JP2018146308A - Detection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate the need for position control of cell species to be fixed as a sensor.SOLUTION: A detection device of the present invention comprises: a detection unit having an arrangement area in which multiple kinds of cells for each of which a specific receptor is expressed and a signal processing unit for outputting a detection signal obtained from a cell when a molecule is captured in a specific receptor in the order in which the cells are arranged in the arrangement area; an introduction unit for exercising control so that multiple kinds of reference molecules including the multiple kinds of cells are introduced into a sensor array in which the multiple kinds of cells are arranged in the arrangement area; a storage unit for storing, for each cell, a first detection signal string obtained from the multiple kinds of cells when, with regard to each reference module, the reference molecule is captured in the specific receptor; and a learning unit for identifying the arrangement positions in the sensor array and kinds of the cells on the basis of the first detection signal string and a second detection signal string outputted for each reference molecule from the detection unit as a result of each of the multiple kinds of reference molecules is introduced into the sensor array by the introduction unit and corresponding to each arrangement position in the arrangement area.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a detection device that detects a detection target.

  Biomimetic sensing systems can be applied to olfactory sensors and taste sensors, and are expected to be applied to prosthetic organs, robot detection systems, sensor networks, and the like. The sensing system refers to a system that combines a sensor group that detects molecules and an information processing system that processes and outputs the sensor group so that the output of the sensor group can be used for a desired function. One of the characteristics of the odor sensing method of a living body is that a plurality of sensor elements having sensitivity to a wide range of molecular species are prepared, so that a very wide range of molecular species can be comprehensively detected with a small number of sensor types. Is a point. The detected information is converted into desired data and utilized. Although the method of data conversion is complex, it can be flexibly adapted to the environment through later learning. Due to these characteristics, it is possible to detect the target molecule by the later learning, not the sensing configuration capable of detecting only the target molecule determined innately.

  As a method of obtaining identification information from outputs obtained from a plurality of sensors, for example, there is Patent Document 1. The pattern recognition device of Patent Document 1 includes a conversion unit that converts an input pattern to be identified into time-series data having a predetermined bit length, and the time-series data of the input pattern and the time-series data of a predetermined bit length of the reference pattern. A comparison means for performing a comparison operation and outputting an identification result signal for the input pattern, time series data of the input pattern based on the identification result signal, and time series data of a reference pattern that most closely approximates the input pattern; And learning means for updating the time-series data of the reference pattern by performing an operation between them.

  Moreover, the existing semiconductor planar process like patent document 2 is applicable to a sensor device, for example. A semiconductor memory device disclosed in Patent Document 2 includes a memory cell that includes a transistor formed on a semiconductor substrate and a variable resistance element whose resistance value is changed by voltage application connected between the source and drain terminals of the transistor. A memory cell array is three-dimensionally arranged in a plurality of directions and in an array. The semiconductor memory device of Patent Document 2 is characterized by a memory cell structure having a double channel structure in which the inside of a switching transistor is filled with a variable resistance element, particularly a phase change material. In the semiconductor memory device of Patent Document 2, a voltage is applied, the switching transistor is turned off to increase the channel resistance, and a current is passed to the internal phase change material side to operate the memory. Non-Patent Document 1 discloses BioFET.

JP 2000-268017 A JP 2010-165982 A

Analyst, (2002), 127, 1137-1151: Recent advances in biologically sensitive field-effect transistors (BioFETs)

  When constructing a sensing system using a sensor array, a mapping function for data conversion is implemented in a part that performs analysis for identification (hereinafter referred to as an analysis unit). At the time of implementation, it is necessary to know the input feature, that is, the type of sensor for each pixel. Cell fixation to the sensor array is performed after the sensing system is configured. This is because cells must be fixed in the electrolyte, the electrolyte must be filled after fixation, and the cells are living and it is difficult to construct a sensing system while maintaining a living environment. Derived from. Under these conditions, since the assignment of the sensor type for each pixel is determined in advance when the mapping function is implemented in the analysis unit, it is necessary to accurately fix the cell type to a desired pixel when creating the sensor array. . This is extremely difficult with current technology.

  An object of the present invention is to make it unnecessary to control the position of a cell type fixed as a sensor.

  The detection device according to one aspect of the invention disclosed in the present application includes an arrangement region in which a plurality of types of cells each expressing a specific receptor can be arranged, and when a molecule is captured by the specific receptor, A signal processing unit that outputs detection signals obtained from the cells in the arrangement order of the cells in the arrangement region, and a plurality of types of reference molecules including the plurality of types of cells, the plurality of types of cells. Obtained from the plurality of types of cells when the reference molecule is captured by the specific receptor for each reference molecule, and an introduction part that controls the introduction to a sensor array arranged in the arrangement region. The first detection signal sequence stored for each cell, the first detection signal sequence stored in the storage unit, and the introduction unit introduces the plurality of types of reference molecules onto the sensor array. The As a result, the arrangement position and type of the cell on the sensor array are identified based on the second detection signal sequence corresponding to each arrangement position of the arrangement region output for each reference molecule from the detection unit. And a learning unit.

  According to a typical embodiment of the present invention, position control of a cell type to be fixed as a sensor can be made unnecessary. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments.

FIG. 1 is an explanatory diagram illustrating a configuration example of the detection apparatus according to the present embodiment. FIG. 2 is an explanatory diagram illustrating an analysis task example of the analysis unit. FIG. 3 is an explanatory diagram illustrating a structure example of a sensor device. FIG. 4 is an explanatory diagram illustrating a specific configuration example of the sensor device. FIG. 5 is a side sectional view 1 of the sensor device. FIG. 6 is a side sectional view 2 of the sensor device. FIG. 7 is a block diagram illustrating a detailed configuration example of the detection apparatus. FIG. 8 is an explanatory diagram illustrating an example of learning processing by the learning unit. FIG. 9 is an explanatory diagram illustrating an example of assignment setting processing by the second control unit. FIG. 10 is an explanatory diagram illustrating a specific example of the molecule to be detected by the conversion unit. FIG. 11 is a flowchart illustrating an example of a detection processing procedure performed by the detection device. FIG. 12 is an explanatory diagram illustrating another example of the analysis task of the analysis unit. FIG. 13 is an explanatory diagram illustrating an implementation example of the introduction unit.

  The detection device of the present embodiment does not focus on a single molecule and detects its odor or taste, but can detect various types of molecules for a detection target that does not know what odor or taste it is. Multiple types of sensors are prepared, and the detection results of each sensor are comprehensively determined, what kind of smell or taste the detection target has, that is, what kind of molecules the detection target is composed of Is detected.

