the inner tube 1470 and out of the organoid chamber 112 through the outer tube 1475. In a preferred embodiment of the present invention, the inner tube 1470 and the outer tube 1475 are preferably connected to a common reservoir or separate reservoirs (not shown) from which fluid flows into the outer tube 1475 and to which fluid is pumped back from the inner tube 1470, respectively.

[0456] As shown in drawing "#2", in a preferred embodiment of the present invention, a horizontal shear force is induced by a parallel flow into the organoid chamber 112 through the outer tube 1475 and a flow out of the organoid chamber 112 through the inner tube 1480 (or vice versa). This horizontal shear force is applied to the surface of the organoid 113 and may preferably shear single cells 1400 away from the organoid 113 so that they may then be aspirated out of the organoid chamber 112 through the cell extraction nanotube 202 and preferably may then be analyzed for genomic or epigenetic analysis or other desired analysis.

[0457] A third embodiment of the chip-on-chip cell shear detachment is shown diagrammatically by the drawing designated "#3". In this embodiment, the cell extraction nanotube 202 may preferably include two tubes, a first tube 1480 and a second tube 1485. One or more shear flow pumps 1460 operate to pump fluid into the organoid chamber 112 through the first tube 1480 and to pump fluid out of the organoid chamber 112 through the second tube 1485. As the drawing "#3" shows, in a preferred embodiment of the present invention, horizontal shear forces are induced by a parallel flow into the organoid chamber 112 through the first tube 1480 and a flow out of the organoid chamber 112 through the second tube 1485. This horizontal shear force is applied to the surface of the organoid 113, and may preferably shear single cells 1400 away from the organoid 113 so that they may then be aspirated from the organoid chamber 112 through the cell extraction nanotube 202, and preferably may then be analyzed for genomic or epigenetic analysis or other desired analysis. In a preferred embodiment of the present invention, the first tube 1480 and the second tube 1485 are preferably connected to a common reservoir or separate respective reservoirs (not shown) from which fluid flows into the first tube 1480 and to which fluid is pumped back from the second tube 1485, respectively.

[0458] In all three embodiments above, it will be understood that the one or more shear flow pumps 1460 may include different types of devices or methods for actuating fluid flow in a microfluidic channel, including but not limited to a friction wheel (as shown), various piezoelectric pumps known in the art, or electroosmotic driven fluid flow as is known in the art.

[0459] It is further understood that each of the above embodiments may be operated by a single shear flow pump 1460 or by multiple shear flow pumps 1460. As an example, in embodiment "#1", a single shear flow pump 1460 may be used, in which case the direction is reversed to achieve inflow vs. outflow, or two separate pumps may be used for this purpose (in which case the pumps do not need to change their flow direction). In embodiments "#2" and "#3", a single friction wheel pump type shear flow pump 1460 may be used to operate the inflow and outflow in each of the inner and outer tubes of embodiment "#2" and each of the first and second tubes of embodiment "#3". For embodiments "#1", "#2" and "#3" of the present invention in which the shear flow pump 1460 utilizes electro-osmosis, different electric fields may preferably be applied to the outer and inner tubes (of "#2") or to the first and second tubes (of "#3").

[0460] Reference is now made to FIG. 15 which illustrates generally a flow chart of an overview of a drug discovery process utilizing the chip-on-chip system 100 .

[0461] In a preferred embodiment of the present invention, the chip-on-chip system 100 is used as the centerpiece of a drug discovery process as shown in FIG.

[0462] The method includes a step 1500 of predicting a biological effect of a formulation 1800. This prediction may be an artificial intelligence (AI) process, in which formulations from a large pool of possible formulations are evaluated by an AI process, preferably a deep learning process. The biological effect may be various biological effects, desired suspected biological activities of the formulation (e.g., its likelihood to kill a certain type of cancer cell), or undesired biological activities (e.g., its likelihood to have a certain undesired side effect in a certain tissue). The predicted biological effect of the formulation preferably outputs a likelihood score for every formulation in the "pool" of evaluated formulations, for each desired or undesired "biological effect" of interest.

[0463] The method then includes step 1510 of biologically screening a subset of formulations, evaluating the scores of all formulations generated during step 1500, and selecting the formulation with the highest score. The selected formulations are then screened on the chip-on-chip system 100 with the aim of validating these predictions.

