Information processing system, information processing method, program, information processing device and computing device

Abstract

The main purpose of the present technology is to provide a technology for automatically presenting a better combination of fluorescent labels. The present technology provides an information processing system, including an information processing device, the information processing device including: a calculation processing unit that calculates an objective function regarding a combination of a biomolecule and a labeled fluorescent substance based on first data listing information regarding a biomolecule that is a labeled target in a sample and second data listing information regarding a labeled fluorescent substance; and an allocation information output unit that outputs information regarding allocation of the labeled fluorescent substance to each biomolecule, the information being generated based on combination information acquired from the combination optimization processing unit based on a coefficient of the objective function.

Description

Technical Field

[0001] The present technology relates to an information processing system, an information processing method, a program, an information processing device, and a computing device. More specifically, the present technology relates to an information processing system that performs a process for assigning a fluorescent body to a biomolecule, an information processing method for the process, a program for causing an information processing device to perform the process, and an information processing device and a computing device included in the information processing system. Background Art

[0002] For example, a particle group such as a cell is marked with a fluorescent dye, and the intensity and / or pattern of fluorescence generated by the fluorescent dye excited by irradiating each particle in the particle group with a laser beam is measured to measure the characteristics of the particles. A typical example of a particle analysis device that performs this measurement is a flow cytometer. For example, from the perspective of an optical system for fluorescence measurement, a flow cytometer can be roughly classified into a filter type and a spectral type.

[0003] The excitation wavelength at which a signal can be obtained at a peak is related to the fluorescent dye. Therefore, in a filter-type flow cytometer of the prior art, particles labeled with a fluorescent dye are irradiated with light having its excitation wavelength, and a fluorescent signal is obtained by using an optical filter corresponding to the peak wavelength range of the fluorescent dye.

[0004] On the other hand, in a spectral flow cytometer, for example, the fluorescence of each cell is collectively acquired as a spectrum. Using a spectral reference of each fluorescent dye, the acquired fluorescence spectrum is subjected to fluorescence separation processing (also called unmixing processing), and a fluorescence signal is acquired.

[0005] As a technology regarding fluorescence separation processing, for example, the following PTL1 discloses a fluorescence intensity correction method including the following process: receiving fluorescence generated from fluorescent dyes that are excited by irradiating microparticles multiply labeled with a plurality of fluorescent dyes having overlapping fluorescence wavelength bands with light by using a larger number of photodetectors arranged with different light receiving wavelength bands than the number of fluorescent dyes, and approximating a measurement spectrum obtained by collecting detection values ​​from each photodetector using a linear sum of individual staining spectra obtained in microparticles individually labeled with each fluorescent dye.

[0006] [Citation List]

[0007] [Patent Document]

[0008] [PTL1]

[0009] JP2011-232259A. Summary of the invention

[0010] [Technical issues]

[0011] The expression of various antigens can be assessed simultaneously by the unmixing process of fluorescence spectrum. However, the performance of unmixing process may deteriorate, depending on the combination of fluorescent markers. On the other hand, for example, for researchers studying antigen expression, the inspection for selecting a better combination of fluorescent markers is often onerous. Therefore, the purpose of this technology is to provide a technology for automatically presenting a better combination of fluorescent markers.

[0012] [Solution to the problem]

[0013] The present technology provides an information processing system including an information processing device, the information processing device including:

[0014] A calculation processing unit configured to calculate an objective function related to a combination of a biological molecule and a labeled fluorescent body based on first data listing information about a biological molecule as a labeled target in a sample and second data listing information about the labeled fluorescent body; and an allocation information output unit configured to output information about the allocation of the labeled fluorescent body to each biological molecule, the information being generated based on combination information obtained from the combination optimization processing unit based on coefficients of the objective function.

[0015] The information processing device may acquire first data input by a user.

[0016] The information processing system may further include a database configured to store the second data, and

[0017] The information processing device may acquire the second data by referring to a database based on the first data.

[0018] The information processing device may acquire identification information of an algorithm used to acquire the allocation information, and

[0019] The algorithm may include at least a combinatorial optimization algorithm executed by a combinatorial optimization processing unit.

[0020] The objective function may be an objective function formulated as a constrained optimal selection problem of basis vector candidates.

[0021] Candidates for basis vectors selected in the optimal selection problem may correspond to labeled fluorescent bodies that label biomolecules as labeling targets.

[0022] The objective function may be an objective function of a binary variable of quadratic form.

[0023] The objective function may be configured to indicate linear independence between the fluorescence spectra of the labeling fluorophores assigned to each biomolecule.

[0024] The objective function can be configured to include constraints on the number of fluorophores assigned to a biomolecule.

[0025] The objective function can be configured to include the constraint that a fluorophore is assigned to a biomolecule.

[0026] The objective function may be configured to include constraints regarding fluorescence intensity with respect to a combination of a biomolecule and a fluorophore assigned to the biomolecule.

[0027] The fluorescence intensity may be based on the amount of the biomolecule and the fluorescence intensity of the fluorophore assigned to the biomolecule.

[0028] The objective function can be configured to include the dispersion of fluorescence intensity information.

[0029] The information processing device may include a communication device capable of transmitting coefficients of an objective function obtained through calculation processing to an Ising machine that specifically optimizes binary quadratic variables, and receiving values ​​of binary variables selected by the ionizer.

[0030] The Ising machine may have a function of selecting one or more combinations of spin variables corresponding to binary variables based on coefficients of an objective function.

[0031] The information processing apparatus can acquire data on the combination of the biomolecule and the fluorescent substance based on the value of the spin variable acquired by the Ising machine.

[0032] The information processing device can acquire data on the combination of the biomolecule and the fluorescent body by converting the spin variable into a binary variable.

[0033] The biomolecule can be an antigen or an antibody.

[0034] In addition, the present technology provides an information processing method, including:

[0035] a calculation processing step of calculating a target function related to a combination of a biomolecule and a labeled fluorescent body based on first data listing information of a biomolecule as a labeled target in a sample and second data listing information of a fluorescent signal of the labeled fluorescent body; and

[0036] The output step outputs information on the allocation of the labeled fluorescent body to each biomolecule, the information being generated based on the combination information acquired from the combinatorial optimization processing unit based on the coefficient of the objective function.

[0037] Furthermore, the present technology provides a program for causing an information processing device to execute the following steps:

[0038] a calculation processing step of calculating a target function related to a combination of a biomolecule and a labeled fluorescent body based on first data listing information of a biomolecule as a labeled target in a sample and second data listing information of a fluorescent signal of the labeled fluorescent body; and

[0039] The output step outputs information on the allocation of the labeled fluorescent body to each biomolecule, the information being generated based on the combination information acquired from the combinatorial optimization processing unit based on the coefficient of the objective function.

[0040] In addition, the present technology provides an information processing device, including:

[0041] a calculation processing unit that calculates an objective function related to a combination of the biomolecule and the labeled fluorescent body based on first data listing information of the biomolecule as a labeled target in the sample and second data listing information of a fluorescent signal of the labeled fluorescent body; and

[0042] The allocation information output unit is configured to output allocation information of the labeled fluorescent substance to each biomolecule based on the combination information obtained from the combinatorial optimization processing unit based on the coefficient of the objective function.

[0043] Furthermore, the present technology provides a computing device including a combinatorial optimization processing unit configured to perform combinatorial optimization processing using coefficients of an objective function formulated as a constrained optimal selection problem of basis vectors from basis vector candidates.

[0044] According to an embodiment of the present technology, the combinatorial optimization processing unit may perform combinatorial optimization processing of waveform patterns based on coefficients of a combinatorial optimization objective function, wherein the combinatorial optimization objective function is used to assign a waveform pattern to each piece of communication information based on first data including communication information and second data including a waveform pattern of a communication carrier.

[0045] According to another embodiment of the present technology, a combinatorial optimization processing unit performs combinatorial optimization processing of waveform patterns based on first data including spread codes and second data including waveform patterns of spread codes, based on coefficients of a combinatorial optimization objective function for assigning waveform patterns to each spread code. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of the configuration of filter-type and spectral-type flow cytometers.

[0047] Figure 2 Graph showing the performance degradation of the demixing process.

[0048] Figure 3 This is a schematic diagram of the Ising model.

[0049] Figure 4 is a diagram showing thermal fluctuations and quantum fluctuations.

[0050] Figure 5 is a block diagram of an example of an information processing system according to the present technology.

[0051] Figure 6 This is a block diagram of an example of an information processing device included in an information processing system according to the present technology.

[0052] Figure 7 The present invention is a flowchart of an information processing method.

[0053] Figure 8 The present invention is a flowchart of an optimization process performed in an information processing method according to the present technology. DETAILED DESCRIPTION

[0054] [Description of Embodiments]

[0055] Hereinafter, suitable embodiments for implementing the present technology will be described. The embodiments described below show typical embodiments of the present technology, and the scope of the present technology is not limited to these embodiments. The technology will be described in the following order.

[0056] 1. First Embodiment (Information Processing System)

[0057] (1) Description of the First Embodiment

[0058] (2) First Example of the First Implementation Method

[0059] (2-1) Configuration example of information processing system of the present technology

[0060] (2-2) Example of the information processing system of the present technology configured as a particle analysis system

[0061] (2-3) Example of processing in the information processing system of the present technology

[0062] 2. Second Embodiment (Information Processing Method)

[0063] 3. Third Implementation Method (Procedure)

[0064] 4. Fourth Embodiment (Information Processing Device)

[0065] 5. Fifth Embodiment (Calculation Device)

[0066] 1. First Embodiment (Information Processing System)

[0067] (1) Description of the First Embodiment

[0068] An information processing system according to the present technology includes an information processing device, the information processing device including: a calculation processing unit configured to calculate an objective function regarding a combination of a biomolecule and a labeled fluorescent body based on first data listing information regarding a biomolecule as a labeled target in a sample and second data listing information regarding a labeled fluorescent body; and an allocation information output unit configured to output information regarding allocation of the labeled fluorescent body to each biomolecule, the information being generated based on combination information acquired from a combination optimization processing unit based on coefficients of the objective function. Using the calculation processing unit and the allocation information output unit, a better combination of a biomolecule and a labeled fluorescent body can be automatically selected. In addition, the calculation processing unit and the allocation information output unit can prevent, for example, degradation of performance of an unmixing process.

[0069] The basic concept of the present technology will be described in detail below.

[0070] As described above, for example, from the viewpoint of the optical system used for fluorescence measurement, flow cytometers can be roughly divided into filter type and spectral type. Figure 1 Describe these two types of flow cytometers.

[0071] For example, as in Figure 1 As shown in , the filter type flow cytometer uses a plurality of half mirrors HM to branch fluorescence into multiple ones, and performs measurement through different filters using, for example, a photomultiplier tube PMT. For example, the filter type also acquires a generalized spectrum. However, in the filter type, it is necessary to prepare an optical system of a half mirror for each wavelength, and there is a problem that it is difficult to increase the number of channels due to the constraints of the optical arrangement. On the other hand, as Figure 1 As shown in 2, one feature of the spectral type flow cytometer is that fluorescence is acquired as a spectrum by using a prism splitting optical element P. Specifically, the spectral fluorescence is acquired as multidimensional (multi-channel) data using a photomultiplier tube PMT prepared for each wavelength. The spectral type is suitable for increasing the number of channels in principle because the spectrum itself is easy. The spectral type has the advantage that fluorescence can be analyzed in detail because the spectrum can be measured in multiple channels.

[0072] Because spectral flow cytometer can analyze fluorescence spectrum in detail as described above, spectral flow cytometer can increase the type of fluorescent labeling and check the expression of a large number of different antigens at the same time. However, the fluorescence spectrum actually obtained includes the fluorescence of the mixture of different fluorescent labels. Therefore, in order to evaluate the fluorescence of each fluorescent labeling, it is necessary to decompose (separate) the mixed fluorescence. This decomposition process is called unmixing processing. The most commonly used scheme in unmixing processing is the (weighted) least square method. In this method, the fluorescence spectrum actually obtained is regarded as the linear sum of the fluorescence spectrum of the dye of each fluorescent labeling, and the coefficient with the minimum (weighted) square error is set to the amount of the fluorescent labeling bound to the antigen.

