Stable conformation search system, method, and medium storing stable conformation search program

Abstract

Stable conformation search systems, methods, and media storing stable conformation search programs are disclosed. The system executes the following steps: obtaining N (N: gt; an integer of 1) main chain particles and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space; calculating the coordinates of the (i + 1)-th main chain particle by using the coordinates of the i-th main chain particle and a first state variable represented by the relative coordinates between the i-th main chain particle and the (i + 1)-th main chain particle; calculating the coordinates of an (i + 1)-th side chain particle corresponding to the (i + 1)-th main chain particle by using the coordinates of the (i + 1)-th main chain particle and a second state variable represented by the relative coordinates between the (i + 1)-th main chain particle and the (i + 1)-th side chain particle; calculating a value of energy of the model when any one of the first state variable and the second state variable changes; a first state variable and a second state variable having a local minimum are identified.

Description

Technical Field

[0001] The embodiments discussed herein relate to a stable conformation search system, a stable conformation search method, and a non-transitory computer-readable recording medium storing a stable conformation search program. Background Art

[0002] In recent years, in the field of drug discovery, middle molecules (molecular weight of 500 to 3000) with few side effects have attracted much attention, and development of search methods for searching for stable conformations of middle molecules is underway.

[0003] As an example, the applicant has studied a search method for searching for stable conformations of a coarse-grained model by using the interaction potential between coarse-grained particles. For example, the applicant has studied a search method for reducing the amount of calculation for searching for stable conformations of a coarse-grained model.

[0004] Japanese Patent Application Publication No. 2021-82165 is disclosed as a related art. Summary of the invention

[0005] Technical issues

[0006] However, the above-mentioned search method in the example is intended to search for stable conformations of main chain particles, and the search is performed without considering the characteristics of side chain particles.

[0007] According to one aspect, the object is to further improve the search accuracy when searching for stable conformations of the model.

[0008] Solution to the problem

[0009] According to one aspect of the embodiment, there is provided a stable conformation search system, comprising: a first computing unit, configured to: obtain a model in which N (N is an integer of 2 or greater) main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, and calculate the coordinates of the (i+1)th main chain particle in the sequence in the lattice space by using the coordinates of the i-th main chain particle in the sequence in the lattice space and a first state variable, the first state variable being represented by the relative coordinates between the i-th main chain particle and the (i+1)th main chain particle in the lattice space; and a second computing unit, configured to: obtain a model in which N (N is an integer of 2 or greater) main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, and calculate the coordinates of the (i+1)th main chain particle in the sequence in the lattice space by using the coordinates of the i-th main chain particle in the sequence in the lattice space and a first state variable, the first state variable being represented by the relative coordinates between the i-th main chain particle and the (i+1)th main chain particle in the lattice space; and A calculation unit configured to calculate the coordinates of an (i+1)th side chain particle corresponding to the (i+1)th main chain particle in the lattice space by using the coordinates of the (i+1)th main chain particle in the lattice space and a second state variable, the second state variable being represented by the relative coordinates between the (i+1)th main chain particle and the (i+1)th side chain particle; a calculation unit configured to calculate the value of energy of the model when any one of the first state variable and the second state variable changes; and an identification unit configured to identify the first state variable and the second state variable whose energy values ​​are local minimum values.

[0010] Advantageous Effects of the Invention

[0011] The search accuracy when searching for stable conformations of the model can be further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a diagram showing an application example of the stable conformation search system;

[0013] Figure 2 is a diagram showing an overview of a stable conformation search process for a two-bead model;

[0014] Figure 3A and Figure 3B is a diagram showing a representation method using relative coordinates;

[0015] Figure 4A is a diagram showing a specific example of a first state variable;

[0016] Figure 4B is a diagram showing a specific example of the second state variable;

[0017] Figure 5 is a diagram showing details of the energy calculation process;

[0018] Figure 6 is a first diagram showing details of terms in an energy calculation process;

[0019] Figure 7 is a second diagram showing details of terms in the energy calculation process;

[0020] Figure 8 is a third diagram showing details of terms in the energy calculation process;

[0021] Fig. 9 is a diagram showing the magnitude of the attractive force between amino acid residues;

[0022] Fig.10 is a diagram showing an example of a system configuration of a stable constellation search system and a functional configuration of a terminal device;

[0023] Fig.11 is a diagram showing an example of a system configuration of a stable conformation search system and a functional configuration of an Ising device;

[0024] Fig.12 is a diagram showing a specific example of processing performed by an Ising device;

[0025] Fig.13 is a diagram showing an example of a hardware configuration of a terminal device;

[0026] Fig.14 is a flow chart showing the sequence of the stable conformation search process;

[0027] Fig.15 is a flowchart showing details of the search process;

[0028] Fig.16 is a flowchart showing details of the determination process; and

[0029] Fig.17 is a diagram showing an example of the effect of reducing the amount of calculation. DETAILED DESCRIPTION

[0030] The embodiments will be described below with reference to the accompanying drawings. In this specification and the drawings, constituent elements having substantially the same functional configuration will be denoted by the same reference numerals, and redundant descriptions thereof will be omitted.

[0031] [First embodiment]

[0032] <Application examples of the stable conformation search system>

[0033] First, an application example of the stable conformation search system according to the first embodiment will be described. Figure 1 is a diagram showing an application example of the stable conformation search system.

[0034] In the field of drug discovery, in general, the use of small molecules has become the mainstream so far. However, small molecules are likely to bind to proteins other than the target protein and have the potential to inhibit functions other than the target function. For this reason, middle molecules with small side effects have been attracting attention in recent years.

[0035] In the case of middle molecules as drug candidates, searching for stable conformations involves a lot of calculations. Therefore, the applicant has been developing a search method that first narrows the scope to a specific middle molecule (e.g., a cyclic peptide) and then searches for stable conformations of the narrowed middle molecule in the following two stages.