<Configuration example of detection device>
FIG. 1 is an explanatory diagram illustrating a configuration example of the detection apparatus according to the present embodiment. (A) is the graph 10 which shows the signal sensitivity characteristic of various molecules, (B) is a block diagram which shows the structural example of the detection apparatus 100 concerning a present Example. The horizontal axis of the graph 10 of (A) is the type of molecule, that is, the molecular weight, and the vertical axis is the signal sensitivity corresponding to the ion concentration change, the ion flow rate, and the like. scA to scZ are sensitivity characteristics of cells A to Z, respectively. The sensitivity characteristics scA to scZ of the cells A to Z are congested with the sensitivity characteristics of other cells having similar molecular weights. The range from the cell A showing the smallest molecular weight of the cells A to Z to the cell Z showing the largest molecular weight is the type of molecules that can be detected by the detection device 100 of (B). This range is called a window. The range of the window, whether natural or artificial, may be the range of all molecules present in nature, or may be narrowed down to a specific molecular weight range.

  In (B), in order to realize odor or taste sensing similar to that of a living body, the detection apparatus 100 molecules a plurality of types of cells A, B, C, D,..., Z each expressing a specific receptor. The sensor array 110 is prepared as a sensor and the molecular sensors are dispersed on the array. Each molecular sensor type, that is, a cell type detects, for example, a molecular type within a window range. The molecular species is given to the pixel array PA as a reference molecule or a molecule to be detected. The pixel array PA is a matrix pixel group in which cells are not arranged, and the pixel array PA in which the cells are fixed to the pixels P (i, j) is the sensor array 110.

  The receptor is a substance that exists on the cell surface or inside the cell and affects the function of the cell by specifically binding to a specific substance outside the cell. For example, olfactory receptors are present in olfactory neurons and bind to the structure of odor molecules. For example, the Drosophila olfactory receptor Or13a binds to the structure of the odor molecule 2,3-butanediol or 1-hexen-3-ol.

As shown in (A), since the sensitivity characteristics of the molecular species in the detectable window range are converging, the sensor array 110 can comprehensively detect the molecular species in the window range. The type of molecule that can be detected depends on the receptor to be expressed. When a cell odor expressing a certain type of receptor is captured by the cell's receptor, ions (Na ⁺ , k ⁺ , Ca ^2+, etc.) are exchanged between inside and outside the cell. Using this phenomenon, the detection apparatus 100 detects changes in ion concentration and the ion flow itself using means such as current, potential, and fluorescence.

  Specifically, using (B), (1) a user prepares a plurality of types of cells A, B, C, D,..., Z each expressing a specific receptor as a molecular sensor, And the sensor array 110 which spread | dispersed the said molecular sensor on the array is comprised on the sensor device 101. FIG. (2) The learning unit 103 outputs a learning instruction to the introduction unit 102. (3) Upon receiving the learning instruction (2), the introduction unit 102 sequentially introduces teacher signals ts1, ts2, ts3,... Serving as reference molecules onto the sensor array 110. Thereby, the reference molecule binds to a receptor expressed in cells on the sensor array 110.

  (4) The sensor device 101 outputs a detection signal from the sensor array 110 to the learning unit 103. The detection signal is a signal such as an electric current, a potential, and fluorescence corresponding to an ion exchange phenomenon caused by binding of the reference molecule to the receptor. As a result, the learning unit 103 identifies the type of cell arranged at each pixel position of the sensor array 110.

  (5) A molecule to be detected is introduced into the sensor array 110. As a result, the molecule to be detected binds to the receptor expressed in the cells on the sensor array 110. (6) The sensor device 101 that is a detection unit outputs a detection signal to the analysis unit 104. This detection signal is a signal such as current, potential, and fluorescence corresponding to an ion exchange phenomenon caused by binding of a molecule to be analyzed to a receptor. (7) The analysis unit analyzes the detection signal using the learning result of the learning unit 103, specifies the cells constituting the molecule to be detected, and outputs the analysis result.

  In this way, the detection apparatus 100 comprehensively determines the detection results from the sensor array 110 for detection target molecules that do not know what kind of odor or taste, and what kind of odor or taste the detection target molecules have. That is, it is possible to detect what kind of molecule the detection target is composed of.

  Further, it is not necessary to control the position and number of cell types (receptor types) fixed to the pixel array PA. Therefore, the number of manufacturing steps of the manufacturing process of the sensor array 110 and the development load of the manufacturing technology can be reduced, and the manufacturing cost can be reduced. Furthermore, it can respond to the replacement of cells, and can contribute to extending the life of the detection apparatus 100.

<Example of analysis task of analysis unit 104>
FIG. 2 is an explanatory diagram illustrating an example of an analysis task of the analysis unit 104. As a premise of the analysis in the analysis unit 104, which cell is arranged at each pixel position of the sensor array 110 is determined, and the analysis unit 104 needs to grasp this. That is, the order of the detection signal sequence of each pixel P (i, j) from the sensor device 101 needs to match the order of the cells grasped by the analysis unit 104. Among the analysis task examples (A) to (C) shown below, (A) and (B) are analysis task examples by the detection device 100 without the learning unit 103, and (C) has the learning unit 103. It is an example of an analysis task by the detection apparatus 100. The example of FIG. 2 shows an example in which three types of molecules (represented by ◯, Δ, +) can be identified. In each of (A) to (C), the total number of ◯, Δ, and + is the number of trials of analysis and the number of introductions of the molecule to be analyzed.

  The mapping function 200 is, for example, an expression that expresses the graph 10 showing the sensitivity characteristics scA to scZ of various molecules shown in FIG. By giving the signal sensitivity for each cell to the mapping function 200, the molecular weight of the molecule is uniquely specified. The mapping function 200 is set prior to placement assuming the cell type placed at each pixel. The number of axes in the solution space corresponds to the number of cells in the window.

  In (A), the pixel P (i, j), which is the position where N cells are to be placed, is grasped by the analysis unit 104, and the N cells are placed at positions where the analysis unit 104 knows. This is an example analysis task. That is, an example of an analysis task with complete position control. In this case, when detection signal sequences from N cells Sensor 1 to N are input from the sensor device 101 to the analysis unit 104, the analysis unit 104 enters the solution space via the mapping function 200 implemented in the analysis unit 104. Map. By setting the optimal mapping function 200, data groups corresponding to a plurality of molecular species can be separated and identified.

  (B) is an example of an analysis task when the analysis unit 104 grasps the pixel position where the N cells are to be arranged, but the N cells are not arranged at the position grasped by the analysis unit 104. is there. In (B), the failure example of the position control in which the cell Sensor3 is disposed at the pixel position of the sensor array 110 where the cell Sensor1 is to be disposed, and the cell Sensor1 is disposed at the pixel position of the sensor array 110 where the cell Sensor3 is to be disposed. It is. In this case, the analysis unit 104 assigns the detection signal from the cell Sensor3 to the variable of the mapping function 200 to which the detection signal from the cell Sensor1 is to be substituted, and the variable to the mapping function 200 to which the detection signal from the cell Sensor3 is to be substituted. The detection signal from the cell Sensor1 is substituted. As a result, a situation occurs in which the data group mapped to the solution space cannot be separated well compared to the case of (A).