[0073] In the present specification, “fluorescent labeling” refers to an antibody labeled with a fluorescent body, and is also referred to as, for example, a fluorescently labeled antibody.

[0074] The unmixing process can be described as follows.

[0075] Assume that a fluorescence spectrum is obtained from one cell by flow cytometry as shown in the following equation 1.

[0076] [Mathematical formula 1]

[0077] t=(t1...t N ) T …(1)

[0078] Here, N is the number of channels of the flow cytometer. On the other hand, it is assumed that there are M types of fluorescent markers, and as shown in the following equation 2, the fluorescence spectrum of the mth marker is obtained in advance.

[0079] [Mathematical formula.2]

[0080]

[0081] Furthermore, these fluorescence spectra are arranged to form a matrix of the following Equation 3.

[0082] [Math.3]

[0083] F=(f (1) ...f (M) )...(3)

[0084] This matrix F is called the spillover matrix. Here, the quantity to be obtained is the quantity of fluorescent markers bound to antigens on the cells. The quantity of the i-th fluorescent marker is x i , and these are arranged side by side, and a vector is obtained as shown in Equation 4 below.

[0085] [Math.4]

[0086] x=(x1...xM ) T …(4) In the unmixing process, as shown in the following equation 5, the amount x of the fluorescent label is obtained as a solution of the weighted least square method.

[0087] [Math.5]

[0088]

[0089] However, k represents the weight of the squared error of the kth channel and is determined by the magnitude of the noise in this channel.

[0090] The solution x of this weighted least squares method is * can be obtained analytically as shown in Equation 6 below.

[0091] [Math.6]

[0092] x * =(F T WF) -1 F T Wt …(6)

[0093] Here, w is a matrix in which weights are arranged as diagonal components, and is as shown in the following Equation 7.

[0094] [Math.7]

[0095] W=diag(w) …(7)

[0096] According to the above, the amount of each fluorescent label attached to the antigen can be calculated by unmixing processing of the fluorescence spectrum, and the expression of various antigens can be evaluated simultaneously.

[0097] However, the performance of the demixing process may deteriorate, for example, in the following case.

[0098] The fluorescence spectra of two or more different fluorescent labels used are the same or similar, and

[0099] The case where the combination of fluorescence spectra of two or more different fluorescent labels used is the same as or similar to the spectra of other fluorescent labels.

[0100] Assume a situation in which, for cells expressing antigens A and B, e.g. Figure 2 As shown in , antibody a labeled with fluorescent dye α2 is used as a fluorescent marker conjugated to antigen A, and antibody b labeled with fluorescent dye β2 is used as a fluorescent marker conjugated to antigen B. In this case, since the fluorescence spectra of fluorescent dye α2 and fluorescent dye β2 are similar, the performance of the unmixing process may be deteriorated.

[0101] The selection of the combination of the fluorescent markers, that is, the selection of the combination of the antibody and the antibody-labeled fluorescent body, is checked so as to prevent the deterioration of the performance of the unmixing process. For example, multiple types of antibodies labeled with different fluorescent bodies are prepared for each antigen, and a better fluorescent marker is selected from the multiple types of fluorescent markers to prevent the deterioration of the performance of the unmixing process. More specifically, the user first specifies multiple antigens to be checked, and then selects a fluorescent marker to be attached to each antigen, and this selection can be made so as not to correspond to the above situation.

[0102] For measurement in an apparatus or system that performs measurement using a combination of an antibody and a fluorophore, such as a flow cytometer, it is conceivable to examine the following two things in order to assign a fluorescent label with good fluorescence separation performance to each of a plurality of antigens that are measurement targets.

[0103] (i) Provide indicators for evaluating which combination is good

[0104] (ii) provide a mechanism for searching a large number of combinations at high speed;

[0105] Regarding (i) above, it is considered to use, for example, the correlation between fluorescence spectra as an indicator. More specifically, it is considered to calculate the correlation between fluorescence spectra and to combine the phosphors in order from the phosphor with the smallest correlation so that more than two phosphors with the same or similar fluorescence spectra are not used. However, in this scheme, as the combination process proceeds, highly correlated combinations still exist, and overall optimization may not be achieved in the end. In addition, in this scheme, it is sometimes impossible to avoid the situation where the combination of the spectra of the above-mentioned two or more different phosphors is the same or similar to the spectra of other phosphors.

[0106] For (ii) above, as the number of antigen types and fluorescent marker types increases, the demand for inspection increases. More specifically, when the number of types of antigens to be evaluated and / or the number of types of fluorescent markers for each type of antigen increases simultaneously, the number of combinations of fluorescent markers that can be adopted increases dramatically. For example, when the number of antigen types is M and the number of fluorescent marker types for each antigen is K, it is necessary to select from M to K. K The number of combinations of types to obtain a good combination is often increased. It is often necessary to evaluate more types of antigens simultaneously, and the number of types of fluorescent labels for each antigen is also increasing. Therefore, the number of combinations to be examined increases exponentially.

[0107] As an index regarding the performance of the unmixing process, the present inventors focused on the determinant of a matrix F regarding the fluorescent spectra of arranged fluorescent bodies in the following Equation 8.

[0108] [Math.8]

[0109] det(F TF) …(8)

[0110] It is considered that the performance of the demixing process can be improved by adopting a combination that maximizes the determinant in the above Equation 8. Hereinafter, the determinant will be described in more detail.

[0111] When considering two situations in which the performance of the unmixing process may deteriorate, it can be assumed that the condition of "the combination of fluorescence spectra of two or more different fluorescent markers is the same or similar to the spectra of other fluorescent markers" should be met in order to prevent the performance of the unmixing process from deteriorating. This condition means that the fluorescence spectra of all fluorescent markers are mainly independent. The main independence of vectors is usually indicated by the determinant of the matrix in which the vectors are arranged to be non-zero. However, because the overflow matrix F of equation 3 (which is a matrix in which the fluorescence spectra of the above-mentioned fluorescent markers are arranged) is not a square matrix, the determinant cannot be calculated. Therefore, the determinant in equation 8 above is the determinant of the square matrix of equation 9 below, which is considered to replace the overflow matrix F.

[0112] [Math.9]

[0113] F T F …(9)

[0114] In the following, when the determinant det(F T F) is not 0, which shows that the fluorescence spectra of the fluorescent labels are largely independent.

[0115] First, when the fluorescence spectrum is mainly dependent,

[0116] [Math.10]

[0117] det(F T F)=0 …(10)

[0118] is shown in Equation 10 above.

[0119] The fact that fluorescence spectra are primarily dependent means that the kth fluorescence spectrum f k Other spectra f can be written as m A linear combination of . That is, f k and f m The relationship shown in the following equation 11 is satisfied.

[0120] [Mathematical formula.11]

[0121]

[0122] Therefore, the square matrix F of equation 9 above is T F can be written as shown in Equation 12 below.

[0123] [Math.12]

[0124]

[0125] As shown in Equation 12 above, because the kth column is written as a linear combination of the other columns, the square matrix F T F is the primary slave. Therefore, as shown in equation 10 above.

[0126] Next, it will be shown that when the determinant det(F T When F) is 0, the matrix is ​​primarily dependent.

[0127] When the relationship in equation 10 above is satisfied, F T The kth column of F is written as a linear combination of the other columns. Therefore, this yields Equation 13 below.

[0128] [Math.13]

[0129]

[0130] In Equation 13 above, since the kth spectrum f k The other spectrum f is written as m The linear combination of T F is the primary affiliation.

[0131] From the above, it can be understood that the spectrum of the fluorescent label is mainly independent of the following conditions:

[0132] [Math.14]

[0133] det(F T F)≠0…(14)

[0134] Next, we show that when the determinant det(F T When F) is larger, the decomposition performance tends to be good.

[0135] First, the determinant det(F T F) is equal to or greater than 0, because the matrix F T F is a positive semidefinite matrix.

[0136] First, prove that the matrix F T F is a positive semidefinite matrix.

[0137] When F T When F is a semi-positive definite matrix, the following equation 15 is established due to the properties of the semi-positive definite matrix.

[0138] [Math.15]

[0139]

[0140] In the above Equation 15, when the left side is deformed, the following Equation 16 is obtained.

[0141] [Math.16]

[0142] x T (F T F)x=(Fx) T (Fx)=Fx·Fx≥0 …(16)

[0143] As shown in Equation 16 above, F T F is a positive semidefinite matrix.

[0144] Then, prove that the determinant of a semidefinite matrix is ​​equal to or greater than 0.

[0145] According to the characteristics of semi-positive definite matrices, about the matrix F T The following eigenvalues ​​of F are all non-negative, i.e., as shown in Equation 18 below.

[0146] [Math.17]

[0147] (λ1...λ M ) …(17)

[0148] [Math.18]

[0149] 0≤λ i (i=1...M) …(18)

[0150] On the other hand, according to the relationship between the determinant and the eigenvalue, a relationship as shown in the following Equation 19 can be obtained.

[0151] [Math.19]

[0152]

[0153] Therefore, the determinant in the above Equation 8 has a value equal to or greater than 0.

[0154] It is shown above that the determinant in the above equation 8 is equal to or greater than 0.

[0155] Next, it will be shown below that when the value of the determinant deviates from 0, the performance of the decomposition tends to improve.

[0156] First, the synthesis of the fluorescence spectrum is modeled as follows Equation 20.

[0157] [Math.20]

[0158] y=Fx+n …(20)

[0159] Modeling in this way makes it possible to express the decomposition of the fluorescence spectrum as shown in Equation 21 below.

[0160] [Math.21]

[0161] x′=x+(F T F) -1 F T n …(21)

[0162] However, here, the weight matrix is ​​the identity matrix.

[0163] First, consider the case where there are only two fluorescent labels to be separated. Furthermore, for simplicity, assume that the fluorescence spectra are normalized by the square norm. In this case, when calculating the matrix F T When F

[0164] The matrix is ​​expressed as in Equation 22 below.

[0165] [Math.22]

[0166]

[0167] In Equation 22, when the spectrum is normalized, the diagonal terms are all 1, and as shown in Equation 23 below.

[0168] [Math.23]

[0169] f1·f1=f2·f2=1 …(23)

[0170] Furthermore, in Equation 22 above, all off-diagonal terms are correlation coefficients r. That is, the matrix F T F is equivalent to the correlation matrix R as shown in Equation 24 below.

[0171] [Math.24]

[0172]

[0173] Furthermore, the matrix F can be obtained as shown in the following equations 25 and 26 T The inverse matrix and determinant of F.

[0174] [Math.25]

[0175]

[0176] [Math.26]

[0177] det(F T F)=|R|=1-r 2 …(26)

[0178] Consider using these to calculate the noise component after unmixing. As shown in Equation 27 below, the noise n is a random number in Equation 28 below, which is a standard normal distribution,

[0179] [Math.27]

[0180]

[0181] [Math.28]

[0182] ∈ i (i=1...N)…(28)

[0183] For expression, the noise after unmixing is as shown in Equation 29 below.

[0184] [Math.29]

[0185]

[0186] As described above, since the fluorescence spectrum is normalized by the square norm, the random numbers of the following equation 30 of the standard normal distribution are newly used to obtain the following equation 31.

[0187] [Math.30]

[0188] ∈′1,∈′2…(30)

[0189] [Math.31]

[0190]

[0191] In addition, the random numbers of the following equation 32 are standard normal distribution, which is newly used to obtain the following equation 33.

[0192] [Math.32]

[0193] ∈″1,∈2″ …(32)

[0194] [Math.33]

[0195]

[0196] The noise decreases monotonically with respect to the following equation 34,

[0197] [Math.34]

[0198] |R|=det(F T F)…(34)

[0199] The minimum value is reached when the maximum value is 1. That is, when the noise added to the fluorescence spectrum is Gaussian noise having the same intensity for each channel, the performance of the unmixing process is better when the determinant in the above equation 8 is larger.