[0036] The first stage is a stage of generating a two bead model and searching for a stable conformation of the two bead model. The two bead model is a model in which one amino acid residue is represented by two types of particles (coarse-grained particles). In the following, a plurality of atoms (atomic groups) in the atomic groups included in the amino acid residue that are used to form an amino bond are referred to as main chain particles. A plurality of atoms (atomic groups) in the side chain corresponding to the main chain particle other than the atoms in the atomic groups included in the amino acid residue that are used to form an amino bond are referred to as side chain particles. The stable conformation search system according to the first embodiment is applied to Figure 1 The first stage is shown in the dashed box and the system searches for stable conformations of the double-bead model.

[0037] The second stage is a stage of searching for stable conformations of all atoms based on the double-bead model with the searched stable conformations, and further verifying the efficacy of the drug based on the all-atom model with the searched stable conformations.

[0038] If the desired drug efficacy is not obtained in the second stage, the method returns to the first stage and further searches for a stable conformation of the double bead model. The above cycle of these stages is repeated until the desired drug efficacy is obtained. In the case of obtaining the desired drug efficacy, for example, the pharmacist synthesizes a mid-molecule drug based on an all-atom model with a stable conformation at that time, and performs activity evaluation on it, etc. (so-called wet method experiment).

[0039] The stable conformation search system according to the first embodiment, which will be described in detail below, is a system applied to the first stage in a series of sequences for mid-molecule drug discovery as described above.

[0040] <Overview of the stable conformation search process for the double-bead model>

[0041] Next, an overview of the stable conformation search process for the double-bead model performed by the stable conformation search system according to the first embodiment will be described. Figure 2 2 is a diagram showing an overview of the stable conformation search process for the double-bead model.

[0042] The stable conformation search system according to the first embodiment generates a two-bead model as amino acid molecules linked in a linear chain by sequentially arranging as follows in a space (lattice space) such as a face-centered cubic lattice (FCC): [a] backbone portion ( Figure 2201 to 206 in the figure), each of which is centered on the α carbon of the amino acid to form the main chain particle of the peptide; and [b] side chain particles ( Figure 2 211 to 216 in the figure). The stable conformation search system searches for stable conformations of the generated double-bead model according to a combinatorial optimization problem by using an Ising device.

[0043] like Figure 2 As shown, the stable conformation search system according to the first embodiment, when searching for a stable conformation of the double-bead model, iterates the state transition of the double-bead model by the following operations: [a] in the space of the face-centered cubic lattice, the position of each main chain particle is moved, and [b] in the space of the face-centered cubic lattice, the position of the corresponding side chain particle is moved to a position corresponding to the position to which the main chain particle is moved. In the process of iterative state transition, the stable conformation search system according to the first embodiment determines whether each transition (also called change) is acceptable based on the difference between the energy value before the transition and the energy value after the transition, and identifies the minimum energy value (to be precise, the local minimum value of the energy value) among the energy values ​​after the transition (after the change) that are determined to be acceptable. In this way, the stable conformation search system according to the first embodiment searches for a double-bead model of a stable conformation.

[0044] For example, according to the first embodiment, the stable conformation search system identifies the minimum energy value without falling into a local solution by using a Markov chain Monte Carlo (MCMC) method to determine that a transition is acceptable if the transition satisfies predetermined conditions, such as [a] if the energy value after the transition is less than the energy value before the transition, then the transition is determined to be acceptable, and [b] if the difference between the energy value after the transition and the energy value before the transition is less than a predetermined thermal noise, then the transition is determined to be acceptable even if the energy value after the transition is greater than the energy value before the transition.

[0045] The stable conformation search system according to the first embodiment effectively identifies the minimum energy value by iterating the state transition of the double-bead model using a parallel annealing method (PT).

[0046] The stable conformation search system according to the first embodiment identifies the minimum energy value with a small amount of calculation by the following operations:

[0047] [a] calculating the coordinates of the (i+1)th main chain particle in the sequence among a plurality of main chain particles arranged in sequence in the space of a face-centered cubic lattice by using a first state variable represented by relative coordinates to the coordinates of the i-th main chain particle in the sequence,

[0048] [b] calculating the coordinates of the (i+1)th side chain particle by using the second state variable represented by the relative coordinates to the coordinates of the (i+1)th main chain particle in the sequence, and

[0049] [c] Whenever any one of the first state variable and the second state variable changes, the energy value of the double-bead model is calculated (the details of the energy value calculation method will be described later).

[0050] The “first state variable” and the “second state variable” herein are variables for specifying the state of the double bead model (the coordinates of each main chain particle and the coordinates of each side chain particle of the double bead model).

[0051] exist Figure 2 , reference numerals 201 ′ to 206 ′ and 211 ′ to 216 ′ schematically show stable conformations of a two-bead model (main chain particles and side chain particles) searched by the stable conformation search system according to the first embodiment by identifying minimum energy values.

[0052] <Advantages of Relative Coordinates for Representing the First State Variable and the Second State Variable>

[0053] As described above, the stable conformation search system according to the first embodiment calculates the coordinates of the (i+1)th main chain particle by using the first state variable represented by the relative coordinates to the coordinates of the i-th main chain particle. The stable conformation search system according to the first embodiment calculates the coordinates of the (i+1)th side chain particle by using the second state variable represented by the relative coordinates to the coordinates of the (i+1)th main chain particle. The advantages of representing the state variables by relative coordinates will be described below.

[0054] Figure 3A and Figure 3B is a diagram showing a method of expressing using relative coordinates. Figure 3A and Figure 3B In the example shown, the space of the face-centered cubic lattice is represented by a two-dimensional plane for simplicity of description.

[0055] Figure 3A FIG. 4 shows the case where the first state variable is represented by relative coordinates. Figure 3A As shown, when determining the position of the first backbone particle included in the double-bead model, the number of candidate positions of the second backbone particle adjacent to the first backbone particle is four.

[0056] When determining the position of the second main chain particle included in the double bead model, the number of candidate positions of the third main chain particle adjacent to the candidate position of the second main chain particle is 3. When determining the position of the third main chain particle included in the double bead model, the number of candidate positions of the fourth main chain particle adjacent to the candidate position of the third main chain particle is 3.

[0057] As described above, in the case of specifying the coordinates of each main chain particle by using a first state variable represented by relative coordinates, the number of candidate positions of the (i+1)th main chain particle is equal to or less than a certain number regardless of the number of main chain particles included in the double-bead model, and therefore the number of bits of the first state variable does not increase.