  (C) is an example of an analysis task in which the analysis unit 104 dynamically grasps a pixel P (i, j) that is a position where N cells are arranged. That is, the detection apparatus 100 learns which cells are arranged in which pixels P (i, j) by the learning unit 103, and changes the arrangement of the detection signal sequence from the sensor array 110 according to the learning result. Thus, as in (A), identification data in the case of complete position control is obtained. Since it is difficult to arrange cells at predetermined positions, the learning unit 103 learns which cells are arranged at which pixel positions, thereby improving the ease of constructing the sensor array 110 and improving the analysis accuracy. Can be planned.

<Structural example of sensor device 101>
FIG. 3 is an explanatory diagram showing a structural example of the sensor device 101. (A) shows a sensor device 101 having a pixel array PA, and (B) shows an arbitrary pixel P (i, j) in the pixel array PA. i is an index indicating the row direction of the pixel array PA, and j is an index indicating the column direction of the pixel array PA. In FIG. 3, as an example, a pixel array PA of 6 rows and 7 columns is configured.

  The sensor device 101 includes a first substrate 301 and a second substrate 302 stacked on the first substrate 301. The first substrate 301 is a semiconductor or an insulating substrate typified by silicon, glass, or sapphire. The second substrate 302 is an insulating substrate in which a signal processing unit such as an FET (Field Effect Transistor) sensor or a CMOS (Complementary MOS) sensor is incorporated. The second substrate 302 has a pixel array PA, which is a set of a plurality of pixels P (i, j) arranged in a matrix, on the surface (substrate surface) 302a. The pixel array PA is a cell arrangement region. Each pixel is an arrangement position of a cell and has an open hole H (i, j) that is opened. The hole diameter of H (i, j) is adjusted to, for example, several tens of micrometers, and one to several cells are fixed. In FIG. 3, fixed cells are not drawn. In actual use, the sensor device 101 in FIG. 3 is filled with an electrolytic solution, and cells serving as sensors are fixed in a state close to a biological environment.

A cell fixed in the opening hole H (i, j) is a cell in which a specific receptor is expressed in advance. In a cell expressing a certain type of receptor, when the target molecular species is captured by the receptor, ion exchange (Na ⁺ , k ⁺ , Ca ^2+, etc.) occurs between the inside and outside of the cell. Using this phenomenon, the detection apparatus 100 detects ion concentration change and ion flow itself using means such as current, potential, and fluorescence. Therefore, by preparing a plurality of types of cells and spreading them on the sensor device 101, each cell is fixedly arranged in the opening hole H (i, j). The fixed position control accuracy varies greatly depending on the processing method of the surface 302a of the second substrate 302 and the cell spreading method. However, as will be described later, the position control during cell distribution to the sensor device 101 is performed by the learning unit 103. It becomes unnecessary.

  FIG. 4 is an explanatory diagram illustrating a specific configuration example of the sensor device 101. A box-shaped chamber 400 is provided on the second substrate 302. An opening edge 401 a of the opening 401 of the chamber 400 is coupled to the surface 302 a of the second substrate 302 so as to cover the sensor array 110. An excitation light source 407 is provided on the bottom surface 406 of the opening 401 of the chamber 400. The excitation light source 407 irradiates the sensor array 110 with excitation light. The chamber 400 is filled with an electrolyte solution. The first side surface 402 of the chamber 400 and the second side surface 403 opposite to the first side surface 402 are respectively provided with a first opening hole 404 and a second opening hole 405 across the opening edge 401a. The first opening 404 is a micro flow channel through which cells, reference molecules, and detection target molecules to be fixed to the sensor array 110 flow into the chamber 400. The second opening hole 405 is a micro flow path for discharging the solution in the chamber 400. When the cells constituting the sensor array 110 deteriorate, the second opening holes 405 remove and discharge the deteriorated cells from the chamber 400.

  FIG. 5 is a side sectional view 1 of the sensor device 101. (A) shows a side cross section for three adjacent pixels P (i−1, j), P (i, j), and P (i + 1, j) for the sake of simplicity. (B) is an equivalent circuit of the FET sensor 503 of each pixel P (i−1, j), P (i, j), P (i + 1, j). In (A), the electrolyte solution 500 is filled on the second substrate 302 on which the sensor array 110 is configured. Cells A to C are fixed in the opening hole H (i, j) of each pixel P (i, j). Receptors RA to RC are expressed on each surface of cells A to C. In the second substrate 302, around the opening hole H (i, j) of each pixel P (i, j), there are a wiring layer 501, a semiconductor layer 502, and a pair of field effect transistors (FETs) 503a and 503b. FET sensor 503 comprised by these is comprised. In FIG. 5, the FET sensor 503 is given as an example, but a CMOS sensor or a CCD sensor may be used as long as it can detect current, potential, and fluorescence.

  Each FET sensor 503 reads changes in current, potential, and fluorescence around the cells A to C. For example, in a cell that expresses a receptor, when a target molecular species is captured by the receptor of the cell, exchange of ions (Na +, k +, Ca2 +, etc.) occurs inside and outside the cell. When the excitation light EL is irradiated from the excitation light source 407 above the sensor device 101110 toward the sensor array 110, the fluorescence FL is induced by exciting the ions, which are fluorescent substances in the cells A to C, and the source S , A current between the source S and the drain D flows in the channel between the drain D. The current between the source S and drain D is read from the gate electrode G. Thereby, the FET sensor 503 detects this change in fluorescence.

  When the sensor device 101 is formed so that the FET sensor 503 is parallel to the surface 302a, that is, the direction d in which the current between the source S and drain D flows is parallel to the surface 302a, the excitation light EL is directly applied to the FET sensor. Incident at 503. Therefore, the FET sensor 503 detects the excitation light EL that should not be detected originally. On the other hand, in the configuration of FIG. 5, in this embodiment, the direction d in which the current between the source S and drain D of the FETs 503a and 503b flows is perpendicular to the surface 302a of the second substrate 302. The direction d is the incident direction of the excitation light EL. When such a device configuration is used, there is an advantage especially when reading a fluorescence change.

  Specifically, the excitation light EL is almost absorbed by the semiconductor layer 502 on the surface 302a side of the FET sensor 503 that is not related to the FET sensor 503. For this reason, the risk that the FET sensor 503 detects the excitation light EL is significantly reduced. That is, the FET sensor 503 can exclusively detect the fluorescence FL. Further, for example, if the gate electrode G of the FET sensor 503 in the pixel P (i, j) is a metal layer that is opaque to the excitation light EL and the fluorescence FL, the neighboring pixels P (i−1, j), P The influence of the fluorescence FL leaking from (i + 1, j) can also be reduced, and between pixels P (i, j) and P (i-1, j) and pixels P (i, j), P (i + 1, j). The crosstalk between them can be suppressed.

  In the configuration illustrated in FIG. 5, the configuration in which the excitation light EL is irradiated from above the sensor array 110 has been described. However, the excitation light EL from the excitation light source 407 disposed below the sensor array 110, that is, on the first substrate 301 side. It is good also as a structure which irradiates.