[0200] As described above, in the case where there are only two fluorescent labels to be separated, it can be seen that the performance of the unmixing process becomes better when the determinant in the above equation 8 is larger. In addition, even in the case where there are three or more fluorescent labels to be separated, when the determinant in the above (8) is larger, the performance of the unmixing process tends to become better, and this can be shown, for example, by performing simulations.

[0201] As described above, the combination of fluorescent labels that is expected to improve the performance of the unmixing process is considered to be a combination that maximizes the determinant in equation 8 above about the matrix F that arranges the fluorescence spectra. Here, it is expected that an algorithm called quantum annealing or simulated annealing or a simulated branching algorithm similar to quantum annealing will be effective for the problem of combinatorial optimization. For any algorithm, by treating the combinatorial optimization problem as a binary quadratic optimization problem, processing using the algorithm becomes possible.

[0202] Quantum annealing and the Ising machine will be described below.

[0203] First, quantum annealing will be described.

[0204] This quantum annealing is known as a hardware system that uses a quantum device to perform combinatorial optimization using multiple quantum bits. Here, the quantum bit is an information unit (bit) that can realize quantum superposition and quantum entanglement. Quantum superposition is a phenomenon in which two states (values) are maintained in parallel and probabilistically (quantum probability amplitude) with 1 bit (1Qbit), and quantum entanglement is a phenomenon in which a correlation state between 2 bits that cannot be realized conventionally is realized.

[0205] Because quantum annealing is a type of calculation using qubits, quantum annealing is often confused with quantum computing itself. However, the operation is very different from the quantum computers that have been examined traditionally. However, because the calculation is still a calculation using qubits, this is now often positioned as a quantum computing scheme.

[0206] A computation that can be accomplished by quantum annealing is to obtain a solution to a combinatorial optimization problem for minimizing an objective function in quadratic format of Equation 35 below.

[0207] [Math.35]

[0208]

[0209] In the above equation 35, σ i , σ j and σ k is a variable that takes one of two values, 0 and 1, and is a variable for which an optimal solution is desired.ij and h k are parameters and are given externally. Here, for ease of explanation, these parameters are given in σ i The parameter is ±1 instead of 0 or 1 (for convenience).

[0210] In quantum annealing, first, a physical model called the Ising model is realized using quantum devices and circuits. Figure 3 Schematic diagram illustrating the Ising model.

[0211] The Ising model is a physical model proposed as a simple model for explaining the magnetism of a magnet. Figure 3 , each grid point represents an electron spin for which only an upward or downward spin torque is observed.

[0212] In statistical physics, an energy function is used to describe the behavior of a physical model, but the energy function of the physical model is described as shown in the following Equation 36.

[0213] [Math.36]

[0214]

[0215] σ i , σ j and σ k is a variable indicating whether the spin is up (-1) or down (1). In addition, J ij is the bonding energy that determines the magnitude of the interaction between the spins, h k is the component of the external magnetic field parallel to the spin. As for the bonding energy between spins, a model in which only the bonding energy between adjacent spins has a non-zero value has been adopted in the model used to explain the ferromagnetism of magnets, but it is generalized here so that the bonding energy between any spins has a non-zero value. This model is called spin glass.

[0216] It is easy to understand that the energy function of the Ising model is equivalent to the objective function of the combinatorial optimization. The details of the variable conversion of ±1 and 0, 1, etc. are omitted. Therefore, when the Ising model with the same parameters as the objective function of the combinatorial optimization is artificially configured and the spin of the basic state (minimum energy state) of the energy function is measured, the starting point of the idea of ​​quantum annealing is to obtain the optimal solution (not local optimization but global optimization).

[0217] The minimum energy state is achieved by a stepwise adiabatic transition from the superposition state to the fundamental state (a state transition slow enough to be considered a stable state) using ultra-low temperatures. This process is called quantum annealing.

[0218] Specifically, first, initially, a transverse magnetic field (horizontal magnetic field) for achieving a superposition state of all combinations in the following equation 37 is applied.

[0219] [Math.37]

[0220]

[0221] As shown in Equation 38 below, the superposition term is gradually relaxed,

[0222] [Math.38]

[0223]

[0224] Finally the ground state is achieved. In the next step, the spin is measured in the ground state and used as the optimal solution.

[0225] This process is "slow" in an adiabatic process, but in a practical machine the whole process can be completed in microseconds.

[0226] In addition, for example Figure 4 As shown in FIG. 1 , since the tunneling effect caused by quantum fluctuations is used, it is considered that the process can be realized at a higher speed than simulated annealing using thermal fluctuations. The effect has also been experimentally confirmed in a specific problem setting.

[0227] Regarding quantum devices that realize qubits, for example, quantum devices using ions, light, and single electrons (quantum dots) are known, but the most effective quantum devices at present are superconducting devices. Extremely low temperatures are required to achieve superconductivity. Therefore, a large number of cooling devices were required in the past. However, with the development of cloud technology in recent years, these cooling devices are installed on the server side, so that the calculation results can only be obtained on the edge side.

[0228] In addition, in the case of implementing quantum annealing, when combinatorial optimization using the Ising model has become a problem, for example, a device that uses existing digital circuits and gate circuits to accelerate the combinatorial optimization itself represented by the Ising model without using quantum annealing has appeared. This can also be achieved on the edge side because it can be achieved by digital circuits without using superconductivity.

[0229] The inventors have found that the problem of optimal selection of fluorescent markers for flow cytometry can be formulated as the problem of optimal selection of basis vector candidates. Now, assume that there are L basis vector candidates, and the i-th basis vector is represented by u i Indicates. Using these basis vector candidates, the matrix U in Equation 39 below is generated.

[0230] [Math.39]

[0231] U=(u1...u L )…(39)

[0232] Then, it indicates whether to select the i-th basis vector u i The value of is shown in Equation 40 below.

[0233] [Math.40]

[0234] b i ∈{0,1} …(40)

[0235] That is, when the basis vector u i When selected, the binary value b i is 1 if the value is true, otherwise it is 0.

[0236] where b i The matrix arranged with diagonal entries is B of the following equation 41.

[0237] [Math.41]

[0238] B=diag(b1...b L ) …(41) The constraints on how to choose the basis vectors are imposed on the binary values ​​b i For example, when the number of selected basis vectors is M (M < L), the following constraints are imposed.

[0239] [Math.42]

[0240]

[0241] Furthermore, when the number of basis vectors to be selected is M (M < L) and the mth basis vector is M in the following equation 43 m When one of the candidates is selected, the M constraints of the following equation 44 are imposed.

[0242] [Math.43]

[0243]

[0244] [Math.44]

[0245]

[0246] When the matrices U and B of the above equations 39 and 41 are used, the next matrix Q becomes a matrix configured by the selected basis vectors or the 0 vector.

[0247] [Math.45]

[0248] Q=UB=(b1u1..) …(45)

[0249] On the other hand, the following matrix configured only by the selected basis vectors is F as in the following Equation 46 as before.

[0250] [Math.46]

[0251]

[0252] Here, the following Equation 47 indicates a combination of numbers of M basis vector candidates selected as basis vectors among L basis vector candidates.

[0253] [Math.47]

[0254] (i1...i M ) …(47) In the present description, a good combination of basis vectors is one that maximizes the determinant in equation 8 above. However, this determinant is difficult to convert into a binary quadratic form because the variable b indicating the choice of basis vectors i does not appear clearly. So, for example, try to maximize the determinant Q of equation 48 below T Q, instead of the determinant F in Equation 9 T F.

[0255] [Math.48]

[0256]

[0257] However, in practice, this is as shown in Equation 49 below.

[0258] [Math.49]

[0259] QUR T Q=0 …(49)

[0260] This is because in the matrix Q T In Q, the row and column components of i (as shown in equation 50 below) are all zero.

[0261] [Math.50]

[0262] b i =0 …(50)

[0263] Therefore, for row i and column i where all components are 0, only the diagonal components are replaced with 1. This operation can be implemented as shown in Equation 51 below.

[0264] [Math.51]

[0265] A=I-B+Q T Q …(51)

[0266] Thus, the relationship of the following equation 52 is established.

[0267] [Math.52]

[0268] detA=detF TF …(52)

[0269] This relationship can be seen by considering the adjustment matrix (the matrix without rows and columns) with the diagonal components replaced by 1.

[0270] Thus, the problem of optimal selection of basis vectors is replaced by the maximization problem of Equation 53 below.

[0271] [Math.53]

[0272] det(I-B+Q T Q) …(53)

[0273] That is, the problem of optimal selection of basis vectors is formulated as a combinatorial optimization problem of Equation 54 below.

[0274] [Math.54]

[0275]

[0276] Subsequently, the combinatorial optimization problem is considered to be a combinatorial optimization problem of binary quadratic form. When the matrix A is written down in more detail, the following Equation 55 is obtained.

[0277] [Math.55]

[0278]

[0279] Here,

[0280] [Math.56]

[0281] r ij =u i ·u j …(56)

[0282] is the basis vector u i and u j In addition, R is a matrix defined by the variable b to be optimized, which is represented by the following equation 57.

[0283] [Math.57]

[0284]

[0285] Using this, the combinatorial optimization problem can be transformed as shown in Equation 58 below.

[0286] [Math.58]

[0287]

[0288] In Equation 58, since the determinant still exists, it is necessary to write down the determinant. However, when the determinant is written as it is, the number of terms increases as an exponential function of the rank of the matrix, making calculation difficult. Therefore, consider the Taylor expansion of the determinant as shown in Equation 59 below.

[0289] [Math.59]

[0290]

[0291] For each differential coefficient, the first-order derivative is obtained according to the definition of differential as shown in the following equation 60.

[0292] [Math.60]

[0293]

[0294] Here, F(x) is defined as the following Equation 61.

[0295] [Math.61]

[0296] F(x)=A -1 (I+Ax) …(61)

[0297] Similarly, when the second-order derivative is obtained, the second-order derivative is as shown in the following Equation 62.

[0298] [Math.62]

[0299] f (2) (x) = f′(x)tr(F -1 )-f(x)tr(F -2 )

[0300] =f(x)(tr(F -1 )) 2 -tr(F -2 )…(62)

[0301] Similarly, the third order derivative is shown in Equation 63 below.

[0302] [Math.63]

[0303] f (3) (x) = f′(x)((tr(F -1 )) 2 -tr(F -2 )+f(x)(-2tr(F -1 )tr(F -2 )+2tr(F -3 ))

[0304] =f(x)((tr(F -1 ))3 -3tr(F -1 )tr(F -2 )+2tr(F -3 ))…(63)

[0305] On the other hand, when x = 0, replace f(x) and F -1 (x), the following equations 64 and 65 are obtained.

[0306] [Math.64]

[0307] f(0)=0 …(64)

[0308] [Math.65]

[0309] F -1 (0) = A …(65)

[0310] From the above, the following equations 66, 67 and 68 can be obtained.

[0311] [Math.66]

[0312] f (1) (0) = tr(A) …(66)

[0313] [Math.67]

[0314] f (2) (0) = tr(A) 2 -tr(A 2 ) …(67)

[0315] [Math.68]

[0316] f (3) (0) = tr(A) 3 -3tr(A)tr(A 2 )+2tr(A 3 ) …(68)

[0317] Using this result, the Taylor expansion of f(x) is written as shown below in Equation 69. Refer to the above reference web page until Equation 69 is obtained.

[0318] [Math.69]

[0319]

[0320] Then,

[0321] [Math.70]

[0322]

[0323] As shown in the following equations 71, 72 and 73, tr(A), tr(A) are calculated for the actual R. 2 ) and tr(A 3 ).

[0324] [Math.71]

[0325] tr(A)=0 …(71)

[0326] [Math.72]

[0327]

[0328] [Math.73]

[0329]

[0330] As described above, the following Equation 74 can be obtained.

[0331] [Math.74]

[0332]

[0333] That is, the optimal selection problem of basis vector candidates can be considered as a problem of minimizing the objective function of the following equation 75.

[0334] [Math.75]

[0335]

[0336] Considering the above constraints on basis vectors, the combinatorial optimization equation is formulated as shown in Equation 76 below.