[0058] For example, when the first state variable is Figure 3A When the number of bits of the first state variable is expressed as shown in the relative coordinates, there is an advantage that the number of bits of the first state variable can be reduced. When searching for a solution to a problem replaced by the Ising model, the number of parameters such as the first state variable increases as the scale of the problem becomes larger. For this reason, since it is possible that the operation unit that performs the solution search is not allowed to store all the parameters used in the solution search in the storage unit used as a cache, the efficiency of the operation may be reduced. As described above, the advantage of reducing the number of bits of the first state variable makes it easier to ensure the storage capacity required to store various parameters required for the solution, thereby improving the efficiency of the calculation.

[0059] Similarly, Figure 3B FIG. 4 shows the case where the second state variable is represented by relative coordinates. Figure 3B As shown, when determining the position of the first main chain particle included in the double-bead model, the number of candidate positions of the first side chain particle adjacent to the first main chain particle is three in addition to the position of the second main chain particle.

[0060] When determining the position of the second main chain particle included in the double bead model, in addition to the positions of the first main chain particle and the third main chain particle, the number of candidate positions of the second side chain particle adjacent to the candidate position of the second main chain particle is two. When determining the position of the third main chain particle included in the double bead model, in addition to the positions of the second main chain particle and the fourth main chain particle, the number of candidate positions of the third side chain particle adjacent to the candidate position of the third main chain particle is two.

[0061] As described above, in the case where the coordinates of each side chain particle are specified by using the second state variable represented by the relative coordinates, the number of candidate positions of the i-th side chain particle when the coordinates of the i-th main chain particle are determined is equal to or less than a certain number regardless of the number of side chain particles included in the double bead model. Therefore, the number of bits of the second state variable does not increase.

[0062] For example, when the second state variable is Figure 3BWhen the number of bits of the second state variable is expressed as shown in the relative coordinates, there is an advantage that the number of bits of the second state variable can be reduced. When searching for a solution to a problem replaced by the Ising model, the number of parameters such as the second state variable increases as the scale of the problem becomes larger. For this reason, since it is possible that the operation unit that performs the solution search is not allowed to store all the parameters used in the solution search in the storage unit used as a cache, the efficiency of the operation may be reduced. As described above, the advantage of reducing the number of bits of the second state variable makes it easier to ensure the storage capacity required to store various parameters required for the solution, thereby improving the efficiency of the calculation.

[0063] <Specific examples of the first state variable and the second state variable>

[0064] Next, specific examples of the first state variable and the second state variable expressed by relative coordinates will be described. Figure 4A and Figure 4B is a diagram showing a specific example of the first state variable and the second state variable.

[0065] like Figure 4A As shown, the center position of the face-centered cubic lattice (FCC) 410 is the origin (coordinate = (0,0,0)), and the i-th main chain particle is located at the origin. In this case, the number of candidate positions of the (i+1)th main chain particle adjacent to the i-th main chain particle is 12 (see <0> to <11> ). The relative coordinates of the 12 candidate positions of the main chain particles and the origin are as follows:

[0066] Location <0> Relative coordinates: (-1,-1,0)

[0067] Location <1> Relative coordinates: (-1,1,0)

[0068] Location <2> Relative coordinates: (1,-1,0)

[0069] Location <3> Relative coordinates: (1,1,0)

[0070] Location <4> Relative coordinates: (-1,0,-1)

[0071] Location <5> Relative coordinates: (-1,0,1)

[0072] Location <6> Relative coordinates: (1,0,-1)

[0073] Location <7> Relative coordinates: (1,0,1)

[0074] Location <8> Relative coordinates: (0,-1,-1)

[0075] Location <9> Relative coordinates: (0,-1,1)

[0076] Location <10> Relative coordinates: (0,1,-1)

[0077] Location <11> The relative coordinates are: (0,1,1).

[0078] For example, the coordinates of the (i+1)th backbone particle adjacent to the i-th backbone particle are relative coordinates of any candidate position among the above-mentioned 12 candidate positions (see reference numeral 420 ).

[0079] As an example, Figure 4A (b) presents an example of a double-bead model including eight main chain particles, in which the coordinates of each main chain particle are calculated by using the first state variable represented by the relative coordinates. For example, when the coordinates of the zeroth main chain particle are set at the origin (coordinates = (0,0,0)) and the first state variable is set to [0,1,4,6,3,2,7], the coordinates can be calculated as follows:

[0080] ·The coordinates of the first main chain particle = the coordinates of the zeroth main chain particle + position <0> The relative coordinates of

[0081] Coordinates of the second main chain particle = coordinates of the first main chain particle + position <1> The relative coordinates of

[0082] Coordinates of the third main chain particle = coordinates of the second main chain particle + position <4> The relative coordinates of

[0083] Coordinates of the fourth main chain particle = coordinates of the third main chain particle + position <6> The relative coordinates of

[0084] Coordinates of the fifth main chain particle = coordinates of the fourth main chain particle + position <3> The relative coordinates of

[0085] ·Coordinates of the sixth main chain particle = coordinates of the fifth main chain particle + position <2> The relative coordinates of ; and

[0086] The coordinates of the seventh main chain particle = the coordinates of the sixth main chain particle + position <7> The relative coordinates of .

[0087] Note that the number of values ​​included in the first state variable is (the number of main chain particles - 1) as described above.

[0088] Similarly, if Figure 4B As shown in (a), the center position of the face-centered cubic lattice (FCC) 410 is the origin (coordinate = (0,0,0)), and the i-th main chain particle is located at the origin. In this case, the number of candidate positions of the i-th side chain particle adjacent to the i-th main chain particle is 12 (see <0> to <11> ). The relative coordinates of the 12 candidate positions of the side chain particles and the origin are as follows:

[0089] Location <0> Relative coordinates: (-1,-1,0)

[0090] Location <1> Relative coordinates: (-1,1,0)

[0091] Location <2> Relative coordinates: (1,-1,0)

[0092] Location <3> Relative coordinates: (1,1,0)

[0093] Location <4> Relative coordinates: (-1,0,-1)

[0094] Location <5> Relative coordinates: (-1,0,1)

[0095] Location <6> Relative coordinates: (1,0,-1)

[0096] Location <7> Relative coordinates: (1,0,1)

[0097] Location <8> Relative coordinates: (0,-1,-1)

[0098] Location <9> Relative coordinates: (0,-1,1)

[0099] Location <10> Relative coordinates: (0,1,-1)

[0100] Location <11> The relative coordinates are: (0,1,1).