  FIG. 6 is a side sectional view 2 of the sensor device 101. (A) shows a side cross section for three pixels P (i−1, j), P (i, j), and P (i + 1, j) for the sake of simplicity. (B) is an equivalent circuit of the FET sensor 503 (one side only) of each pixel. FIG. 6 shows a sensor device 101 having an FET sensor 503 in which two FETs are configured in series in the direction d of the current between the source S and drain D perpendicular to the surface 302a. Since the FET has a two-stage configuration, the excitation light EL1 is incident on the upper FET 503a1 on the surface 302a side, but is absorbed by the FET 503a1, so that the excitation light EL2 transmitted through the FET 503a1 is weaker than the excitation light EL1. It is done. Therefore, the lower FET 503a2 detects the fluorescence FL and is difficult to detect the excitation light EL2. Thereby, the detection accuracy of fluorescence FL can be improved.

  The gate electrodes G1 and G2 can be controlled independently. For example, the influence of the excitation light EL1 can be eliminated by opening and controlling the gate electrode G1 of the upper FET 503a1 on the incident side of the excitation light EL1 to the on state. Specifically, for example, in (B), since the excitation light EL1 incident from above is absorbed in the second substrate 302, its intensity decreases exponentially.

  Since the upper FET 503a1 is the closest FET to the excitation light source 407, the intensity of the incident excitation light EL1 is strong, and it is difficult to detect the fluorescence FL that is originally desired to be detected. In the lower FET 503a2, since the intensity of the excitation light EL2 is sufficiently weakened, it is possible to detect the fluorescence FL originally desired to be detected. In such a case, the gate electrode G1 is opened to turn on the FET 503a1, and the gate electrode G2 is closed to turn off the FET 503a2. The resistance value between the source S1 and the drain D1 of the FET 503a1 is a value determined by the gate electrode G1, and is several orders of magnitude lower than the resistance value between the source S2 and the drain D2 of the FET 503a2. The resistance value between the source S2 and the drain D2 of the FET 503a2 is a value corresponding to the incident light intensity, and in this case, a value corresponding to the fluorescence intensity due to the photoelectric effect.

  The sensor current SI output from the drain D2 of the FET 503a2 is a value corresponding to the series resistance value of the FETs 503a1 and 503a2, but since the resistance value of the FET 503a2 is several orders of magnitude higher, a value corresponding to the fluorescence is output. In order to sufficiently suppress the influence of the excitation light EL1, the number of FETs may be increased to three or more. However, since the series resistance value is increased and the driving voltage of the FET sensor 503 is increased, optimization is performed. What is necessary is just to select the number of stages. As the number of stages increases, it becomes easier to detect fluorescence FL from ions on the cell surface close to the FET. When the second substrate 302 is a transparent substrate, the excitation light EL2 can be irradiated from the lower surface bonded to the first substrate 301. In that case, the above description may be reversed.

  The FET shown in FIGS. 5 and 6 has a structure equivalent to the BioFET described in Non-Patent Document 1. Therefore, by mounting the reference electrode on the BioFET, it can function as a sensor that detects a change in the potential of the electrolyte accompanying ion exchange of cells. In this case, since the reaction of the cell can be confirmed by both the fluorescence and the potential change, it is possible to generate a reliable sensor output. The sensor device 101 having the structure shown in FIGS. 5 and 6 can be realized by applying an existing semiconductor planar process.

<Cell fixation method>
In the sensor array 110 used in the detection apparatus 100, a group of cells each expressing a specific receptor is spread on the surface 302a in the opening hole H (i, j) opened for each pixel P (i, j). A matrix sensor group fixed to the opening hole H (i, j). As a method of spreading the cell group, there is a method of strictly controlling the fixing position and the number of cells using a dispenser. However, if the number of pixels and the types of cells to be fixed increase, the process becomes complicated accordingly. In a simple and efficient spraying method, a plurality of types of cells are seeded in an electrolyte solution, dropped onto the surface 302a, and after waiting for the cells to settle into the open holes H (i, j), the open holes H (i , J) is a method of washing away excess cells distributed on the surface 302a.

  The problem with this spraying method is that the type of cell and the position of the cell are not controlled. In some cases, the pixel P (i, j) in which the cell is not fixed in the opening hole H (i, j) or two A pixel P (i, j) to which the above cells are fixed is generated. However, the problem is solved by learning the type and position of the cells fixedly arranged on the sensor array 110 by the learning unit 103 described above. In other words, by providing the learning unit 103, it is not necessary to control the cell type and the cell position, and the cells can be dispersed easily and efficiently.

<Detailed Configuration Example of Detection Device 100>
FIG. 7 is a block diagram illustrating a detailed configuration example of the detection apparatus 100. The detection apparatus 100 includes a sensor device 101, an introduction unit 102, a learning unit 103, an analysis unit 104, a timing generator 711, and a switch 712. The sensor device 101 includes a pixel array PA, a peripheral circuit 701, and a control circuit 702. The pixel array PA is a set of matrix-like pixels P (i, j), and a group of cells is dispersed on the surface 302a, and the cells are fixed in the opening holes H (i, j). Become.

  The peripheral circuit 701 reads a detection signal sequence from the sensor array 110. Specifically, for example, pixels P (1,1) to P (1, n), P (2,1) to P (2, n),..., P (m, 1) to P (m, n) The control circuit 702 reads out the detection signal sequence from the peripheral circuit 701 according to an instruction from the timing generator 711, and compares the read detection signal sequence with the comparison unit or conversion unit to which the switch 712 is connected. 743.

  The introduction unit 102 includes a first memory 721, a first control unit 722, a valve / flow rate control unit 723, a reference molecular source MS, and a valve group Vs. The first memory 721 stores teacher signal information. For each reference molecule, the teacher signal information is the signal sensitivity that defines the opening and closing information of the electromagnetic control valve corresponding to the reference molecule over time and the signal sensitivity when each reference molecule binds to the receptor of various cells in the window. Information. The time-dependent opening / closing information is information indicating that the electromagnetic control valves of other reference molecules are closed during a period in which the electromagnetic control valve of a certain reference molecule is opened. The opening period and the closing period are set in advance. The signal sensitivity information is, for example, a table showing the signal sensitivity in the combination of the reference molecule and the cell as shown in FIG.

  The first control unit 722 reads the opening / closing information from the first memory 721 and outputs it to the valve / flow rate control unit 723 as an instruction to introduce a teacher signal (reference molecule). Thereby, a desired teacher signal (reference molecule) at each timing or a teacher signal given a time-series change as necessary can be introduced into the sensor device 101, and its reaction can be detected. The first control unit 722 reads the signal sensitivity from the first memory 721 and outputs the signal sensitivity to the comparison unit. In addition, when adding, changing, or deleting a reference molecule to the reference molecule source MS, the first control unit 722 writes or deletes corresponding teacher signal information according to an instruction from the outside.

  In addition, the first control unit 722 controls the timing generator 711 to instruct generation of a timing signal. Specifically, for example, when the reference molecule is introduced as a teacher signal, the first control unit 722 instructs the timing generator 711 to generate a timing signal. The first control unit 722 controls ON / OFF of the gate electrode of the FET included in the FET sensor 503.