[0337] [Math.76]

[0338]

[0339] In Equation 76, λ m is a positive real parameter and is large so that the second term is always satisfied, or is a dynamically determined value.

[0340] The constraint term of the second term in equation 76 is a binary quadratic form, as can be easily understood. However, L(b) in the first term has three or higher terms and does not have a binary quadratic form. Therefore, equation 76 cannot be implemented in an Ising machine.

[0341] Therefore, here, L(b) is approximated using up to quadratic terms. In this case, when the combinatorial optimization is formulated, the combinatorial optimization can be written as follows.

[0342] [Math.77]

[0343]

[0344] Instead of Equation 77, as the derivative form that can be easily considered from Equation 77,

[0345] [Math.78]

[0346]

[0347] A monotonically increasing function of the following equation 79 may be used.

[0348] [Math.79]

[0349] r ik ≥0 …(79)

[0350] For example, this is the following equation 80 etc.

[0351] [Math.80]

[0352] (r ik ) n , 1≤n …(80)

[0353] In the following, when the conversion is performed without changing Equation 78, this combination is optimized to a binary quadratic form and can be easily implemented in an Ising machine. Specifically, the following Equation 81

[0354] [Math.81]

[0355] σ i =1-2b i …(81)

[0356] is used for transformation so that Equation 77 can be formulated as a combinatorial optimization problem as shown below in Equation 82. Therefore, the objective function formulated as a constrained optimal selection problem of basis vector candidates can be used for the selection of fluorescent markers.

[0357] [Math.82]

[0358]

[0359] The objective function of the above equation 82 becomes the Hamiltonian of the Ising model. By using the program, the coefficients of the objective function of the above equation 82 can be easily obtained. Therefore, by replacing the Ising machine with the obtained coefficients, the spin can be optimized. By returning the obtained optimized spin to a binary variable, the optimized solution can be obtained, as shown in the following equation 83.

[0360] [Math.83]

[0361]

[0362] The above approximates L(b) using terms up to quadratic terms for the actual R, but it is also possible to perform the approximation using terms up to slightly higher dimensions. Specifically, consider simplifying the three terms of equation 84 below to quadratic terms.

[0363] [Math.84]

[0364]

[0365] First, the following Equation 85 is obtained by using the above Equation 81.

[0366] [Math.85]

[0367]

[0368] In the above Equation 85, the following Equation 86 is expressed as the third term with respect to spin.

[0369] [Math.86]

[0370]

[0371] To reduce this dimension, the following auxiliary spin of Equation 87 is prepared.

[0372] [Math.87]

[0373] σ kl =σ k σ l …(87)By using the auxiliary spin of Equation 87 above, the size can be reduced to a quadratic term as shown in Equation 88 below.

[0374] [Math.88]

[0375]

[0376] Regarding the above Equation 88, the following Equation 89 needs to be optimized to 0 as a constraint not to cause a mismatch between the auxiliary rotation and the original rotation.

[0377] [Math.89]

[0378] H kl =3σ kl +σ k σ l -2σ k σ kl -2σ l σ kl …(89)

[0379] This term is a constraint term added to the objective function, but since this term is a quadratic term of the spin, this term can be implemented in the Ising machine. By using the program, the primary and secondary coefficients of the spin of the Ising machine to be replaced can be easily obtained. When the Ising machine is replaced by the coefficients obtained by the program, the spin in the Ising machine can be optimized. The optimized solution can be obtained by returning the obtained optimized spin to a binary variable, as shown in the following equation 90.

[0380] [Math.90]

[0381]

[0382] As described above, in order to automatically select a better combination of a biomolecule and a labeled fluorescent body, an objective function related to the combination of the above-mentioned biomolecule and the labeled fluorescent body can be used. As described above, the objective function, specifically the coefficient of the objective function, can be obtained, for example, using a program. For example, a general information processing device such as a computer owned by a user who selects the combination can calculate the objective function, specifically the coefficient of the objective function. The combinatorial optimization processing unit that performs combinatorial optimization processing generates combination information by using the objective function, and uses the generated combination information, so that the general information processing device can generate information about the allocation of the labeled fluorescent body to each biomolecule.

[0383] (2) First Example of the First Implementation Method

[0384] (2-1) Configuration example of information processing system of the present technology

[0385] The following will refer to Figure 5 An information processing system according to the present technology is described. Figure 5 A configuration example of an information processing system according to the present technology is shown.

[0386] Figure 5 The information processing system 1 shown in FIG. 1 includes an information processing device 2, a database 3, a communication device 4, an input device 5, an output device 6, and a calculation device 7. The information processing system 1 may be a system for presenting a combination of a biomolecule (specifically, an antigen or an antibody) and an antibody-labeled fluorescent body used in a particle analysis device such as a flow cytometer to a user.

[0387] Figure 6 A block diagram of the information processing device 2 is shown in FIG. Figure 6 As shown, the information processing device 2 includes a processing unit 10. The processing unit 10 includes a calculation processing unit 11 and an allocation information output unit 12. The calculation processing unit 11 calculates the target function for the combination of the biomolecule and the labeled fluorescent body, and the allocation information output unit 12 outputs information about the allocation of the labeled fluorescent body to each biomolecule. The information processing device 2 may further include a storage unit 13.

[0388] The calculation processing unit 11 can calculate the target function related to the combination of the biomolecule and the labeling fluorescent body based on the first data listing the information of the biomolecule as the labeling target of the sample and the second data listing the information of the labeling fluorescent body.

[0389] The first data includes information about a biomolecule that is a target of labeling of the sample. The first data may be data input by a user. The information about the biomolecule may be, for example, the name or abbreviation of the biomolecule, or a number (e.g., an ID number) for specifying the biomolecule. The biomolecule may be, for example, a biomolecule designated as a target of labeling by a user. The biomolecule may be, for example, an antibody or an antigen. The sample may be, for example, a sample containing particles to be described below, or may be, more specifically, a bioparticle, and even more specifically, a sample comprising cells. The biomolecule may be a cell surface marker.

[0390] The second data contains information of the labeled fluorescent body. The information preferably includes the fluorescence spectrum of each labeled fluorescent body. For example, the information can be the name or abbreviation of the labeled fluorescent body, or a number (for example, an ID number) for specifying the labeled fluorescent body.

[0391] The objective function is preferably an objective function formulated as a constrained optimal selection problem of basis vector candidates. The basis vector candidates selected in the optimal selection problem may correspond to a marker fluorophore that marks a biomolecule as a marker target. For example, the candidate may correspond to a fluorescence spectrum of a fluorophore that may be used as a marker fluorophore.

[0392] As described in the above “(1) Description of the first embodiment”, in the unmixing process, the fluorescence spectrum actually obtained is regarded as the linear sum of the fluorescence spectra of each fluorescently labeled dye. That is, the fluorescence spectrum of D channels (D frequencies) can be regarded as a D-dimensional vector (D can be any positive number), and the actual fluorescence spectrum is the linear sum of the fluorescence spectrum of each labeled fluorescent body as a reference. Therefore, as described above, in the constrained optimal selection problem, the labeled fluorescent body, specifically the fluorescence spectrum of the labeled fluorescent body, can be treated as a basis vector candidate.

[0393] The objective function is preferably an objective function of a binary variable of quadratic form. Since the objective function is an objective function of a binary variable of quadratic form, the objective function can be implemented in an Ising machine.

[0394] The objective function may be configured to indicate linear independence between the fluorescence spectra of the labeled fluorescent bodies assigned to the respective biomolecules. For example, the objective function may be an objective function based on a determinant indicating linear independence. The objective function based on a determinant indicating linear independence may be the determinant det(F) of equation 8 described in the above “(1) Description of the first embodiment”. T F) or an objective function of equivalent determinant.

[0395] The objective function may preferably be configured to include a constraint on the number of fluorophores assigned to one biomolecule, and may more preferably be configured to include a constraint that one fluorophore is assigned to one biomolecule.

[0396] The objective function may be configured to include constraints regarding fluorescence intensity regarding a combination of a biomolecule and a fluorophore assigned to the biomolecule.The fluorescence intensity may be based on the amount of the biomolecule and the fluorescence intensity of the fluorophore assigned to the biomolecule.

[0397] Based on these constraints, for example, brighter marker fluorophores can be assigned to less abundant biomolecules and dimmer marker fluorophores can be assigned to more abundant biomolecules.

[0398] The objective function may be, for example, an objective function obtained by converting equation 76 described in "(1) Description of the first embodiment" above into a binary quadratic form. The objective function converted into a binary quadratic form is, for example, the objective function of equation 82 described in "(1) Description of the first embodiment" above.

[0399] In a preferred embodiment of the present technology, the objective function can be configured to include the dispersion of fluorescence intensity. For example, the objective function can include a term regarding the dispersion of fluorescence intensity. More specifically, the objective function can be based on, for example, a term regarding the dispersion of fluorescence intensity that has been added to the objective function in equation 76 described in "(1) Description of the first embodiment" above. An example of a term regarding the dispersion of fluorescence intensity will be described below.

[0400] For example, consider that the assumed intensity of the kth fluorescent marker is labeled a k . In this case, the intensity of the fluorescent label relative to the mth antigen is expressed by the following equation 91.

[0401] [Math.91]

[0402]

[0403] Furthermore, the dispersion of fluorescence intensity is expressed by the following Equation 92.

[0404] [Math.92]

[0405]

[0406] Equation 92 is a quadratic equation for b. By incorporating this dispersion of fluorescence intensity into Equation 76 above, an objective function can be obtained as formulated as shown in Equation 93 below. The objective function of Equation 93 can be employed in the present technology.

[0407] [Math.93]

[0408]

[0409] The fluorescence intensity (e.g., peak value or square norm) of the selected fluorescent marker can be aligned by a dispersed objective function combined with the fluorescence intensity. Fluorescence intensity can be based on, for example, the fluorescence intensity of the fluorophore included in the fluorescent marker and the amount of the molecule detected by the fluorescent marker (e.g., the amount of the antigen, specifically, the amount of the expression of the antigen). For example, the amount of the molecule can be, for example, divided into a plurality of stages (e.g., stages 2 to 10, specifically, stages 2 to 8, and more specifically, stages 2 to 5) and more specifically, divided into the amount of three stages including "large", "medium" and "small". Similar to the amount, the fluorescence intensity of the fluorophore can be a fluorescence intensity divided into a plurality of stages (e.g., stages 2 to 10, specifically, stages 2 to 8, and more specifically, stages 2 to 5) and more specifically, divided into the degree of brightness of the three stages including "bright", "medium" and "dark". The number of these stages can be set by the user, such as a researcher.

[0410] The allocation information output unit 12 can output the information of the allocation of the labeled fluorescent body to each biomolecule based on the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function. As will be described below, the optimization processing unit obtains the combination information based on the coefficient of the objective function. The allocation information output unit 12 can generate the allocation information of the labeled fluorescent body to each biomolecule based on the combination information.

[0411] For example, the combination information may include a spin variable. The spin variable may preferably be a spin variable optimized by the optimization processing unit. The allocation information output unit 12 is preferably capable of generating information related to the allocation of the labeled fluorescent body to each biomolecule based on the value of the spin variable. For example, the allocation information output unit can obtain information about the allocation of the labeled fluorescent body to each biomolecule by converting the spin variable into a binary variable, specifically, data about the combination of the biomolecule and the fluorescent body. For example, as described with reference to equation 90 in "(1) Description of the first embodiment" above, the conversion of the spin variable to the binary variable can be performed.

[0412] The allocation information output unit 12 outputs the generated allocation information. For example, the allocation information output unit 12 can output the allocation information to the output device 6 connected to the information processing device 2. When the output device 6 is a display device, the display device can display the allocation information so that the user can visually recognize the allocation information.

[0413] The storage unit 13 may be configured to store various types of data used in the present technology. The configuration of the storage unit 13 may be appropriately selected by those skilled in the art. For example, the storage unit 13 may store the first data, the second data, the target function, the combination information, and the allocation information.