[0101] For example, the coordinates of the ith sub-chain particle adjacent to the ith main-chain particle are relative coordinates of any candidate position among the 12 candidate positions (see reference numeral 420 ).

[0102] As an example, Figure 4B (b) presents an example of a double-bead model including eight main chain particles, in which the positions of all main chain particles are determined and the coordinates of the side chain particles corresponding to each main chain particle are calculated by using the second state variable represented by the relative coordinates. For example, when the second state variable = [11, 8, 3, 4, 9, 5, 7, 10] is set relative to the coordinates of the zeroth main chain particle to the seventh main chain particle, the coordinates can be calculated as follows:

[0103] ·The coordinates of the zeroth side chain particle = the coordinates of the zeroth main chain particle + position <11> The relative coordinates of

[0104] ·The coordinates of the first side chain particle = the coordinates of the first main chain particle + position <8> The relative coordinates of

[0105] ·Coordinates of the second side chain particle = coordinates of the second main chain particle + position <3> The relative coordinates of

[0106] ·Coordinates of the third side chain particle = coordinates of the third main chain particle + position <4> The relative coordinates of

[0107] Coordinates of the fourth side chain particle = coordinates of the fourth main chain particle + position <9> The relative coordinates of

[0108] · Coordinates of the fifth side chain particle = coordinates of the fifth main chain particle + position <5> The relative coordinates of

[0109] ·Coordinates of the sixth side chain particle = coordinates of the sixth main chain particle + position <7> The relative coordinates of ; and

[0110] ·The coordinates of the seventh side chain particle = the coordinates of the seventh main chain particle + position <10> The relative coordinates of .

[0111] Note that the number of values ​​included in the second state variable is equal to (the number of main chain particles) as described above.

[0112] <State Transition of Double Bead Model>

[0113] Next, an example of the state transition of the double bead model will be described. In the first embodiment, the state transition of the double bead model is performed by moving the main chain particles as follows:

[0114] Select state S i (i = 2 to N), and by moving the i-th main chain particle, the state S i Transition to state S i Another state of being different;

[0115] The i-th backbone particle may move randomly, or may move so that the first state variable is incremented by 1; and

[0116] As the i-th main chain particle moves, the (i+1)-th main chain particle to the N-th main chain particle and the i-th side chain particle to the N-th side chain particle are transferred together with the i-th main chain particle.

[0117] In a first embodiment, the state transition of the two-bead model is performed by moving the side chain particles as follows:

[0118] Select state R i (i = 2 to N), and transform it into the state R i Another state of being different;

[0119] The i-th side chain particle can be selected arbitrarily, or can move when the i-th main chain particle moves;

[0120] The i-th side chain particle may move randomly, or may move so that the second state variable is incremented by 1; and

[0121] Move the i-th side chain particle to a destination where it may not collide with another main chain particle or side chain particle,

[0122] wherein, for example, the coordinates of the destination of movement of the i-th side chain particle are selected to be inconsistent with the coordinates of all main chain particles and the coordinates of all side chain particles except the i-th side chain particle, or

[0123] The coordinates of the destination of the i-th side chain particle are selected so as not to coincide with the coordinates of the main chain particle adjacent to the i-th main chain particle (the (i-1)th main chain particle or the (i+1)th main chain particle).

[0124] As described above, in the first embodiment, the state transition is performed in consideration of the side chain particles.

[0125] <Details of Energy Calculation Processing>

[0126] Next, the details of energy calculation processing performed by the stable conformation search system according to the first embodiment based on the interaction between main chain particles, the interaction between main chain particles and side chain particles, etc. when searching for a stable conformation of the double bead model will be described. Figure 5 is a diagram showing details of energy calculation processing.

[0127] like Figure 5 As shown, the energy calculation process (see reference numeral 510) performed when searching for a stable conformation of the double-bead model includes the following processes and is represented, for example, by the following formula 1:

[0128] a process for calculating energy based on the interaction between the main chain particles, the energy including: energy depending on the angle between the main chain particles; energy depending on the dihedral angle between the main chain particles; energy depending on the repulsive or attractive force between the main chain particles; and energy depending on the distance between the main chain particles at both ends in the case where the double bead model has a ring structure; and

[0129] Processing for calculating side chain particle correlation energies.

[0130] For example, the energy calculation process is represented by the following equation 1:

[0131]

[0132] Where K θ , K τ and K endIn the above formula 1: θ represents the angle formed by three consecutive (i-1)th main chain particles to the (i+1)th main chain particles; τ represents the dihedral angle formed by four consecutive (i-2)th main chain particles to the (i+1)th main chain particles; θ0 represents the target value of the angle; τ0 represents the target value of the dihedral angle; r ij represents the distance between the i-th main chain particle and the j-th main chain particle; σ ij represents the distance between the i-th main chain particle and the j-th main chain particle in the natural state; r 1N Represents the distance between the two ends of the main chain particles; R th Indicates the target value of the distance between the main chain particles at both ends; E res represents the energy term associated with the side chain particles.

[0133] In order to perform the above-mentioned energy calculation processing, the stable conformation search system according to the first embodiment calculates the difference between the energy values ​​generated when the positions of the main chain particles are moved as the calculation object. Therefore, the stable conformation search system according to the first embodiment makes it possible to reduce the amount of calculation in the energy calculation processing.

[0134] Figures 6 to 8 The first to third figures show the details of the respective items in the energy calculation process. Figure 6 As indicated by reference numeral 610 in FIG. 1 , when calculating energy depending on the angle between main chain particles, the stable conformation search system according to the first embodiment calculates the difference between energy values ​​generated when the positions of the main chain particles are moved as a calculation target.