  The valve / flow rate control unit 723 specifies the electromagnetic control valve to be opened and the electromagnetic control valve to be closed from the valve group Vs based on the opening / closing information from the first control unit 722. The valve / flow rate control unit 723 controls opening / closing of the identified electromagnetic control valve for the opening / closing time included in the opening / closing information.

  The reference molecule source MS stores a plurality of types of reference molecules. The valve group Vs is a set of electromagnetic control valves corresponding to the reference molecules of the reference molecule source MS. Each electromagnetic control valve of the valve group Vs controls the opening and closing of the piping between them in order to introduce each reference molecule of the reference molecule source MS into the sensor device 101. The piping from the valve is connected to the first opening hole 404 of the sensor device 101.

  The timing generator 711 generates a timing signal according to an instruction from the first control unit 722 and outputs the timing signal to the control circuit 702 of the sensor device 101. Accordingly, the control circuit 702 can control the peripheral circuit 701 to read the detection signal sequence from the signal processing unit of the sensor device 101 and output the detection signal string to the comparison unit or conversion unit 743 that is the connection destination via the switch 712. .

  The switch 712 selects the output destination of the detection signal sequence from the sensor device 101, switches the signal path, and outputs it. Specifically, for example, when the reference molecule is introduced into the sensor device 101 as a teacher signal by the first control unit 722, the switch 712 selects the comparison unit as an output destination under the control of the first control unit 722. To do. On the other hand, when the detection target molecule is introduced into the sensor device 101, for example, when it is not controlled by the first control unit 722, the switch 712 selects the conversion unit 743 as an output destination. Thus, the detection signal sequence when the reference molecule is introduced into the sensor device 101 is output to the comparison unit, and the detection signal sequence when the detection target molecule is introduced into the sensor device 101 is output to the conversion unit 743. The

  The learning unit 103 detects signal sensitivity information, which is a first detection signal sequence for each cell corresponding to a plurality of types of reference molecules, and a plurality of types of reference molecules introduced onto the sensor array 110 by the introduction unit 102, as a result Based on the second detection signal sequence corresponding to each pixel of the pixel array output from the device 101 for each reference molecule, the pixel of the cell arranged on the sensor array 110 and the type of the cell are identified.

  Specifically, for example, the learning unit 103 includes a comparison unit 730, a second control unit 741, and a second memory 742. The second control unit 741 and the second memory 742 are shared with the analysis unit 104. The comparison unit 730 has a temporary memory 731. The comparison unit 730 stores the signal sensitivity information from the first control unit 722 in the temporary memory 731. The comparison unit 730 stores the second detection signal string input from the sensor device 101 via the switch 712 in the temporary memory 731. The second detection signal sequence is stored for each reference molecule. The comparison unit 730 identifies the arrangement position of the cell on the sensor array 110 and the type of the cell based on the second detection signal sequence for each reference molecule and the signal sensitivity information. The arrangement position and type of the identified cell are the learning result. The learning unit 103 outputs the learning result of the comparison unit 730 to the second control unit 741.

  The second control unit 741 stores the learning result from the comparison unit 730 in the second memory 742. The second memory 742 stores the learning result written by the second control unit 741. Here, the learning process by the learning unit 103 will be specifically exemplified.

  FIG. 8 is an explanatory diagram illustrating an example of learning processing by the learning unit 103. Here, in order to simplify the description, the sensor array 110 of 2 rows and 3 columns is used, the cells in the window of FIG. 1A are referred to as cells A to F, and the reference molecules are referred to as reference molecules a to d. Further, the value of the signal sensitivity is an example for facilitating the explanation. Further, as a default, the cells A to F correspond to the pixels P (1, 1) to P (2, 3).

  Signal sensitivity information 801 is stored in the first memory 721, and is stored in the temporary memory 731 by the first control unit 722. The signal sensitivity information 801 has a known signal sensitivity that has been detected when the reference molecule is bound to a receptor expressed in each of the cells AF. For example, the signal sensitivity detected when the reference molecule a is bound to the receptor expressed in the cell A is “8”.

  The second detection signal sequence for each reference molecule output from the sensor device 101 is stored in the first memory 721 as a data set 802. The comparison unit 730 compares the vertical column (that is, the column of cells) of the signal sensitivity information 801 with the vertical column of the data set 802 (that is, the column of pixels P (i, j)). If they match, the cell is placed at the pixel P (i, j). For example, the cell A column of the signal sensitivity information 801 is (8, 4, 0, 0), which matches the column of the pixel P (1, 2) of the data set 802. Therefore, it is found that the cell arranged in the pixel P (1,2) is the cell A. Similarly, the cell arranged in the pixel P (1,1) is arranged in the cell F, the cell arranged in the pixel P (1,3) is arranged in the cell C (no change), and the pixel P (2,1). It is found that the cell is the cell B, the cell arranged in the pixel P (2, 2) is the cell E, and the cell arranged in the pixel P (2, 3) is the cell D. In this way, a learning result 803 is obtained.

  Returning to FIG. 7, the analysis unit 104 performs mapping conversion of the third detection signal sequence from the detection unit input according to the pixel in which a plurality of types of cells are arranged on the sensor array 110 using the mapping function 200, thereby detecting the detection target. Analyze molecules.

  Specifically, for example, the analysis unit 104 includes a conversion unit 743, a second control unit 741, and a second memory 742. The conversion unit 743 includes a register 744744. The conversion unit 743 stores the third detection signal string output from the sensor device 101 via the switch 712 in the register 744. The conversion unit 743 stores the learning result output from the second memory 742 via the second control unit 741 in the register 744. The conversion unit 743 changes the order of the third detection signal sequence to be given to the mapping function 200 based on the learning result 803. Thereby, the conversion unit 743 outputs identification data (for example, molecular weight) for specifying the detection target molecule as an analysis result by performing mapping conversion on the third detection signal sequence whose order has been changed by the mapping function 200.

  The conversion unit 743 may display the analysis result on a display device (not illustrated) included in the detection apparatus 100 or may transmit the analysis result to another apparatus connected to the detection apparatus 100. Here, the analysis processing by the analysis unit 104 will be specifically exemplified.

  FIG. 9 is an explanatory diagram illustrating an example of assignment setting processing by the second control unit 741. (A) is correspondence information 901 indicating the correspondence between cells and pixels P (i, j) before and after the learning result, and (B) is a third detection signal sequence (before and after the learning result to the mapping function 200). It is an example of substitution of x1, x2, x3, x4, x5, x6). A to F in the formula (B) are coefficients specific to the cells A to F. Before obtaining the learning result 803, the detection signal x1 corresponds to the pixel P (1,1), the detection signal x2 corresponds to the pixel P (1,2), and the detection signal x3 corresponds to the pixel P (1,3). Correspondingly, the detection signal x4 corresponds to the pixel P (2,1), the detection signal x5 corresponds to the pixel P (2,2), and the detection signal x6 corresponds to the pixel P (2,3). .