[0464] The method then includes a step 1520 of collecting real-time genomic nanosensing. In a preferred embodiment of the present invention, nanowire sensors are used to achieve non-amplified detection of specific nucleic acid molecules of interest that are present in the fluid of the organoid chamber 112 and are therefore associated with the organoid 113. Of particular interest is the detection of cell-free nucleic acid molecules, particularly circular RNA (circRNA) molecules, that are not degraded outside the cell and are associated with disease. As an example, certain circRNAs are associated with autism, and early evidence indicates the possible effectiveness of cannabis in treating autism, and therefore the use of nanowire non-amplified detection of circRNAs in high-throughput multi-organ chip systems may be useful in developing novel drugs for autism. It is understood that non-amplified genomic sensing and specific non-amplified sensing of circRNAs in human-on-chip systems have not been performed before. It is further understood that real-time nanosensing, in which sensors are not directly associated with the organoid chamber 112 in the human-on-chip system, but rather are "shared" by all organoid chambers 112 and delivered by chip-on-chip architecture, has also not been performed before. Real-time genomic nanosensing 1820 is defined hereinabove.

[0465] The method then includes step 1530 of pumping a flexible circulation, for example using actively pumped and fully controlled microfluidic flow, thus flexible in its routing. This circulation allows for both short-circuiting and on-chip organ scaling. As an example, it provides a microfluidic flow that passes from the intestine chamber to the liver chamber, bypassing all other organs, thereby mimicking the "liver first-pass circulation". It is understood that this capability is unique and different from existing systems that rely on gravity flow and have fixed flow routing.

[0466] The method then includes a step 1540 of refining the prediction based on results from the multi-human chip 150 and using the deep matching engine 170. This prediction may then be fed back to step 1500.

[0467] It is understood that these steps are performed until sufficient validation is achieved to obtain a drug, as shown in step 1550.

[0468] 16A-16E show a plurality of organoid chambers with a circulation tunnel connecting the plurality of organoid chambers according to an exemplary embodiment of the subject matter. The plurality of organoid chambers are disposed within a multi-human chip 110. Some of the organoid chambers within the multi-human chip 110 may be connected via the circulation tunnel, while other organoid chambers are not connected to any other organoid chambers.

[0469] Figure 16A shows an array of human chips. The array is bounded by a frame 1600. The array includes multiple groups 1610, 1615, 1618 of interconnected organoid chambers. Each group is separated such that the chambers in the group are not connected to the chambers in the other groups. The number of chambers in each group may be the same or different. The number of chambers may be determined according to the organs represented by the tissues in the chambers of each group.

[0470] 16B shows an isometric view of a group of interconnected organoid chambers 1622, 1625, 1628. Each of the organoid chambers 1622, 1625, 1628 contains biological material, such as cells. The shape of the organoid chambers 1622, 1625, 1628 allows a sensor to pass through (reach) the organoid chambers 1622, 1625, 1628 to extract a sample. The organoid chambers 1622, 1625, 1628 may include barriers configured to prevent leakage from the group into another group of organoid chambers. 1620 The barrier may be mounted on a surface 1660. The surface 1660 may be part of a body carrying the organoid chambers 1622, 1625, 1628.

[0471] FIG. 16C shows a side view of a group of interconnected organoid chambers 1622, 1625, 1628. The organoid chambers 1622, 1625, 1628 are connected via a main chamber 1640. For example, contents can flow from the organoid chamber 1622 to the organoid chamber 1625 only via the main chamber 1640. The main chamber 1640 may be expandable. For example, when the main chamber 1640 expands, the main chamber 1640 presses against the organoid chambers 1622, 1625, 1628, drawing contents out of the organoid chambers 1622, 1625, 1628. The contents are then flowed into the organoid chambers 1622, 1625, 1628, creating an internal circulation between the organoid chambers 1622, 1625, 1628, regardless of the sensor chip 120. The main chamber 1640 may be connected to an air tube 1630 for inflating and deflating the main chamber 1640 .

[0472] 16D shows a top view of a group of interconnected organoid chambers 1622, 1625, 1628. This top view also shows a main chamber 1640. This top view also shows circulation tunnels 1642, 1645, 1648 that connect the organoid chambers 1622, 1625, 1628 to the main chamber 1640.

[0473] 16E shows a bottom view of the group of interconnected organoid chambers 1622, 1625, 1628. This bottom view also shows the air tube 1630 and the main chamber 1640, as the air tube 1630 is located below the surface 1660.

[0474] Those skilled in the art will understand that the present invention is not limited by what has been particularly shown and described above, but rather the scope of the present invention includes both combinations and subcombinations of the various features described hereinabove, as well as variations and modifications that will occur to those skilled in the art upon reading this specification, but which are not in the prior art.

Claims

1. 1. A method for assessing a characteristic of a drug, comprising: treating the tissue sample with said drug; extracting from the tissue sample at least one characteristic of the processed tissue sample; providing said at least one feature in an engine; obtaining a prediction from the engine; associating said prediction with said drug; The method includes:

2. 2. The method of claim 1, wherein the at least one feature comprises a plurality of features, and each feature from the plurality of features is extracted from the tissue sample at a predetermined time after treating the tissue with the drug.

3. The method of claim 1 , wherein the at least one characteristic is obtained by applying a chemical or biological test to the tissue.

4. The method of claim 1 , wherein the at least one feature is obtained by capturing an image of the tissue and extracting features from the image using image analysis techniques.