[0414] Hereinafter, a configuration example of the information processing device 2 will be described. The processing in the processing unit 10 of the information processing device 2 can be implemented, for example, by the following configuration. For example, the information processing device 2 may include a central processing unit (CPU), a RAM, and a ROM. The CPU, the RAM, and the ROM may be connected to each other via a bus. The input and output interfaces may be further connected to the bus. The communication device 4, the input device 5, and the output device 6 may be connected to the bus via the input and output interfaces. For example, the information processing device 2 may be configured by a general-purpose computer.

[0415] The database 3 can store information used as the second data by the information processing device 2, and has, for example, second data listing information about the labeled fluorescent body. For example, in the database 3, the information is as described above. The database 3 can be set inside the information processing device 2 or can exist outside the information processing device 2, and in this case, the database 3 can be connected to the information processing device 2 by wire or wirelessly.

[0416] For example, the information processing device 2 may refer to the database 3 based on the first data and obtain the second data from the database 3. For example, the information processing device 2 may obtain information about the labeled fluorescent substance associated with each of the biomolecule groups listed in the first data. The obtained information about the labeled fluorescent substance may be used by the calculation processing unit.

[0417] For example, the database 3 may transmit the second data to the information processing device 2 in response to receiving the first data from the information processing device 2. The database 3 generates a fluorescent body list for each biomolecule included in the first data, for example, by selecting a labeled fluorescent body that can be labeled. That is, a plurality of fluorescent body lists are generated. The database 3 transmits the plurality of fluorescent body lists to the information processing device 2. The information processing device 2 may use the plurality of fluorescent body lists as the second data.

[0418] The communication device 4 transmits the objective function calculated by the information processing device 2 (specifically, the calculation processing unit 11) to the calculation device 7 (specifically, the optimization processing unit included in the calculation device 7). For example, when the Ising machine is used as the calculation device 7, the communication device 4 may be configured so that the coefficients of the objective function obtained by the calculation processing unit 11 can be transmitted to the Ising machine. In addition, the communication device 4 may be configured to be able to receive the value of the binary variable selected by the Ising machine.

[0419] The communication device 4 can connect the information processing device 2 to the network by wire or wirelessly. For example, the information processing device 2 can be connected to the computing device 7 via the network.

[0420] The communication device 4 may be provided inside the information processing device 2 or may be provided outside the information processing device 2. The configuration of the communication device 4 may be appropriately selected by those skilled in the art.

[0421] The input device 5 may be configured to receive measurement data of the sample, for example, fluorescence measurement data of the sample. In an embodiment, data acquired by a particle analysis device such as a flow cytometer, in particular, optical data may be configured to be received. The input device 5 may include an interface capable of receiving data.

[0422] In addition, the input device 5 may include an interface configured to be able to receive input of various types of data. For example, the input device 5 may be configured to receive a user's biomolecule selection operation. The input device 5 may include, for example, a mouse, a keyboard, and a touch panel as a device for receiving such an operation.

[0423] The output device 6 may be configured to be able to output various types of data. In an embodiment, the output device 6 includes a display device.

[0424] The display device may display, for example, an image for requesting the user to select a biomolecule included in the first data. For example, the display device may display a list of biomolecules included in the first data.

[0425] For example, the display device may display a list of the labeled phosphors included in the second data. In addition, the display device may display the fluorescence spectrum of the labeled phosphors.

[0426] In a preferred embodiment of the present technology, the display device may display a screen for requesting the user to select an optimization scheme for optimizing the combination of the biomolecule and the labeled fluorescent body. The screen may be a screen that enables one or more optimization schemes to be selected from a plurality of optimization schemes. An optimization scheme based on an objective function may be included as one of the optimization schemes that can be selected on the screen.

[0427] The display device can display the allocation information output by the allocation information output unit 12. More specifically, a combination of a biomolecule and a labeled fluorescent body can be displayed.

[0428] The computing device 7 includes a combinatorial optimization processing unit that generates combined information about the combination of the biomolecule and the labeled fluorescent body based on the coefficient of the objective function. The computing device including the combinatorial optimization processing unit is also called a combinatorial optimization solver. The computing device 7 including the combinatorial optimization processing unit can be configured as, for example, an Ising machine, and more specifically, can be configured as an Ising machine that optimizes binary quadratic variables. The computing device 7 can be configured as, for example, a semiconductor CMOS circuit, a digital circuit, a GPU, or an FPGA.

[0429] The Ising machine has a function of selecting one or more combinations of spin variables based on the coefficients of the objective function, and more specifically, has a function of selecting one or more combinations of spin variables corresponding to binary variables based on the coefficients of the objective function.

[0430] The computing device 7 may be an Ising machine for solving a multivariable combinatorial optimization problem. For example, the computing device 7 may use an Ising model (also known as quadratic unconstrained binary optimization (QUBO)) to replace the problem of the computational target, and perform the computation, the Ising model being a model indicating the behavior of the spin of a magnetic material. The computing device 7 may be an Ising machine that calculates a combination of variable values ​​that minimizes the value of an objective function by performing simulated annealing using a digital circuit, or may be an Ising machine that performs the same computation by performing quantum annealing using a superconducting circuit. Examples of the computing device 7 may include an Ising machine that reproduces the operation of the Ising model using a semiconductor CMOS circuit, an Ising machine configured by a digital circuit inspired by quantum phenomena, and an Ising machine that implements an analog branching algorithm through an FPGA. Through such a machine, the function of a combinatorial optimization processing unit may be realized.

[0431] (2-2) Example of the information processing system of the present technology configured as a particle analysis system

[0432] In an embodiment of the present technology, the information processing system of the present technology may be configured as a particle analysis system. The particle analysis system may be configured as a system that performs flow cytometry, for example, or may be configured as a system that performs cell sorting. In addition to the configuration of the above-mentioned information processing device and computing device, the particle analysis system may include, for example, a light irradiation unit that irradiates particles as analysis targets with light, a chip provided with a flow path through which particles (specifically, cells) flow, and a detection unit that detects light generated by light irradiation. For example, the light irradiation unit and the detection unit may be configured as a particle analysis device, and the particle analysis device may be combined with an information processing device and configured as a particle analysis system. The particle analysis device may be connected to the information processing device by wire or wirelessly, or may be connected via a network.

[0433] The light irradiation unit is configured to irradiate a predetermined position in the flow path of the chip with light. When the particle passes through the light irradiation position in the flow path, the particle is irradiated with light, and as a result, fluorescence is generated. That is, the light can act on the particle as an excitation light, specifically a fluorescent body that marks the particle. For example, the light irradiation unit may include a laser source that emits laser light (excitation light) to the particle. Those skilled in the art can appropriately select the configuration of the light irradiation unit.

[0434] The chip can be configured as, for example, a flow unit. The chip is equipped with a flow path. For example, the flow path structure provided in the chip is configured to form a flow (particularly, a laminar flow) in which particles flow substantially in a row. As the flow path structure provided in the chip, for example, a flow path structure known in the technical field of a flow cytometer or a cell sorter can be used.

[0435] The particles emitted from the chip can be sorted. For example, by vibrating the chip with a vibration element such as a piezoelectric vibration element, a droplet containing one particle can be generated from the discharge port. By charging the droplet with a charging unit, the traveling direction can be controlled and the particles can be sorted. Therefore, the particle analysis system can be configured as a system with a sorting function.

[0436] In addition, as a chip, a chip in which a sorting mechanism is provided can be used. Examples of such a chip may include a microchip for sorting particles described in JP2019-174192A. With this chip, particles in a sample liquid can be sorted without contact with the outside air, that is, a closed sorting operation becomes possible.

[0437] As described above, the particle analysis system may include a sorting unit for sorting particles. The sorting unit may sort the particles based on the fluorescence detection result of the detection unit.

[0438] The detection unit detects light generated by the light irradiation unit irradiating the particles with light. For example, the detection unit may be configured to detect light generated by irradiating particles flowing in the flow path of the chip with light. For example, the light detected by the detection unit is light containing fluorescence, and may be light containing fluorescence and light other than fluorescence. In addition to the detection of fluorescence, the detection unit 3 may be configured to further detect scattered light (e.g., one or more of forward scattered light, backscattered light, and side scattered light).

[0439] The detection unit includes at least one photodetector that detects light generated by the light irradiation unit irradiating particles with light. The number of photodetectors included in the detection unit can be, for example, 1 to 20, 1 to 15, or 1 to 10. Each photodetector includes one or more light receiving elements and has, for example, an array of light receiving elements. For example, each photodetector may include one or more photomultiplier tubes (PMTs) and / or photodiodes as light receiving elements, and specifically, one or more PMTs. The photodetector may include, for example, a PMT array, in which a plurality of PMTs are arranged in a one-dimensional direction. At least one photodetector can detect fluorescence and can be configured as a fluorescence detector.

[0440] The number of light receiving elements included in each photodetector (e.g., the number of PMTs) may be, for example, 2 or more, 5 or more, 8 or more, 10 or more, 15 or more, 20 or more, 22 or more, 24 or more, or 26 or more. The number of light receiving elements included in each photodetector (e.g., the number of PMTs) may be, for example, 50 or less, 45 or less, or 40 or less.

[0441] The detection unit may include a spectroscopic unit that disperses light. The spectroscopic unit may include, for example, a prism-type spectroscopic optical element. The spectroscopic unit may be included in each photodetector. The spectroscopic unit may be configured, for example, to disperse light (e.g., fluorescence) and allow light with a predetermined detection wavelength to reach a light receiving element (e.g., a PMT) assigned with the predetermined detection wavelength.

[0442] The detection unit may include a signal processing unit. The signal processing unit converts the electrical signal obtained by the photodetector into a digital signal. The signal processing unit may include, for example, an A / D converter as a device for performing the conversion. The optical signal detected by the photodetector may be converted into a digital signal by the signal processing unit and transmitted to the information processing device, for example, via an input device. The information processing device processes the digital signal into optical data and may be a target of an unmixing process. The optical data may include data on fluorescence intensity.

[0443] In the information processing system of the present technology configured as a particle analysis system, the processing unit of the information processing device processes optical data obtained by irradiating particles with light using a light irradiation unit. The processing may include unmixing processing. The optical data may be optical data including, for example, fluorescence data. More specifically, the optical data may be light intensity data, and the light intensity may be light intensity data of light including fluorescence.

[0444] The processing unit preferably performs unmixing processing using spectral reference data. Through this processing, the fluorescence intensity of each fluorescent dye can be obtained from the optical data. In addition, through this processing, the leakage of fluorescence that becomes a problem in the filter type flow cytometer of the related art is eliminated, and the separation performance of fluorescence is improved. According to the present technology, the performance degradation of the unmixing processing can be prevented by using an optimized combination of antibodies and labeled fluorescent dyes.

[0445] In this specification, spectral reference data (also referred to as SR data) is spectral data of fluorescence generated when each phosphor is irradiated with predetermined excitation light. For example, SR data can be obtained by detecting fluorescence generated by irradiating particles labeled with each phosphor individually with predetermined excitation light through a fluorescence detector.

[0446] The spectral reference data used in the unmixing process includes spectral data of fluorescence generated when the fluorescent body of the labeled particles is irradiated with predetermined excitation light. In order to obtain the spectral reference data to be used in the unmixing process, for example, first, a particle group as an analysis target is marked with each of the multiple fluorescent bodies of the labeled particle group, and a plurality of single-stained particle groups are obtained. Next, for each of the multiple single-stained particle groups, spectral data of fluorescence generated by light irradiation (specifically, laser irradiation) is obtained. The obtained spectral data is used as spectral reference data. Thus, the spectral data of the fluorescence of the particle group labeled by each fluorescent body of the multiple fluorescent bodies can be used as spectral reference data in the unmixing process. This makes it possible to improve the fluorescence separation performance.

[0447] The unmixing process can be performed according to, for example, a fluorescence intensity correction method or a fluorescence intensity calculation method described in JP 2011-232259A (PTL 1 described above).

[0448] The unmixing process can be performed by using, for example, a least square method (LSM) (more preferably, a weighted least square method (WLSM)). For example, by using the fluorescence intensity correction method described in JP5985140B, the unmixing process using the least square method can be performed. The fluorescence intensity correction method can be performed, for example, using the following formula 94 of WLSM.