  By obtaining the learning result 803, the arrangement positions of the cells A to F are changed as shown in the correspondence information 901. Therefore, after learning, the third detection signal sequence (x1, x2, x3, x4, x5, x6) from the sensor device 101 is (x6, x1, x3, x2, x2) in the register 744 according to the correspondence information 901. x5 and x4) are substituted into the mapping function 200 of the conversion unit 743. As a result, as shown in FIG. 2C, identification data in the case of complete position control is obtained, and it is possible to improve the ease of constructing the sensor array 110 and improve the analysis accuracy.

  FIG. 10 is an explanatory diagram illustrating a specific example of a molecule to be detected by the conversion unit 743. In FIG. 10, the sensitivity characteristic graph 10 is used for convenience. Here, it is assumed that sA, sB, and sC are obtained as detection signals of the third detection signal sequence. The detection signal sA is a detection signal from the cell A, the detection signal sB is a detection signal from the cell B, and the detection signal sC is a detection signal from the cell C. The analysis unit 104 identifies the molecular weight mw corresponding to the detection signals sA, sB, and sC. The molecular weight mw is a value obtained as a result of substituting the detection signals sA, sB, and sC into the mapping function 200 after learning, and corresponds to the identification data in FIG.

<Example of Detection Processing Procedure by Detection Device 100>
FIG. 11 is a flowchart illustrating an exemplary detection processing procedure performed by the detection apparatus 100. The introduction unit 102 sets the variable i to i = 1 (step S1101), and determines whether i exceeds NT (step S1102). NT is the number of reference molecules in the reference molecule source MS. If i> NT is not satisfied (step S1102: NO), the introduction unit 102 causes the first control unit 722 to output an introduction instruction for the i-th teacher signal to the valve / flow rate control unit 723 (step S1103). As a result, the electromagnetic control valve corresponding to the reference molecule that is the i-th teacher signal is opened, and the i-th reference molecule is introduced into the sensor array 110. Then, the introduction unit 102 increments the variable i and returns to step S1102.

  On the other hand, if i> NT (step S1102: Yes), the introduction unit 102 stores the second detection signal sequence from the sensor array 110 in the temporary memory 731 of the comparison unit 730 (step S1105). The introduction unit 102 stores the signal sensitivity information 801 in the temporary memory 731 of the comparison unit 730 by the first control unit 722, and the learning unit 103 stores the second detection signal stored in the temporary memory 731 by the comparison unit 730. By comparing the column with the signal sensitivity information 801, the cell type for each pixel P (i, j) is identified (step S1106) and stored in the second memory 742 as the learning result 803 as shown in FIG. To do.

  Then, as illustrated in FIG. 9, the second control unit 741 sets assignment of the third detection signal sequence output from the sensor device 101 to the mapping function 200 using the learning result 803 (step S1107). ). Then, the analysis unit 104 outputs the analysis result by changing the order of the third detection signal sequence according to the assignment setting (step S1107) and substituting it into the mapping function 200 by the conversion unit 743 (step S1108). . As a result, a series of processing ends.

<Machine Learning Example of Learning Unit 103 and Analysis Unit 104>
According to recent AI (Artificial Intelligence) technology, it is possible to learn a relationship between a plurality of inputs and a desired output by machine learning and to formulate a mapping method. If machine learning is introduced into the second control unit 741, the mapping function 200 can be formulated and implemented after the system configuration without preparing the mapping function 200 in the conversion unit 743 in advance.

  A machine learning example of the learning unit 103 will be described. For example, an appropriate mapping function 200 is prepared in advance, and the second control unit 741 gives the mapping function 200 a second detection signal sequence when the reference molecule a is introduced into the sensor array 110, for example. First identification data is output. Next, the coefficient of the mapping function 200 corresponding to the pixel of the cell in which the receptor to which the reference molecule a binds is expressed is changed, and the second detection signal sequence is given to the mapping function 200 after the coefficient change, so that the second Output identification data. If the second identification data is closer to the molecular weight of the reference molecule a than the first identification data, the coefficient change is correct.

  The learning unit 103 can formulate the optimum mapping function 200 by the second control unit 741 by repeatedly executing such processing for each of the reference molecules a to d. The second control unit 741 stores the formulated mapping function 200 in the register 744 of the conversion unit 743.

  In addition, when cells are spread on the pixel array PA, other cells are arranged on the pixel P (i, j) where specific cells are to be arranged, or no cells are arranged, or a plurality of types of cells are arranged. May be placed.

  FIG. 12 is an explanatory diagram illustrating another example of the analysis task of the analysis unit 104. In FIG. 12, the pixel P (1,1) that should output Sensor1 has output Sensor3, and the pixel P (1,3) that should output Sensor3 outputs Sensor1 and Sensor2. This is an example. That is, not the cell that outputs Sensor1 but the cell that outputs Sensor3 is fixed to the pixel P (1,1), and the sensor1 is not a cell that outputs Sensor3 to the pixel P (1,3). And cells that output Sensor2 are fixed.

  In such a case, the second control unit 741 stores the combination of the second detection signal sequence and the identification data (molecular weight) as the analysis result in the second memory 742 each time a reference molecule is introduced. Then, the second control unit 741 calculates the unknown coefficient by substituting each of the plurality of combinations into the mapping function 200 of the unknown coefficient. Then, the second control unit 741 updates the original mapping function 200 to the mapping function 200 having the calculated coefficient, and stores it in the register 744.

  Thus, since the mapping function 200 can be formulated after the cells are fixed, it is possible to change a desired output, that is, a function desired to be exhibited by the detection apparatus 100 after the configuration of the detection apparatus 100. This indicates that, for example, when an odor is taken as an example, the odor species to be identified can be changed as appropriate within the range possible with the prepared cell type.

  The above-described formulation can also be performed in the analysis unit 104. For example, the second control unit 741 stores the combination of the third detection signal string and the identification data (molecular weight) as the analysis result in the second memory 742 each time a detection target molecule is introduced. Then, the second control unit 741 calculates the unknown coefficient by substituting each of the plurality of combinations into the mapping function 200 of the unknown coefficient. Then, the second control unit 741 updates the original mapping function 200 to the mapping function 1200 having the calculated coefficient, and stores it in the register 744. Thus, the mapping function 200 can be optimized even after the mapping function 200 is established.

<Example of implementation of introduction unit 102>
FIG. 13 is an explanatory diagram illustrating an implementation example of the introduction unit 102. The introduction unit 102 includes a supply source 1300, a circulation system 1304, a disposal system 1305, a control unit 1310, and electromagnetic control valves 1311 to 1313. The supply source 1300 is connected to the first opening 404 of the sensor device 101 by piping and an electromagnetic control valve 1311. The supply source 1300 has a first electrolyte solution 1301, a second electrolyte solution 1302, and a surfactant solution 1303, and is a mechanism that supplies any liquid to the sensor device 101 through a pipe. The second electrolyte solution 1302 is an electrolyte solution in which a plurality of types of cells are seeded. The surfactant solution 1303 is a solution containing a surfactant that decomposes a plurality of types of cells arranged in the pixel array PA. The circulation system 1304 is a mechanism that supplies a reference molecule and a detection target molecule into the chamber 400. The discard system 1305 is a mechanism for discarding the liquid in the chamber 400 from the chamber 400.