[0135] For example, in the case of a state transition of the double-bead model by moving the (i+1)th main chain particle relative to the i-th main chain particle, the stable conformation search system calculates energy depending on the difference (=θ-θ') between the following angle (θ) and angle (θ') as energy depending on the angle between the main chain particles:

[0136] the angle (θ) formed by the (i-1)th main chain particle, the i-th main chain particle and the (i+1)th main chain particle before the (i+1)th main chain particle moves (the angle θ centered on the i-th main chain particle before the (i+1)th main chain particle moves), and

[0137] Angle (θ') formed by the (i-1)th main chain particle, the i-th main chain particle and the (i+1)th main chain particle after the (i+1)th main chain particle moves (angle θ' centered on the i-th main chain particle after the (i+1)th main chain particle moves).

[0138] For example, the angle centered on the (i-1)th main chain particle, the angle centered on the (i-2)th main chain particle, etc. remain unchanged before and after the movement of the (i+1)th main chain particle, and thus can be excluded from the calculation of the energy depending on the angle between the main chain particles. When the (i+1)th main chain particle moves, all the main chain particles of the (i+2)th main chain particle and thereafter are transferred while maintaining their angles. For example, the angle centered on the (i+1)th main chain particle, the angle centered on the (i+2)th main chain particle, etc. remain unchanged before and after the movement of the (i+1)th main chain particle, and thus can be excluded from the calculation of the energy depending on the angle between the main chain particles.

[0139] Similarly, if you use Figure 6 As indicated by reference numeral 620 in FIG. 1 , when calculating energy depending on the dihedral angle between main chain particles, the stable conformation search system according to the first embodiment calculates the difference between energy values ​​generated when the positions of the main chain particles are moved as a calculation target.

[0140] For example, in the case of a state transition of the double-bead model by moving the (i+1)th main chain particle relative to the i-th main chain particle, the stable conformation search system calculates energy depending on the difference between the following dihedral angle τ and the dihedral angle τ' (=τ-τ') as energy depending on the dihedral angle between the main chain particles:

[0141] The dihedral angle τ formed between the plane A formed by the (i-2)th main chain particle to the i-th main chain particle and the plane B formed by the (i-1)th main chain particle to the (i+1)th main chain particle before the (i+1)th main chain particle moves, and

[0142] After the (i+1)th backbone particle moves, the dihedral angle τ' formed between the plane A formed by the (i-2)th backbone particle to the i-th backbone particle and the plane B' formed by the (i-1)th backbone particle to the (i+1)th backbone particle.

[0143] For example, before and after the movement of the (i+1)th main chain particle, the dihedral angle formed between the plane formed by the (i-3)th main chain particle to the (i-1)th main chain particle and the plane formed by the (i-2)th main chain particle to the i-th main chain particle, etc. remain unchanged. Therefore, in the calculation of the energy depending on the dihedral angle between the main chain particles, the above dihedral angle can be excluded. When the (i+1)th main chain particle moves, all the main chain particles of the (i+2)th main chain particle and thereafter are transferred while maintaining their angles. For example, before and after the movement of the (i+1)th main chain particle, the dihedral angle formed between the plane formed by the (i-1)th main chain particle to the (i+1)th main chain particle and the plane formed by the i-th main chain particle to the (i+2)th main chain particle, etc. remain unchanged. Therefore, in the calculation of the energy depending on the dihedral angle between the main chain particles, the above dihedral angle can be excluded.

[0144] If used Figure 7 As indicated by reference numeral 710 in FIG. 1 , when calculating the energy depending on the repulsive force or attractive force between the main chain particles, the stable conformation search system according to the first embodiment sets all pairs of main chain particles belonging to different groups as calculation targets. For example, the repulsive force or attractive force between the main chain particles belonging to the same group may be excluded from the energy calculation.

[0145] For example, in the case of a state transition of the double-bead model by moving the (i+1)th main chain particle relative to the i-th main chain particle, the stable conformation search system calculates the energy depending on the difference between:

[0146] before the (i+1)th backbone particle moves, the repulsive or attractive forces calculated for all pairs of backbone particles belonging to group A (a group including the first backbone particle to the i-th backbone particle) and backbone particles belonging to group B (a group including the (i+1)th backbone particle and subsequent backbone particles), and

[0147] After the (i+1)th backbone particle moves, the repulsive or attractive forces calculated for all pairs of backbone particles belonging to group A (a group including the first backbone particle to the i-th backbone particle) and backbone particles belonging to group B (a group including the (i+1)th backbone particle and subsequent backbone particles).

[0148] For example, the energy depending on the repulsive or attractive force between the main chain particles belonging to the A group remains unchanged before and after the movement of the (i+1)th main chain particle, and thus can be excluded from the calculation of the energy depending on the repulsive or attractive force between the main chain particles. Since the main chain particles belonging to the B group are transferred while maintaining the distance between them when the (i+1)th main chain particle moves, the repulsive or attractive force between the main chain particles belonging to the B group remains unchanged before and after the movement of the (i+1)th main chain particle. Therefore, in the calculation of the energy depending on the repulsive or attractive force between the main chain particles, the energy depending on the above repulsive or attractive force can be excluded.

[0149] If used Figure 7 As indicated by the reference numeral 720 in FIG. 1 , when the double-bead model has a ring structure, each time the state of the double-bead model is changed by moving the (i+1)th main chain particle relative to the i-th main chain particle, the distance between the main chain particles at both ends changes.

[0150] Therefore, when the double bead model has a ring structure, each time the (i+1)th main chain particle moves, the energy depending on the distance between the main chain particles at both ends is calculated (when the double bead model has a ring structure, the above energy is not excluded from the calculation object).

[0151] If used Figure 8 As indicated by reference numeral 810 in FIG. 1 , the stable conformation search system according to the first embodiment performs a process of calculating the side chain particle correlation energy according to Formula 2 given below, while applying the following penalty:

[0152] Penalty when the coordinates of the i-th side chain particle are consistent with the coordinates of any main chain particle other than the i-th main chain particle or any side chain particle in the side chain particle other than the i-th side chain particle;

[0153] The penalty for the case where the i-th side chain particle is positioned adjacent to the j-th main chain particle (i≠j); and

[0154] Penalty when side chain particles forming SS bonds are separated from each other (penalty when side chain particles are not adjacent to each other, or penalty when the distance between side chain particles forming SS bonds is equal to or greater than the target value).