5. The method of claim 1 , wherein the engine is an artificial intelligence (AI) engine.

6. The method of claim 5 , wherein the AI ​​engine is a convolutional neural network, a deep neural network, or a random forest engine.

7. The method of claim 1 , wherein the prediction relates to a toxicity level of the drug.

8. Further comprising generating the engine, the generating step comprising: treating a plurality of tissue samples with a plurality of drugs, each drug from said plurality of drugs being associated with a property; extracting at least one feature from at least some of said plurality of tissue samples; providing the at least one feature from at least some of the plurality of tissue samples and the property associated with the drug that treated the tissue samples to an artificial intelligence (AI) training module; and Obtaining a trained AI engine 2. The method of claim 1, comprising:

9. Further comprising generating the engine, the generating step comprising: treating a plurality of tissue samples with a plurality of drugs, each drug from said plurality of drugs being associated with a property; extracting at least one feature from at least some of said plurality of tissue samples; providing the at least one feature from at least some of the plurality of tissue samples and the property associated with the drug that treated the tissue samples to an artificial intelligence (AI) training module; and Obtaining a trained AI engine 2. The method of claim 1, comprising:

10. A human chip device comprising: A first biological chip including a plurality of humanoids, each of the plurality of humanoids including a plurality of organoid chambers, each of the organoid chambers containing at least one organoid; A second biological chip, a plurality of sensors operative to collect measurements from the plurality of organoid chambers of the first chip; and a circulation channel fluidly connecting at least a portion of said plurality of organoid chambers; a second biological chip comprising: an actuator operative to move the second tip relative to the first tip; The device that contains

11. The device of claim 10 , further comprising a controller for controlling operation of the actuator and a memory containing a set of rules for moving the second biological chip relative to the first biological chip.

12. 11. The device of claim 10, further comprising a pump operable to pump a circulating fluid into at least one organoid chamber of the plurality of organoid chambers.

13. The device of claim 10 , further comprising an administration port for delivering fluid to the plurality of chambers via the circulation channel.

14. The device of claim 10, wherein at least a portion of the organoid chamber further comprises a secondary circulation channel and a secondary pump.

15. 15. The device of claim 14, further comprising an exhaust chamber connected to the secondary circulation channel for storing fluid extracted from the corresponding organoid chamber.

16. The device of claim 10, wherein at least a portion of the organoid chamber is connected to a chamber valve, the chamber valve enabling or disabling the flow of contents from the circulation channel to the organoid chamber.

17. The device of claim 10 , wherein at least one of the first biological chip and the second biological chip is microfluidic.

18. The device of claim 1, wherein the second biological chip further comprises an extraction device for extracting a single cell from an organoid located in one of the plurality of organoid chambers.

19. The second biological chip further comprises an analyzer operable to receive at least one cell before applying a treatment to the organoid and at least one cell before applying a treatment to the organoid; 10. The device of claim 9, wherein the analyzer is further operative to determine a correlation between a phenotype and the treatment, the phenotype being selected from the group consisting of genomic data and epigenetic data.

20. 20. The device of claim 19, wherein at least some of the plurality of sensors comprise nanotubes located at the bottom of the second biological chip, the nanotubes operable to extract fluid from or deliver fluid to the organoid chamber.

21. the plurality of humanoids representing a respective plurality of patients; the plurality of organoid chambers containing a plurality of organoids representing a plurality of tissues of the plurality of patients; The device further comprises an analyzer operative to predict efficacy of the drug in the plurality of patients. The device of claim 10.

22. 11. The device of claim 10, further comprising an organoid lid located on top of the plurality of organoid chambers, the organoid lid having an open position in which the second biological chip can access contents inside the organoid chambers and a closed position in which the second biological chip can access contents inside the organoid chambers.

23. The device of claim 10 , wherein at least one of the first biological chip and the second biological chip is microfluidic.

24. Further comprising an analyzer operative to receive at least one measurement of the cells before applying a treatment to the organoids and at least one measurement of the cells before applying a treatment to the organoids; 11. The device of claim 10, wherein the analyzer is further operative to determine a correlation between a phenotype and the treatment, the phenotype being selected from the group consisting of genomic data and epigenetic data.

25. The device of claim 10 further comprising a plurality of different second biological chips having different numbers of probes.

26. The device of claim 10 , wherein the second biological chip further comprises a transmitter for transmitting sensed data to a remote device.

27. The device of claim 10, wherein the second biological chip further comprises a plurality of probes, each probe of the plurality of probes configured to be inserted into an organoid chamber of the plurality of organoid chambers.

28. 30. The device of claim 27, wherein at least one probe of the plurality of probes has a nanosensor at a base of the probe.

29. 30. The device of claim 27, wherein at least one probe of the plurality of probes has a nanosensor on a top portion of the probe.

Images