  The control unit 1310 includes, for example, a first control unit 722 and / or a valve / flow rate control unit 723. The control unit 1310 supplies the first electrolyte solution 1301, the second electrolyte solution 1302, and the surfactant solution 1303 from the supply source 1300 to the chamber 400, the supply of the molecule to be detected from the circulation system 1304 to the chamber 400, And the disposal from the chamber 400 to the disposal system 1305 is controlled. Specifically, for example, the control unit 1310 controls opening / closing of the electromagnetic control valve 1311 so that any one liquid of the first electrolyte solution 1301, the second electrolyte solution 1302, and the surfactant solution 1303 is contained in the chamber 400. And supply of other liquids is stopped.

  In addition, the control unit 1310 controls opening / closing of the electromagnetic control valve 1312 to supply the detection target molecule into the chamber 400 or stop the supply. In addition, the control unit 1310 controls the opening and closing of the electromagnetic control valve 1313 to supply the liquid in the chamber 400 to the waste system 1305 or stop the supply.

  Hereinafter, an example of the cell replacement step will be described. First, the detection apparatus 100 causes the surfactant solution 1303 to flow through the first opening 404 of the chamber 400, introduces it into the sensor array 110, and discards it in the disposal system 1305. As a result, the cells fixed to the sensor array 110 are dissolved and separated, and the surfactant solution 1303 containing the detached cells is discharged into the waste system 1305 in the chamber 400.

  Next, the detection apparatus 100 introduces the first electrolyte solution 1301 into the chamber 400, and purges and cleans the inside of the chamber 400. Although the detection device 100 closes the electromagnetic control valve 1313 on the disposal system 1305 side, it may be opened and filled with the first electrolyte solution 1301 while draining waste.

  Next, the detection apparatus 100 introduces the second electrolyte solution 1302 into the chamber 400. The detection apparatus 100 closes the electromagnetic control valve 1313 of the waste system 1305 to settle and fix the cell group on the pixel array, and then opens the electromagnetic control valve 1313 of the waste system 1305 to remove the second electrolyte solution 1302 from the waste liquid. By draining excess cells.

  Thereafter, the detection apparatus 100 closes the electromagnetic control valve 1311 of the supply source 1300, operates the circulation system 1304, introduces the reference molecule into the chamber 400, and performs sensing. Thereafter, as described above, the detection apparatus 100 performs a learning process by the learning unit 103, identifies a cell fixed for each pixel, and obtains a learning result. Then, the detection apparatus 100 operates the circulatory system 1304, introduces the detection target molecule into the chamber 400, performs sensing, and analyzes the detection target molecule using the learning result.

  Thus, since cells can be replaced in the sensor device 101 in a timely manner, the long-life sensor device 101 with good resistance to deterioration can be provided. Furthermore, when the mapping function 200 described above can be reconfigured, it is possible to change the receptor to be expressed at the time of cell replacement. This makes it possible to change and revise the molecular groups covered by the same detection device 100 in a timely manner, and to revise the molecular groups to be identified and the functions to be realized.

  As described above, the detection apparatus 100 according to the present exemplary embodiment includes the sensor device 101 that is a detection unit, the introduction unit 102, the first memory 721 that is a storage unit, and the learning unit 103. The sensor device 101 includes an arrangement area and a signal processing unit. The arrangement region is a pixel array PA in which a plurality of types of cells each expressing a specific receptor can be arranged. The signal processing unit is, for example, an FET sensor 503, and molecules are captured by the specific receptor. In some cases, detection signals obtained from the cells are output in the order of cell arrangement in the pixel array PA.

  The introduction unit 102 controls to introduce a plurality of types of reference molecules including a plurality of types of cells onto the sensor array 110 in which the plurality of types of cells are arranged in the pixel array PA. For each reference molecule, the storage unit (first memory 721) stores, for each cell, a first detection signal sequence obtained from the cell when the reference molecule is captured by a specific receptor. That is, the first memory 721 stores signal sensitivity information 801.

  The learning unit 103 includes a first detection signal sequence (signal sensitivity information 801) stored in the first memory 721 and a plurality of types of reference molecules introduced into the sensor array 110 by the introduction unit 102, as a result of the sensor device 101. And the second detection signal sequence corresponding to each pixel P (i, j) of the pixel array PA output for each reference molecule from the pixel P (i, j) and the type on the cell sensor array 110 Is identified.

  This eliminates the need to control the position and number of cell types (receptor types) fixed to the pixel array PA. Therefore, the number of manufacturing steps of the manufacturing process of the sensor array 110 and the development load of the manufacturing technology can be reduced, and the manufacturing cost can be reduced.

  In addition, the detection apparatus 100 includes an analysis unit 104. The analysis unit 104 performs mapping conversion of the third detection signal sequence from the sensor device 101 input according to the pixels P (i, j) of the plurality of types of cells on the sensor array 110 using the mapping function 200, thereby detecting the detection target molecule. Analyze. Specifically, when the detection target molecule is introduced onto the sensor array 110, the analysis unit 104 changes the order of the third detection signal sequence based on the identification result by the learning unit 103, and the order is changed. The third detection signal sequence is subjected to mapping conversion by the mapping function 200 to analyze the molecule to be detected.

  As a result, the detection apparatus 100 comprehensively determines the detection results from the sensor array 110 for the detection target molecules that do not know what kind of smell or taste, and what kind of smell or taste the detection target molecules have. That is, it is possible to detect what molecule the detection target is composed of.

  Further, the learning unit 103 obtains the mapping function 200 from the predetermined function based on machine learning using the second detection signal sequence and the output result obtained from the predetermined function to which the second detection signal sequence is given. May be. Thereby, the optimal mapping function 200 can be formulated without preparing the mapping function 200 in advance prior to analysis.

  Further, the analysis unit 104 updates the mapping function 200 based on machine learning using the third detection signal sequence and an analysis result obtained as a result of the third detection signal sequence being given to the mapping function 200. Also good. After the analysis, the mapping function 200 can be optimized.

  The sensor device 101 has a second substrate 302 as an insulating substrate having a pixel array PA. The pixel array PA has a plurality of opening holes H in which a plurality of types of cells formed on the second substrate 302 can be respectively arranged. A set of pixels P (i, j) having (i, j) may be used. Thereby, a cell can be settled to the opening hole H (i, j) of pixel P (i, j) with the dead weight of the cell, and the cell fixation can be facilitated.

  The FET sensor 503 is a field-effect transistor that detects fluorescence from each pixel P (i, j) for each pixel P (i, j) in which cells are arranged, and between the source and drain of the field-effect transistor. The direction of the current may be a direction perpendicular to the surface 302a of the second substrate 302 in which the opening of the opening hole H (i, j) is formed. Thereby, the influence of excitation light EL can be suppressed and the detection accuracy of fluorescence FL from cells can be improved.