[0155] E res =K dup ·∑ i，j(i≠j) {dup(p(i), q(j))+dup(q(i), q(j))}+K adj ·∑ i，j(i≠j) adj(p(i),q(j))+∑ i，j∈{SSB} K ss(dist(q(i),q(j))-D ss ) 4 +K inter ·∑ i，j(i≠j) Inter(A(i), A(j))

[0156] (Formula 2)

[0157] In the above formula 2, K dup , K adj , K ss and K inter In the above formula 2: p(i) represents the coordinates of the ith main chain particle; q(j) represents the coordinates of the jth side chain particle; dup((p(i),q(j)) is a function that takes "1" when the coordinates of the ith main chain particle and the jth side chain particle are consistent with each other, otherwise it takes "0"; adj(p(i),q(j)) is a function that takes "1" when the coordinates of the ith main chain particle and the jth side chain particle are adjacent to each other, otherwise it takes "0"; dist(q(i),q(j)) represents the distance between the ith side chain particle and the jth side chain particle; SSB represents the set of side chain particles that form SS bonds; D SS represents the target value of the distance between the side chain particles forming each SS bond; Inter(A(i), A(j)) represents the energy based on the interaction between the i-th amino acid residue and the j-th amino acid residue.

[0158] Since the energy of repulsive force or attractive force depending on the type of residues is generated between the i-th amino acid residue and the j-th amino acid residue, the energy of repulsive force or attractive force depending on the type of residues is added to the above formula 2. Here, A(i) represents the type of the i-th amino acid residue, and A(j) represents the type of the j-th amino acid residue.

[0159] As described above, in the first embodiment, the energy calculation process taking into account the state transition of the side chain particles and the energy calculation process taking into account the characteristics of the side chain particles are performed.

[0160] Fig. 9 It is a diagram showing the magnitude of the interaction between amino acid residues. Fig. 9 The color of each region where the amino acid residues on the horizontal axis and the amino acid residues on the vertical axis intersect each other represents the magnitude of the interaction between the types of amino acid residues. Fig. 9 The final term Inter(A(i), A(j)) is derived (eg, Inter(A(i), A(j)) is derived by selecting A(i) from the horizontal axis and A(j) from the vertical axis).

[0161] As described above, in the first embodiment, energy calculation processing is performed in consideration of the types of amino acid residues.

[0162] <System configuration of stable conformation search system and functional configuration of device>

[0163] Next, refer to Figures 10 to 12 A specific example of the system configuration, the functional configuration of the device, and the processing of the stable conformation search system according to the first embodiment will be described. Fig.10 is a diagram showing an example of a system configuration of a stable constellation search system and a functional configuration of a terminal device. Fig.11 is a diagram showing an example of a system configuration of a stable conformation search system and a functional configuration of an Ising device. Fig.12 A specific example of the processing performed by the Ising device is shown.

[0164] (1) System configuration of the stable conformation search system and functional configuration of the terminal device

[0165] like Fig.10 As shown, the stable conformation search system 1000 according to the first embodiment includes a terminal device 1010 and an Ising device 1020 .

[0166] A stable constellation search instruction program is installed in the terminal device 1010. The terminal device 1010 functions as a state variable input unit 1011, an energy model input unit 1012, a search instruction unit 1013, and an execution instruction unit 1014 when executing the program.

[0167] The state variable input unit 1011 receives an input of a two-bead model in an initial state as a stable conformation search object. The state variable input unit 1011 identifies the two-bead model by receiving the number N of amino acid residues (N is an integer greater than 2) and the setting of the amino acid residue sequence. The state variable input unit 1011 sets the initial state of the two-bead model by randomly determining N-1 values ​​in the first state variable and N values ​​in the second state variable.

[0168] The state variable input unit 1011 receives an input of a relative coordinate (reference numeral 420) representing a first state variable. The execution instruction unit 1014 is notified of the dual-ball model and the relative coordinate (reference numeral 420) in the initial state, which are received by the state variable input unit 1011.

[0169] The energy model input unit 1012 receives an input of a calculation formula (reference numeral 510) for executing energy calculation processing. The calculation formula (reference numeral 510) received by the energy model input unit 1012 is notified to the execution instruction unit 1014.

[0170] The search instruction unit 1013 receives an input of a search iteration number or a search execution variable for searching for a stable conformation of the double-bead model. The search iteration number or the search execution variable mentioned herein is equivalent to the number of updates of each state variable in the state variable in the Ising device 1020. The search iteration number or the search execution variable is notified to the execution instruction unit 1014, and the input of the search iteration number or the search execution variable is received by the search instruction unit 1013.

[0171] The execution instruction unit 1014 instructs the Ising device 1020 to search for a stable conformation based on all the information notified by the state variable input unit 1011, the energy model input unit 1012, and the search instruction unit 1013. The execution instruction unit 1014 issues a search instruction to the Ising device 1020 by sending the following:

[0172] Search target: a two-bead model with N amino acid residues,

[0173] The initial state of the two-ball model: the first state variable and the second state variable are randomly determined,

[0174] Method of expressing the first state variable and the second state variable: relative coordinates (reference numeral 420 ),

[0175] Energy calculation method in the search process: calculation formula (reference numeral 510), and

[0176] The number of updates to state variables during the search process: the number of search iterations or search execution variables.

[0177] As a result of issuing a search instruction to the Ising device 1020, the execution instruction unit 1014 receives a search result from the Ising device 1020. The search result received by the execution instruction unit 1014 includes a calculation result of the minimum energy and a first state variable and a second state variable that achieve the minimum energy. Therefore, the execution instruction unit 1014 is enabled to output a stable conformation of the double-bead model and energy based on the interaction between the main chain particles in the double-bead model in the stable conformation, the interaction between the main chain particles and the side chain particles, and the like.

[0178] (2) System configuration of the stable conformation search system and functional configuration of the Ising device

[0179] Will refer to Fig.11 Describe the functional configuration of the Ising device 1020 (as mentioned above Fig.10 The system configuration of the stable conformation search system 1000 is described, and thus its description will be omitted).