[0449] [Math.94]

[0450]

[0451] In the above equation 94, x n represents the fluorescence intensity of the nth fluorescent dye, [S T ] represents the transposed matrix of the spectral reference, [L] represents the weight matrix, [S] represents the matrix of the spectral reference, y i represents the measured value in the i-th photodetector, λ i represents the weight of the i-th photodetector, max(y i , 0) represents a larger value when the detection value of the i-th detector is compared with zero, and offset' represents a value determined based on the detection value of each detector.

[0452] The fluorescence wavelength distribution of a fluorophore (e.g., a fluorescent dye) can be wide. Therefore, for example, a PMT for detecting fluorescence generated from a certain fluorophore can also detect fluorescence generated from another fluorophore. That is, the optical data acquired by each PMT can be data in which fluorescence data from multiple fluorophores are superimposed. Therefore, correction is required to separate the optical data into fluorescence data from each fluorophore. Unmixing processing is a scheme for correction, by which data in which fluorescence data from multiple fluorophores are superimposed are separated into fluorescence data from each fluorophore, and fluorescence data from each fluorophore can be obtained.

[0453] The processing unit generates output data using the fluorescence data obtained by the unmixing process. For example, the output data can be a two-dimensional graph of two desired fluorescent bodies in a plurality of fluorescent bodies for marking a particle group, but the present invention is not limited thereto. The vertical axis of the two-dimensional curve graph can be the fluorescence data (specifically, fluorescence intensity) of the fluorescence corresponding to one of the two fluorescent bodies, and the horizontal axis can be the fluorescence data (specifically, fluorescence intensity) of the fluorescence corresponding to another fluorescent body. The two-dimensional curve graph can be, for example, a density curve graph (dot graph), a contour map or a curve graph of both density and contour lines.

[0454] Particles can be, for example, particles with a size that can flow in a flow path arranged in the chip. In the present technology, those skilled in the art can appropriately select particles. In the present technology, particles can include, for example, biological microparticles such as cells, cell clusters, microorganisms and liposomes, and synthetic microparticles such as gel particles, beads, latex particles, polymer particles and industrial particles.

[0455] Biological microparticles (also referred to as biological particles) may include chromosomes, liposomes, mitochondria, cell organs (organelles), etc. that constitute various cells. Cells may include animal cells (e.g., blood cell lineage cells) and plant cells. Specifically, cells may be blood-based cells or tissue-based cells. Blood-based cells may be floating-based cells, such as T cells and B cells. Tissue-based cells may be, for example, sticky cultured cells or sticky cells separated from the tissue. Cell clusters may include, for example, spheroids and organoids. Microorganisms may include, for example, bacteria such as Escherichia coli, viruses such as tobacco mosaic virus, and fungi such as yeast. In addition, biological microparticles may also include biomacromolecules, such as nucleic acids, proteins, and complexes thereof. For example, these biomacromolecules may be extracted from cells or may be included in blood samples or other liquid samples. According to an embodiment of the present technology, the particles are biological particles, specifically cells.

[0456] The synthetic microparticles can be, for example, microparticles made of organic or inorganic polymer materials or metals. Organic polymer materials can include polystyrene, styrene / divinylbenzene, polymethyl methacrylate, etc. Inorganic polymer materials can include glass, silicon dioxide, magnetic materials, etc. Metals can include gold colloids, aluminum, etc. These synthetic microparticles can be, for example, gel particles or beads, and can more specifically be selected from oligonucleotides, peptides, proteins, and enzymes One or more combinations of gel particles or beads attached thereto.

[0457] The particle analysis system may set a particle group as an analysis target. In the particle analysis system, each particle included in the particle group may be irradiated with light by a light irradiation unit. Specifically, the particle group may be a biological particle group, and more specifically, a cell group.

[0458] The particle analysis system can be set to a particle group labeled with multiple fluorescent bodies as the analysis target. The fluorescent body of the labeled particle group can be a fluorescent dye. In addition, the fluorescent dye can bind to the particle (specifically a cell) via a molecule that specifically binds to the particle (e.g., an antibody, an aptamer, DNA or RNA, specifically an antibody).

[0459] When the light irradiation unit performs light irradiation, the particle group may be included in the sample liquid.The type of the sample liquid may be appropriately selected by those skilled in the art, and may be determined based on considerations such as the type of particles (cells).

[0460] For example, a plurality of fluorophores can be a plurality of dyes, specifically, a plurality of fluorescent dyes. The fluorophore can be, for example, a fluorophore (dye) known in the technical field of flow cytometry. The particle group is, for example, a particle group labeled with more than 5 fluorophores, more preferably a particle group labeled with more than 8, more than 10, more than 15 or more than 20 fluorophores. The particle group can be a particle group labeled with more than 22, more than 24, or more than 26 types of fluorophores. The particle group can be, for example, a particle group labeled with less than 50, less than 45, or less than 40 types of fluorophores.

[0461] (2-3) Example of processing in the information processing system of the present technology

[0462] In the following we will refer to Figures 5 to 8 An example of information processing in the information processing system 1 is described. Figure 5 and Figure 6 As described in (2-1) above. Figure 7 and Figure 8 The present invention is an example of a flow chart of an optimization process for a fluorescent dye-labeled antibody used in a particle analysis device (such as a flow cytometer). In the optimization process, the biomolecule may be an antigen or an antibody, and the fluorescent body may be a fluorescent dye.

[0463] exist Figure 7 In step S101, the processing unit 10 of the information processing device 2 starts the optimization process.

[0464] In step S102, the information processing device 2 obtains the first data input by the user. More specifically, the processing unit 10 receives an input of an antigen list from the user. The antigen list can be an antigen that is expected to be set as a marking target by the user relative to a sample that is an analysis target of the particle analysis device. For example, the information processing device 2 causes the display device included in the output device 6 to display a screen for requesting selection of an antigen as a marking target of the particle analysis device. The user operates the input device 5 (for example, a mouse or a keyboard) to select an antigen as a marking target of the particle analysis device on the screen. The information processing device 2 receives the selected antigen as an antigen list.

[0465] In step S103, the information processing device 2 presents to the user a list of fluorescent marker candidates that may correspond to each antigen included in the antigen list input in step S102 (more specifically, a list of fluorescent dye-labeled antibody candidates that specifically bind to each antigen). For example, the processing unit 10 causes the display device to display the list of fluorescent marker candidates. The processing unit 10 may obtain information about the fluorescent marker from the database 3 in order to present the list of fluorescent marker candidates, or may obtain information about the fluorescent marker from the storage unit 13 inside the information processing device 2. The fluorescent marker list may be displayed based on the information about the fluorescent marker.

[0466] In step S104, the information processing device 2 may obtain identification information of an algorithm for obtaining allocation information. The algorithm at least includes a combinatorial optimization algorithm executed by the combinatorial optimization processing unit.

[0467] More specifically, in step S104, the information processing device 2 presents the optimization process candidate list to the user. For example, the information processing device 2 presents the optimization process candidate list to the user, wherein, in addition to the optimization process using the objective function described in (2-1) above, one or more other optimization processes are listed as candidates. Thus, the information processing device 2 requests the user to select the type of optimization process to be performed. For example, one or more other optimization processes may include, for example, rule-based optimization processes or total inspection-based optimization processes. In rule-based optimization processes, good results are often obtained when the number of antigens included in the antigen list is small, but as the number increases, good results may not be obtained. In addition, since all combinations are checked, the total inspection takes time. Therefore, the optimization process has advantages and disadvantages. Therefore, in this step, the user is required to select the optimization process so that appropriate processing can be adopted according to the sample.

[0468] In step S105, the processing unit 10 performs the optimization process selected in step S104. Hereinafter, a case where the optimization process using the objective function described in (2-1) above is selected in step S104 will be described. Figure 8 A more detailed flow of the optimization process using the objective function will be described.

[0469] exist Figure 8 In step S201, the processing unit 10 starts the optimization process using the objective function.

[0470] In step S202, the processing unit 10 prepares first data listing information about antigens based on the antigen list input in step S102. The antigen list itself can be processed as the first data.

[0471] In a preferred embodiment, in step S202, the first data may also include information about the amount of each antigen included in the data (specifically, expression amount data), which enables optimization processing that takes into account the amount of each antigen and the fluorescence intensity of the fluorescent dye.

[0472] In step S203, the processing unit 10 prepares second data listing information of the labeled fluorescent body. In order to prepare the second data, the processing unit 10 may obtain information about the fluorescent marker bound to the antigen contained in the first data prepared in step S202 from the database 3. When the storage unit 13 of the information processing device 2 has information about the fluorescent marker, the processing unit 10 may obtain the information from the storage unit 13 instead of the database 3.

[0473] The information on the fluorescent markers prepared in step S203 includes the fluorescence spectrum of each fluorescent marker, that is, the fluorescence spectrum data of the fluorescent dye constituting each fluorescent marker. The use of the fluorescence spectrum enables optimization processing using an objective function.

[0474] The information about the fluorescent markers prepared in step S203 may include information about the antigens to which each fluorescent marker is combined and / or information about the fluorescence level of each fluorescent marker. The former may be, for example, the name or abbreviation of the antigen or the number (ID number) of the antigen. The latter may include, for example, the brightness level, or may include the brightness value itself. Brightness is also referred to as fluorescence intensity. For example, when a predetermined brightness range is divided into a plurality of ranges, the brightness level may be information about which of the plurality of ranges the brightness belongs to. The combination of information about the fluorescence level and information about the amount of each antigen makes it possible to use a dispersed objective function incorporating the above-mentioned fluorescence intensity, and suppress the variation of the fluorescence intensity of the selected fluorescent marker.

[0475] In step S204, the calculation processing unit 11 calculates the objective function regarding the combination of the biomolecule (specifically, antigen or antibody) and the labeled fluorescent body based on the first data prepared in step S202 and the second data prepared in step S203. The objective function is as described in the above-mentioned "(2-1) Configuration example of information processing system", and its description is also applicable to the present calculation processing.

[0476] For example, the calculation of the objective function can be performed by the calculation processing unit 11 executing a program for calculating the objective function (more specifically, a program for calculating the coefficient of the objective function).

[0477] The program may include, for example, a subroutine for calculating the linear independence between the fluorescence spectra of the labeled fluorophores assigned to each biomolecule (specifically, antigen or antibody) (hereinafter also referred to as a "linear independence subroutine"), and more specifically, may include a subroutine for calculating the objective function based on the determinant in equation 8 described in "(1) Description of the first embodiment" above or a determinant equivalent thereto. More preferably, the subroutine may be a subroutine for calculating a term that enhances the linear independence between the fluorescence spectra of the labeled fluorophores assigned to each biomolecule (specifically, antigen or antibody).

[0478] An example of a linear independence subroutine is described below in Equation 95.

[0479] [Math.95]

[0480]

[0481] In addition to the linear independence subroutine, the program may further include a subroutine regarding constraints on the number of fluorophores assigned to an antigen (hereinafter also referred to as a "fluorophore number constraint subroutine"), and preferably may include a subroutine for calculating the constraints on the assignment of a fluorophore to an antigen.

[0482] An example of the phosphor number limiting subroutine is described in Equation 96 below.

[0483] [Math.96]

[0484]

[0485] In addition to the linear independence subroutine, the program may further include a subroutine regarding a constraint condition regarding dispersion of fluorescence intensity (hereinafter also referred to as a "dispersion constraint subroutine").

[0486] An example of a dispersion constraint subroutine is described below in Equation 97.

[0487] [Math.97]

[0488]

[0489] The program may include a subroutine (hereinafter also referred to as a "coefficient calculation subroutine") for calculating coefficients of an objective function from:

[0490] The objective function obtained by executing the linear independence subroutine,

[0491] An objective function obtained by executing two subroutines including a linear independence subroutine and a phosphor number constraint subroutine, or

[0492] The objective function is obtained by executing three subroutines including a linear independence subroutine, a phosphor number constraint subroutine, and a dispersion constraint subroutine.