  The FET sensor 503 may include a plurality of stages of field effect transistors in the source-drain current direction d. As a result, the lower the FET, the more difficult it is to detect the fluorescence FL and the excitation light EL, thereby improving the detection accuracy of the fluorescence FL.

  Moreover, the detection apparatus 100 has a switch 712 as a switching unit. The switch 712 switches the output destination of the second detection signal sequence from the sensor device 101 to the learning unit 103 and the output destination of the third detection signal sequence from the sensor device 101 to the analysis unit 104. Thereby, the learning of the cell type and the arrangement position and the analysis of the detection target molecule can be executed by one detection device 100. In addition, since the second control unit 741 and the second memory 742 are shared by the learning unit 103 and the analysis unit 104, the number of parts can be reduced and the size of the detection apparatus 100 can be reduced.

  In addition, the detection apparatus 100 includes a chamber 400, a supply source 1300, a circulation system 1304, a disposal system 1305, and a control unit 1310. The chamber 400 is attached to the sensor device 101 and forms a space so as to cover the sensor array 110. The supply source 1300 includes a first electrolyte solution 1301, a second electrolyte solution 1302 in which a plurality of types of cells are seeded, or a surfactant solution 1303 that decomposes a plurality of types of cells arranged in a pixel array in the chamber 400. Supply. The circulation system 1304 supplies the detection target molecule into the chamber 400. The disposal system 1305 discards the liquid in the chamber 400 from the chamber 400. The control unit 1310 supplies the first electrolyte solution 1301, the second electrolyte solution 1302, and the surfactant solution 1303 from the supply source 1300 to the chamber 400, the supply of the molecule to be detected from the circulation system 1304 to the chamber 400, and The disposal from the chamber 400 to the disposal system 1305 is controlled.

  As a result, since cells can be replaced in the sensor device 101 in a timely manner, it is possible to provide the sensor device 101 having a long life and good resistance to deterioration. Furthermore, when the mapping function 200 described above can be reconfigured, it is possible to change the receptor to be expressed at the time of cell replacement. This makes it possible to change and revise the molecular groups covered by the same detection device 100 in a timely manner, and to revise the molecular groups to be identified and the functions to be realized.

  The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. A part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Moreover, you may add the structure of another Example to the structure of a certain Example. Moreover, you may add, delete, or replace another structure about a part of structure of each Example.

  In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, with an integrated circuit, and the processor realizes each function. It may be realized by software by interpreting and executing the program to be executed.

  Information such as programs, tables, and files for realizing each function is recorded on a memory, a hard disk, a storage device such as an SSD (Solid State Drive), or an IC (Integrated Circuit) card, an SD card, a DVD (Digital Versatile Disc). It can be stored on a medium.

  Further, the control lines and the information lines are those that are considered necessary for the explanation, and not all the control lines and the information lines that are necessary for the mounting are shown. In practice, it can be considered that almost all the components are connected to each other.

DESCRIPTION OF SYMBOLS 100 Detection apparatus 101 Sensor device 102 Introduction part 103 Learning part 104 Analysis part 110 Sensor array 200 Mapping function 400 Chamber 503 FET sensor 712 Switch 721 1st memory 722 1st control part 723 Valve / flow rate control part 730 Comparison part 731 Temporary memory 741 Second control unit 742 Second memory 743 Conversion unit 744 Register 801 Signal sensitivity information 802 Data set 803 Learning result 901 Corresponding information 1300 Supply source 1301 First electrolyte solution 1302 Second electrolyte solution 1303 Surfactant solution 1304 Circulating system 1305 Waste system 1310 Control unit

Claims (9)

An arrangement region in which a plurality of types of cells each expressing a specific receptor can be arranged, and a detection signal obtained from the cell when a molecule is captured by the specific receptor is transmitted from the cell in the arrangement region. A detection unit having a signal processing unit for outputting in order of arrangement;
An introduction unit configured to control a plurality of types of reference molecules including the plurality of types of cells so that the plurality of types of cells are introduced onto a sensor array disposed in the arrangement region;
For each reference molecule, a storage unit that stores, for each cell, a first detection signal sequence obtained from the plurality of types of cells when the reference molecule is captured by the specific receptor;
The first detection signal sequence stored in the storage unit, and the arrangement that is output for each reference molecule from the detection unit as a result of the introduction of the plurality of types of reference molecules onto the sensor array by the introduction unit. A learning unit that identifies an arrangement position and a type of the cell on the sensor array based on a second detection signal sequence corresponding to each arrangement position of the region;
A detection apparatus comprising: The detection device according to claim 1,
An analysis unit that analyzes a detection target molecule by performing mapping conversion with a mapping function on a third detection signal sequence from the detection unit that is input according to an arrangement position of the plurality of types of cells on the sensor array;
The analysis unit
When the detection target molecule is introduced onto the sensor array, the order of the third detection signal sequence is changed based on the identification result by the learning unit, and the third detection signal sequence with the changed order is changed. A detection apparatus that analyzes the molecule to be detected by performing mapping conversion with the mapping function. The detection device according to claim 2,
The learning unit
The mapping function is obtained from the predetermined function by machine learning using the second detection signal sequence and an output result obtained from the predetermined function to which the second detection signal sequence is given. apparatus. The detection device according to claim 2,
The analysis unit
The mapping function is updated by machine learning using the third detection signal sequence and an analysis result obtained as a result of applying the third detection signal sequence to the mapping function. The detection device according to claim 1,
The detector is
An insulating substrate having the arrangement region;
The placement area is
2. A detection apparatus comprising: a set of pixels having a plurality of aperture holes in which the plurality of types of cells formed on the insulating substrate can be respectively arranged. The detection device according to claim 5,
The signal processing unit
For each pixel in which the cells are arranged, a field effect transistor that detects fluorescence from each pixel, and the direction of the current between the source and drain of the field effect transistor is such that the opening of the opening hole is formed A detection device characterized by being in a direction perpendicular to a substrate surface of an insulating substrate. The detection device according to claim 6,
The signal processing unit
A detection apparatus comprising a plurality of the field effect transistors in the direction of the source-drain current. The detection device according to claim 2,
A switching unit that switches an output destination of the second detection signal sequence from the detection unit to the learning unit and an output destination of the third detection signal sequence from the detection unit to the analysis unit. Detection device. The detection device according to claim 1,
A chamber forming a space so as to cover the arrangement region;
A supply source for supplying the chamber with a first electrolyte solution, a second electrolyte solution in which the plurality of types of cells are seeded, or a surfactant solution that decomposes the plurality of types of cells arranged in the arrangement region; ,
A circulation system for supplying molecules to be detected into the chamber;
A waste system for discarding the liquid in the chamber from the chamber;
Supply of the first electrolyte solution, the second electrolyte solution, and the surfactant solution from the source to the chamber, supply of the molecule to be detected from the circulation system to the chamber, and the chamber from the chamber A control unit that controls disposal to the disposal system;
A detection apparatus comprising:

Images