[0180] The stable conformation search program is installed in the Ising device 1020. The Ising device 1020 functions as a search unit 1110, an optimal solution identification unit 1120, a state conversion unit 1130, and an energy calculation unit 1140 when executing the program.

[0181] The search unit 1110 searches for a stable conformation of the dual-bead model based on a search instruction from the terminal device 1010, and sends the search result to the terminal device 1010. For example, for the dual-bead model, the search unit 1110 updates (changes) the first state variable and the second state variable within the range of the indicated search iteration number (or search execution variable). For each state transition of the dual-bead model performed by updating the first state variable and the second state variable, the search unit 1110 determines whether the transition is acceptable based on the calculated energy value. The search unit 1110 sends the minimum energy value identified in the following energy values ​​and the first state variable and the second state variable that achieve the minimum energy value as the search result to the terminal device 1010, wherein for each of the above energy values, the transition is determined to be acceptable.

[0182] The optimal solution identification unit 1120 is an example of an identification unit. Whenever the search unit 1110 notifies the optimal solution identification unit 1120 of an energy value, the optimal solution identification unit 1120 compares the notified energy value with the energy value stored in the energy value storage unit 1150, and in the case where the notified energy value is less than the stored energy value, stores the notified energy value in the energy value storage unit 1150. In this way, the optimal solution identification unit 1120 can identify the minimum energy value.

[0183] Whenever the search unit 1110 notifies the optimal solution identification unit 1120 of the updated first state variable and the second state variable, the optimal solution identification unit 1120 saves the updated first state variable and the second state variable. In the case where the optimal solution identification unit 1120 is notified of the corresponding energy value by the search unit 1110 and the optimal solution identification unit 1120 stores the corresponding energy value in the energy value storage unit 1150, the optimal solution identification unit 1120 stores the saved first state variable and the second state variable in the state variable storage unit 1160. In this way, the optimal solution identification unit 1120 can identify the first state variable and the second state variable corresponding to the minimum energy value.

[0184] The state conversion unit 1130 is an example of a first calculation unit and a second calculation unit. Whenever the state conversion unit 1130 is notified of the updated first state variable and the second state variable by the search unit 1110, the state conversion unit 1130 refers to the relative coordinate storage unit 1170 and calculates the coordinates of each main chain particle and each side chain particle in the double bead model after the state transition. Note that the relative coordinate storage unit 1170 stores the relative coordinates transmitted together with the search instruction from the terminal device 1010 in advance.

[0185] The energy calculation unit 1140 is notified of the coordinates of each main chain particle and each side chain particle in the double-bead model after the state transition calculated by the state conversion unit 1130 .

[0186] The energy calculation unit 1140 is an example of a calculation unit. Whenever the energy calculation unit 1140 is notified of the coordinates of each main chain particle and each side chain particle in the double bead model after the state transition by the state conversion unit 1130, the energy calculation unit 1140 performs energy calculation processing. For example, the energy calculation unit 1140 calculates the energy of the reference particle 510 based on the calculation formula (reference numeral 510) transmitted together with the search instruction from the terminal device 1010. Figures 6 to 8 The search unit 1110 is notified of the energy value calculated by the energy calculation unit 1140, and the search unit 1110 performs the next search by regarding the updated first state variable and the second state variable as the first state variable and the second state variable before the next update.

[0187] (3) Specific example of processing by Ising device

[0188] Then, refer to Fig.12 A specific example of the processing performed by the Ising device 1020 will be described. Fig.12 An example is shown in which the number N of amino acid residues is 8, the search unit 1110 updates the first state variable to [0, 1, 4, 6, 3, 2, 7], and notifies the optimal solution identification unit 1120 and the state conversion unit 1130 of the updated first state variable. Fig.12 An example is shown in which the search unit 1110 updates the second state variables to [11, 8, 3, 4, 9, 5, 7, 10] and notifies the optimal solution identification unit 1120 and the state conversion unit 1130 of the updated second state variables.

[0189] like Fig.12 As shown, the state conversion unit 1130 calculates the coordinates of each of the 8 main chain particles and the coordinates of each of the 8 side chain particles based on the relative coordinates as follows in response to the notification of the updated first state variable and the second state variable from the search unit 1110:

[0190] The coordinates of the zeroth main chain particle = (0,0,0),

[0191] The coordinates of the first main chain particle = (0,0,0) + (-1,-1,0) = (-1,-1,0),

[0192] The coordinates of the second main chain particle = (-1,-1,0) + (-1,1,0) = (-2,0,0),

[0193] The coordinates of the third main chain particle = (-2, 0, 0) + (-1, 0, -1) = (-3, 0, -1),

[0194] The coordinates of the fourth main chain particle = (-3, 0, -1) + (1, 0, -1) = (-2, 0, -2),

[0195] The coordinates of the fifth main chain particle = (-2, 0, -2) + (1, 1, 0) = (-1, 1, -2),

[0196] The coordinates of the sixth main chain particle = (-1, 1, -2) + (1, -1, 0) = (0, 0, -2),

[0197] The coordinates of the seventh main chain particle = (0,0,-2) + (1,0,1) = (1,0,-1),

[0198] The coordinates of the zeroth side chain particle = (0,0,0) + (0,1,1) = (0,1,1),

[0199] The coordinates of the first side chain particle = (-1, -1, 0) + (0, -1, -1) = (-1, -2, -1),

[0200] The coordinates of the second side chain particle = (-2, 0, 0) + (1, 1, 0) = (-1, 1, 0),

[0201] The coordinates of the third side chain particle = (-3, 0, -1) + (-1, 0, -1) = (-4, 0, -2),

[0202] The coordinates of the fourth side chain particle = (-2, 0, -2) + (0, -1, 1) = (-2, -1, -1),

[0203] The coordinates of the fifth side chain particle = (-1, 1, -2) + (-1, 0, 1) = (-2, 1, -1),

[0204] The coordinates of the sixth side chain particle = (0,0,-2) + (1,0,1) = (1,0,-1), and

[0205] The coordinates of the seventh side chain particle = (1, 0, -1) + (0, 1, -1) = (1, 1, -2).