[0493] More specifically, the coefficient calculation subroutine may be a subroutine for calculating coefficients of an objective function of a binary variable of a quadratic form, specifically, Ising coefficients.

[0494] Examples of these objective functions are described in the above "(1) Description of the first embodiment", and may be, for example, an objective function based on equation 76 described in the above "(1) Description of the first embodiment". More preferably, the objective function may be an objective function obtained by converting equation 76 into a binary quadratic form. The objective function converted into a binary quadratic form is, for example, the objective function of equation 82 described in the above "(1) Description of the first embodiment".

[0495] In addition, the objective function can be an objective function based on equation 76 to which a term regarding the dispersion of fluorescence intensity has been added, specifically, it can be an objective function in which equation 76 to which a term regarding the dispersion of fluorescence intensity has been added is converted into a binary quadratic form, or it can be an objective function such as equation 93 above.

[0496] An example of the coefficient calculation subroutine is described in Equation 98 below.

[0497] [Math.98]

[0498]

[0499] In a particularly preferred embodiment of the present technology, the program includes a linear independence subroutine, a phosphor number constraint subroutine, a dispersion constraint subroutine, and a coefficient calculation subroutine. An example of a program including these four subroutines is as described in the following equation 99. Each subroutine included in the following equation 99 can be as shown in equations 95 to 98 above.

[0500] [Math.99]

[0501]

[0502] In step S205, the information processing device 2 transmits the objective function calculated in step S204 to the calculation device 7 through the communication device 4. The information processing device 2 may transmit the coefficients of the objective function obtained by the calculation process to the calculation device 7. The calculation device 7 may be an Ising machine specifically used for optimizing binary quadratic variables.

[0503] In step S206 , the computing device 7 receives the target function.

[0504] In step S207, the combination optimization processing unit included in the computing device 7 generates combination information about the combination of the antigen (or antibody) and the labeled fluorescent body based on the objective function, specifically based on the coefficient of the objective function. The combination information includes, for example, binary variables selected based on the coefficient of the objective function. More specifically, the computing device 7 uses the coefficient of the objective function to perform, for example, spin optimization to obtain a combination of optimized spins. Thus, the combination information generated by the computing device 7 includes spin variables, and is preferably an optimized combination of spin variables. The combination information may include one or more combinations of spin variables.

[0505] In step S208 , the calculation device 7 transmits the generated combination information to the information processing device 2 .

[0506] In step S209, the information processing device 2 receives the combination information generated by the calculation device 7 via the communication device 4. For example, the information processing device 2 receives the value of the binary variable selected by the calculation device 7 as an Ising machine.

[0507] In step S210, the allocation information output unit 12 generates allocation information of the labeled fluorescent body to each antigen based on the combination information. For example, the allocation information output unit 12 obtains data related to the combination of the biomolecule (having an antigen or an antibody) and the labeled fluorescent body according to the spin variable obtained by the calculation device 7 as an Ising machine. More specifically, the information processing device 2 obtains data related to the combination of the biomolecule (specifically, an antigen or an antibody) and the labeled fluorescent body by converting the spin variable into a binary variable. The allocation information may include the data obtained in this way.

[0508] In step S211 , the processing unit 10 ends the optimization process.

[0509] In step S106, the allocation information output unit 12 outputs the allocation information generated in step S105 (specifically, step S210). For example, the allocation information output unit 12 may cause the output device 6 to output the allocation information, more specifically, may cause the display device as the output device 6 to display the allocation information, or may cause the printing device as the output device 6 to print the allocation information.

[0510] In step S107, the information processing device 2 receives correction of the output distribution information from the user. That is, the user can correct a part of the combination of the output biomolecule (specifically, antigen or antibody) and the labeled fluorescent body. Step S107 may be omitted.

[0511] In step S108 , the processing unit 10 ends the optimization process.

[0512] Through the above optimization process, the optimal combination of fluorescent markers can be automatically presented to the user. In addition, the performance degradation of the demixing process can be prevented by the combination of fluorescent markers.

[0513] 2. Second Embodiment (Information Processing Method)

[0514] The present technology also provides an information processing method. The information processing method includes: a calculation processing step of calculating an objective function related to a combination of a biomolecule and a labeled fluorescent body based on first data listing information of a biomolecule as a labeled target in a sample and second data listing information of a fluorescent signal of a labeled fluorescent body; and an output step of outputting information about the allocation of the labeled fluorescent body to each biomolecule, the information being generated based on the combination information acquired from the combination optimization processing unit based on the coefficient of the objective function. The information processing method can be performed for the allocation of the labeled fluorescent body to each biomolecule.

[0515] For example, as described in "(2-2) Example of processing in information processing system" of "1. First embodiment" above, the information processing method according to the present technology can be performed. Step S204 in the processing corresponds to a calculation processing step. In addition, step S106 in the processing corresponds to an allocation information output step. The information processing method of the present technology may further include one or more steps other than step S204 and step S106.

[0516] 3. Third Implementation Method (Procedure)

[0517] The present technology also provides a program for causing an information processing device to execute an information processing method (specifically, an optimization processing method) according to the present technology, and more specifically, a program for causing an information processing device to execute the calculation processing steps and output steps described in the above-mentioned "2. Second embodiment (information processing method)". For example, the program can be recorded on any recording medium. The recording medium can be, for example, a micro SD memory card, an SD memory card, a flash memory, a CD, a DVD, or a BD.

[0518] 4. Fourth Embodiment (Information Processing Device)

[0519] The present technology also provides an information processing device including the calculation processing unit and the distribution information output unit described in "1. First embodiment (information processing system)" above. By using the information processing device in combination with the combination optimization processing unit, a better combination of biomolecules and labeled fluorescent bodies can be automatically selected.

[0520] 5. Fifth Embodiment (Calculation Device)

[0521] As described in (1) of the above-mentioned "1. First embodiment (information processing system)", the problem of optimal selection of fluorescent markers for a flow cytometer can be formulated as an optimal selection problem of basis vector candidates. In the present technology, combinatorial optimization processing is performed using the coefficients of the objective function formulated in this manner. As described in "1. First embodiment (information processing system)", the combinatorial optimization processing can be performed by a computing device (such as an Ising machine). That is, the present technology also provides a computing device including a combinatorial optimization processing unit, which is configured to perform combinatorial optimization processing using the coefficients of the objective function, which is formulated as a constrained optimal selection problem of basis vectors from basis vector candidates.

[0522] The computing device including the combinatorial optimization processing unit can also be used in communication technology. The computing device can be used, for example, to associate information or transmission lines with waveform patterns. This will be described in more detail in (1) to (3) below.

[0523] (1) Optimization of communication modulation scheme

[0524] The communication path modulation scheme (specifically, the digital signal modulation scheme) is a scheme for assigning information (bits or bytes) to the waveform pattern of the communication carrier. Examples of currently well-known modulation schemes include amplitude modulation, frequency modulation, phase modulation, pulse modulation, or a combination of these modulation schemes.

[0525] For example, amplitude shift keying modulation (ASK) is a modulation scheme for changing the amplitude of a carrier wave for each time unit (clock) according to the information in the time unit. In this modulation scheme, for example, when the information bit is 0, the amplitude is set to 0, and when the information bit is 1, the amplitude is fixed to 1. Phase shift keying (PSK) is a modulation scheme for changing the phase of a carrier wave according to information. In this modulation scheme, for example, when the information bit is 0, the phase is set to 0, and when the information bit is 1, the phase is fixed to 180 degrees. Quadrature phase amplitude shift keying (QAM) is a modulation scheme that is a combination of amplitude keying and phase shift keying. In this modulation scheme, a carrier wave with a predetermined phase and amplitude is assigned to information.

[0526] Therefore, associating information with a waveform pattern is called a modulation scheme, specifically, a digital modulation scheme. In order to improve the performance of the digital modulation scheme, preferably, the difference of the waveform patterns corresponding to different information is large.

[0527] (2) Optimization of communication multiplexing scheme

[0528] Communication multiplexing (specifically, digital communication multiplexing schemes) is a technique for transmitting multiple information sequences simultaneously. Examples of well-known multiplexing schemes include time division multiplexing, frequency division multiplexing (frequency DM), and code division multiplexing (code DM).

[0529] Code division multiplexing is also called spread spectrum, and includes frequency hopping and direct spread spectrum (direct sequence). In direct spread spectrum, the information (bit sequence) is further modulated with a pulse modulation pattern unique to the transmission line called a spread code, and then the usual (described above) carrier modulation is performed so that the final modulated signal is generated on the transmission line. Here, interference of information in different transmission lines is prevented by changing the pulse modulation pattern of the spread code according to the transmission line. In order to improve the performance of multiplexing, preferably, the waveform patterns corresponding to different spread codes are different, and the combination of waveform patterns of multiple spread codes does not match the waveform pattern (combination) of other spread codes.

[0530] (3) Combination optimization of waveform patterns

[0531] The combinatorial optimization process in the present technology can be used for optimization of the communication modulation scheme or multiplexing scheme described in (1) and (2) above, and more specifically, can be used for association of information or transmission lines with waveform patterns.

[0532] Hereinafter, optimization of the correlation between information and the waveform pattern will be described.

[0533] For example, suppose there are multiple patterns that a waveform pattern can take (e.g., L types), and some of these (e.g., M types) are selected and assigned to information. In this case, the M waveform patterns are difficult to distinguish when the M waveform patterns are the same, and are difficult to distinguish even when the combination of the M waveform patterns is the same. Therefore, preferably, the M types of waveform patterns are linearly independent. However, the number of combinations of selecting M waveform patterns from L waveform patterns increases sharply as the number of L and M is large, so it is expected that the evaluation of the linear independence of each combination will take a long time.

[0534] According to the present technology, by implementing the approximate equation of the determinant calculated from the L-dimensional square matrix generated by the L types of waveform patterns on the Ising machine, it is easy to search for M types of waveform patterns that satisfy linear independence. Hereinafter, a process for implementation will be described.

[0535] Now, suppose there are L waveform pattern candidates, i.e. basis vector candidates, and the i-th basis vector is represented by u i Indicates whether to select the i-th basis vector u i The value of b i ∈{0, 1}. In this case, the determinant for evaluating the linear independence of the selected basis vectors is as described in (1) of “1. First embodiment (information processing system)” above, and is approximated as the following Equation 100.

[0536] [Math.100]

[0537]

[0538] The function of Equation 99 is a sign-reversed version of the approximation equation for the determinant (Equation 74), and preferably has a smaller value. The constraints on how the basis vectors are chosen are imposed on the binary value b i Since the number of selected basis vectors is M (M<L), the constraint of the following equation 101 is imposed.

[0539] [Mathematical formula.101]

[0540]

[0541] By using Equation 100 and Equation 101 above, the optimal selection of basis vectors is formulated as Equation 102 below.

[0542] [Math.102]

[0543]

[0544] The conversion between binary variables and spin variables or the reduction of the dimension of spin variables in the objective function is as described in (1) of the above-mentioned “1. First embodiment (information processing system)”.

[0545] As described above, optimization processing of association of information and waveform patterns can be performed based on the objective function of Equation 102, and more specifically, based on the coefficients of the objective function obtained by converting the objective function of Equation 102 into a binary quadratic form.

[0546] Similarly, this can be used to associate with the waveform pattern of the transmission line based on the objective function of Equation 102, and more specifically, based on the coefficients of the objective function obtained by converting the objective function of Equation 102 into a binary quadratic form.

[0547] That is, in an embodiment of the present technology, a combined optimization processing unit included in a computing device performs combined optimization processing of waveform patterns based on the coefficients of a combined optimization objective function for assigning waveform patterns to each piece of communication information, based on first data including communication information and second data including waveform patterns of communication carriers. For example, this makes it possible to increase the difference in waveform patterns corresponding to different information and improve the performance of a digital modulation scheme.

[0548] In another embodiment of the present technology, the combinatorial optimization processing unit performs combinatorial optimization processing of waveform patterns based on the coefficients of the combinatorial optimization objective function, and the combinatorial optimization objective function is used to assign the waveform pattern to each extension code based on the first data including the extension code and the second data including the waveform pattern of the extension code. This makes it possible to make the waveform patterns corresponding to the respective extension codes different. In addition, it is possible to prevent the combination of waveform patterns of multiple extension codes from matching the waveform patterns of other extension codes or the combination of waveform patterns. Therefore, using the present technology, the performance of multiplexing can be improved.

[0549] The present technology may also have the following configurations.

[0550] [1] An information processing system, comprising an information processing device, the information processing device comprising: a calculation processing unit configured to calculate an objective function regarding a combination of a biological molecule and a labeled fluorescent body based on first data listing information regarding a biological molecule that is a labeled target in a sample and second data listing information regarding a labeled fluorescent body; and an allocation information output unit configured to output information regarding allocation of the labeled fluorescent body to each biological molecule, the information being generated based on combination information obtained from the combination optimization processing unit based on coefficients of the objective function.

[0551] [2] An information processing system according to [1], wherein the information processing device obtains first data input by a user.

[0552] [3] The information processing system according to [1] or [2], further comprising: a database configured to store the second data,

[0553] The information processing device obtains the second data by referring to the database based on the first data.

[0554] [4] The information processing system according to any one of [1] to [3],

[0555] wherein the information processing device obtains identification information of an algorithm for obtaining the allocation information, and

[0556] The algorithm at least includes a combinatorial optimization algorithm executed by a combinatorial optimization processing unit.

[0557] [5] An information processing system according to any one of [1] to [4], wherein the objective function is an objective function formulated as a constrained optimal selection problem of basis vector candidates, and

[0558] The basis vector candidates selected in the optimal selection problem correspond to the labeled fluorophores that label the biomolecules that are the labeling targets.

[0559] [6] An information processing system according to any one of [1] to [5], wherein the objective function is an objective function of a binary variable of a quadratic type.

[0560] [7] The information processing system according to [6], wherein the objective function is configured to indicate linear independence between fluorescence spectra of the labeling fluorescent bodies assigned to the respective biomolecules.

[0561] [8] The information processing system according to any one of [1] to [7], wherein the objective function is configured to include a constraint condition regarding the number of fluorescent bodies assigned to one biomolecule.

[0562] [9] An information processing system according to [8], wherein the objective function is configured to include a constraint that a fluorescent body is assigned to a biological molecule.

[0563]

[10] An information processing system according to any one of [1] to [9], wherein the objective function is configured to include a constraint condition regarding fluorescence intensity regarding a combination of a biomolecule and a fluorophore assigned to the biomolecule.

[0564]

[11] The information processing system according to

[10] , wherein the fluorescence intensity is based on the amount of the biomolecule and the fluorescence intensity of the fluorescent body assigned to the biomolecule.

[0565]

[12] An information processing system according to any one of [1] to

[11] , wherein the objective function is configured to include dispersion of fluorescence intensity information.

[0566]

[13] An information processing system according to any one of [1] to

[12] , wherein the information processing device includes a communication device, which is capable of transmitting coefficients of an objective function obtained through computational processing to an Ising machine that specifically optimizes binary quadratic variables, and receiving values ​​of binary variables selected by the Ising machine.

[0567]

[14] An information processing system according to

[13] , wherein the Ising machine has a function of selecting one or more combinations of spin variables corresponding to binary variables based on coefficients of an objective function.

[0568]

[15] The information processing system according to

[13] or

[14] , wherein the information processing device obtains data on the combination of the biomolecule and the fluorescent substance based on the value of the spin variable obtained by the Ising machine.

[0569]

[16] The information processing system according to

[15] , wherein the information processing device acquires data related to the combination of the biomolecule and the fluorescent body by converting the spin variable into a binary variable.

[0570]

[17] The information processing system according to any one of [1] to

[16] , wherein the biomolecule is an antigen or an antibody.

[0571]

[18] An information processing method comprising: a calculation processing step of calculating a target function related to a combination of a biomolecule and a labeled fluorescent body based on first data listing information about a biomolecule as a labeled target in a sample and second data listing information about a fluorescent signal of the labeled fluorescent body; and

[0572] The output step outputs information on the allocation of the labeled fluorescent body to each biomolecule, the information being generated based on the combination information acquired from the combinatorial optimization processing unit based on the coefficient of the objective function.

[0573]

[19] A program for causing an information processing device to execute: a calculation processing step of calculating an objective function regarding a combination of a biomolecule and a labeled fluorescent body based on first data listing information regarding a biomolecule that is a labeled target in a sample and second data listing information regarding a fluorescent signal of the labeled fluorescent body; and

[0574] The output step outputs information on the allocation of the labeled fluorescent body to each biomolecule, the information being generated based on the combination information acquired from the combinatorial optimization processing unit based on the coefficient of the objective function.

[0575]

[20] An information processing device comprising: a calculation processing unit configured to calculate an objective function related to a combination of a biomolecule and a labeled fluorescent body based on first data listing information about a biomolecule as a labeled target in a sample and second data listing information about a fluorescent signal of the labeled fluorescent body; and

[0576] The allocation information output unit is configured to output allocation information of the labeled fluorescent substance to each biomolecule based on the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function.

[0577]

[21] A computing device includes a combinatorial optimization processing unit configured to perform combinatorial optimization processing using coefficients of an objective function, the objective function being formulated as a constrained optimal selection problem of basis vectors from basis vector candidates.

[0578]

[22] A computing device according to

[21] , wherein a combinatorial optimization processing unit performs combinatorial optimization processing of waveform patterns based on first data including communication information and second data including a waveform pattern of a communication carrier, and based on coefficients of a combinatorial optimization objective function for assigning the waveform pattern to each piece of communication information.

[0579]

[23] A computing device according to

[21] , wherein a combinatorial optimization processing unit performs combinatorial optimization processing of waveform patterns based on first data including extended codes and second data including waveform patterns of extended codes, and based on coefficients of a combinatorial optimization objective function for assigning waveform patterns to each extended code.

[0580] [Reference Numbers List]

[0581] 1 Information Processing System

[0582] 2Information Processing Device

[0583] 3 Database

[0584] 4 communication devices

[0585] 5 Input Devices

[0586] 6 Output Devices

[0587] 7. Computing Device

[0588] 10 processing units

[0589] 11Computing Processing Unit

[0590] 12 Allocation information output unit.

Claims

1. An information processing system comprising an information processing device, the information processing device comprising: a calculation processing unit configured to calculate an objective function related to a combination of the biomolecule and the labeling fluorescent substance based on first data listing information related to the biomolecule as a labeling target in the sample and second data listing information related to the labeling fluorescent substance; as well as an allocation information output unit configured to output allocation information of the labeled fluorescent body to each biomolecule according to the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function; The combination information includes a spin variable, and the allocation information output unit obtains the allocation information of the labeled fluorescent substance to each biological molecule by converting the spin variable into a binary variable.

2. The information processing system according to claim 1, wherein: The information processing device acquires the first data input by a user.

3. The information processing system according to claim 1, further comprising: a database configured to store the second data, The information processing device acquires the second data by referring to the database based on the first data.

4. The information processing system according to claim 1, in, The information processing device acquires identification information of an algorithm used to acquire the allocation information, and The algorithm at least includes a combinatorial optimization algorithm executed by the combinatorial optimization processing unit.

5. The information processing system according to claim 1, in, The objective function is an objective function formulated as a constrained optimal selection problem of basis vector candidates, and The basis vector candidates selected in the optimal selection problem correspond to the labeling fluorescent substance that labels the biomolecule, which is the labeling target.

6. The information processing system according to claim 1, wherein: The objective function is an objective function of a binary variable of quadratic form.

7. The information processing system according to claim 6, wherein: The objective function is configured to indicate linear independence between fluorescence spectra of the labeling fluorophores assigned to respective biomolecules.

8. The information processing system according to claim 1, wherein: The objective function is configured to include a constraint regarding the number of the fluorophores assigned to one biomolecule.

9. The information processing system according to claim 8, wherein: The objective function is configured to include a constraint that a fluorophore is assigned to a biomolecule.

10. The information processing system according to claim 1, wherein: The objective function is configured to include constraints regarding fluorescence intensity regarding a combination of the biomolecule and the fluorophore assigned to the biomolecule.

11. The information processing system according to claim 10, wherein: The fluorescence intensity is based on the amount of the biomolecule and the fluorescence intensity of the fluorophore assigned to the biomolecule.

12. The information processing system according to claim 1, wherein: The objective function is configured to include a dispersion of fluorescence intensity information.

13. The information processing system according to claim 1, wherein: The information processing device includes a communication device capable of transmitting coefficients of the objective function obtained through calculation processing to an Ising machine that specifically optimizes binary quadratic variables, and receiving values ​​of binary variables selected by the Ising machine.

14. The information processing system according to claim 13, wherein: The Ising machine has a function of selecting one or more combinations of spin variables corresponding to binary variables based on coefficients of the objective function.

15. The information processing system according to claim 13, wherein: The information processing device acquires data on a combination of the biomolecule and the fluorescent substance based on the value of the spin variable acquired by the Ising machine.

16. The information processing system according to claim 15, wherein: The information processing device acquires data related to a combination of the biomolecule and the fluorescent body by converting the spin variable into a binary variable.

17. The information processing system according to claim 1, wherein: The biomolecule is an antigen or an antibody.

18. An information processing method, comprising: a calculation processing step of calculating a target function related to a combination of the biomolecule and the labeled fluorescent body based on first data listing information related to the biomolecule as a labeled target in the sample and second data listing information related to a fluorescent signal of the labeled fluorescent body; as well as an output step of outputting information on the allocation of the labeled fluorescent body to each biomolecule based on the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function; The combination information includes spin variables, and the distribution information of the labeled fluorescent substance to each biological molecule is obtained by converting the spin variables into binary variables.

19. A computer-readable storage medium storing a program, wherein when the program is executed, the program is used to cause an information processing device to perform: a calculation processing step of calculating an objective function related to a combination of the biomolecule and the labeled fluorescent body based on first data listing information related to the biomolecule as a labeled target in the sample and second data listing information related to a fluorescent signal of the labeled fluorescent body; and an output step of outputting information on the allocation of the labeled fluorescent body to each biomolecule based on the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function; in, The combined information includes a spin variable, and by converting the spin variable into a binary variable, the distribution information of the labeled fluorescent substance to each biomolecule is acquired.

20. An information processing device, comprising: a calculation processing unit that calculates an objective function related to a combination of the biomolecule and the labeled fluorescent body based on first data listing information related to the biomolecule as a labeled target in the sample and second data listing information related to a fluorescent signal of the labeled fluorescent body; as well as an allocation information output unit, which outputs allocation information of the labeled fluorescent body to each biomolecule according to the combination information obtained from the combination optimization processing unit based on the coefficient of the objective function; The combination information includes a spin variable, and the allocation information output unit obtains the allocation information of the labeled fluorescent substance to each biological molecule by converting the spin variable into a binary variable.

21. A computing device, comprising a combinatorial optimization processing unit, the combinatorial optimization processing unit being configured to perform combinatorial optimization processing using coefficients of an objective function to output combined information, the objective function being formulated as a constrained optimal selection problem of basis vectors from basis vector candidates; in, calculating the objective function related to a combination of the biomolecule and the labeling fluorescent substance based on first data listing information related to the biomolecule as a labeling target in the sample and second data listing information related to the labeling fluorescent substance; The combination information includes spin variables, and the distribution information of the labeled fluorescent substance to each biological molecule is obtained by converting the spin variables into binary variables.

22. The computing device of claim 21, wherein: The combinatorial optimization processing unit performs the combinatorial optimization processing of the waveform pattern based on the coefficients of the combinatorial optimization objective function for assigning the waveform pattern to each piece of communication information, based on first data including communication information and second data including the waveform pattern of the communication carrier.

23. The computing device of claim 21, wherein: The combinatorial optimization processing unit performs the combinatorial optimization processing of the waveform pattern based on the coefficients of a combinatorial optimization objective function for assigning the waveform pattern to each spreading code, based on first data including a spreading code and second data including a waveform pattern of the spreading code.

Images