[0206] <Hardware Configuration of Terminal Device>

[0207] Next, the hardware configuration of the terminal device 1010 will be described. Fig.13 is a diagram showing an example of a hardware configuration of a terminal device.

[0208] like Fig.13 As shown, the terminal device 1010 includes a processor 1301, a memory 1302, an auxiliary storage device 1303, an interface (I / F) device 1304, a communication device 1305, and a drive device 1306. In the terminal device 1010, the processor 1301, the memory 1302, the auxiliary storage device 1303, the I / F device 1304, the communication device 1305, and the drive device 1306 are coupled to each other via a bus 1307.

[0209] The processor 1301 includes various computing devices such as a central processing unit (CPU) and a graphics processing unit (GPU). The processor 1301 loads various programs (eg, a stable conformation search instruction program, etc.) onto the memory 1302 and executes the programs.

[0210] The memory 1302 includes a main storage device such as a read-only memory (ROM) and a random access memory (RAM). The processor 1301 and the memory 1302 form a so-called computer. When the processor 1301 executes various programs loaded on the memory 1302, the computer realizes the various functions described above.

[0211] The auxiliary storage device 1303 stores various programs and various types of information to be used when the processor 1301 executes the various programs.

[0212] The I / F device 1304 is a coupling device by which the terminal apparatus 1010 is coupled to an operation device 1310 and an output device 1320 as examples of external devices.

[0213] The communication device 1305 is a communication device for communicating with the Ising device 1020 via a network.

[0214] The drive device 1306 is a device at which the recording medium 1330 is set. The recording medium 1330 mentioned herein includes a medium that records information optically, electrically or magnetically, such as a CD-ROM, a floppy disk or a magneto-optical disk. The recording medium 1330 may also include a semiconductor memory that records information electrically, such as a ROM or a flash memory.

[0215] For various programs to be installed in the auxiliary storage device 1303, the distributed recording medium 1330 set in the drive device 1306 and the various programs recorded in the recording medium 1330 are read and then installed in the auxiliary storage device 1303. Alternatively, the various programs to be installed in the auxiliary storage device 1303 may be installed by downloading from a network via the communication device 1305.

[0216] Although only the hardware configuration of the terminal device 1010 is described herein and the hardware configuration of the Ising device 1020 is not described, the hardware configuration of the Ising device 1020 may be the same as or similar to, for example, the hardware configuration of the terminal device 1010. Alternatively, the hardware configuration of the Ising device 1020 may be the same as or similar to the hardware configuration of a so-called quantum computer.

[0217] <Order of stable conformation search process>

[0218] Next, the procedure of the stable conformation search process performed by the stable conformation search system 1000 will be described. Fig.14 is a flowchart showing the sequence of the stable conformation search process.

[0219] At step S1401 , the terminal device 1010 receives the number N (N is an integer greater than 2) of amino acid residues and the setting of the amino acid residue sequence, thereby identifying a double-bead model (including N main chain particles and N side chain particles) as a stable conformation search target.

[0220] At step S1402, the terminal device 1010 randomly determines the initial state of the double-bead model including N main chain particles and N side chain particles. For example, the terminal device 1010 randomly determines N-1 values ​​in the first state variable and N values ​​in the second state variable.

[0221] At step S1403, the terminal device 1010 receives an input for representing relative coordinates of the first state variable and the second state variable. The terminal device 1010 receives an input for a calculation formula for an energy calculation process in a search process. The terminal device 1010 receives a search iteration number or a search execution variable as the update number of the first state variable and the second state variable in the search process. The terminal device 1010 sends a search instruction to the Ising device 1020 based on all received information to search for a stable conformation of the double-bead model.

[0222] At step S1404, the Ising device 1020 performs a search process. Details of the search process performed by the Ising device 1020 will be described later.

[0223] At step S1405 , the terminal device 1010 receives the search result from the Ising device 1020 , and outputs the calculation result of the minimum energy value of the double-bead model as the search target and the first state variable and the second state variable that achieve the minimum energy value.

[0224] <Search Processing Order>

[0225] Will refer to Fig.15 The details of the search process (step S1404 ) of the Ising device 1020 are described. Fig.15 is a flowchart showing the details of the search process.

[0226] exist Fig.15 At step S1501, the Ising device 1020 calculates the energy value of the double-bead model in the initial state, and stores the calculated energy value as the minimum energy value (minE) at the current time point.

[0227] At step S1502, the Ising device 1020 randomly changes any one of the N-1 values ​​in the first state variable specifying the coordinates of the N main chain particles in the double-bead model, thereby updating the coordinates of the main chain particles. Note that the first state variable specifying the coordinates of the N main chain particles is represented by relative coordinates. For example, the processing at step S1502 such as randomly changing the first state variable and updating the coordinates of the main chain particles corresponds to the following processing: by using the coordinates of the i-th main chain particle in the sequence in the lattice space and the first state variable represented by the relative coordinates between the i-th main chain particle and the (i+1)-th main chain particle in the lattice space, the coordinates of the (i+1)-th main chain particle in the sequence in the lattice space are calculated.

[0228] At step S1503, the Ising device 1020 updates the coordinates of the side chain particles while updating the coordinates of the main chain particles. Note that the coordinates of the side chain particles are represented by relative coordinates. For example, the process at step S1503, such as updating the coordinates of the side chain particles, corresponds to the following process: by using the coordinates of the (i+1)th main chain particle in the lattice space and the second state variable represented by the relative coordinates between the (i+1)th main chain particle and the (i+1)th side chain particle, the coordinates of the (i+1)th side chain particle corresponding to the (i+1)th main chain particle in the lattice space are calculated.

[0229] At step S1504 , the Ising device 1020 calculates the energy value E′ of the double-bead model based on the current coordinates of the main chain particles and the side chain particles.

[0230] At step S1505, the Ising device 1020 performs a determination process of determining whether the calculated energy value E' is a minimum energy value. The details of the determination process (step S1505) will be described later.

[0231] At step S1506, the Ising device 1020 determines whether the process has reached the search iteration number for the side chain particles. If it is determined at step S1506 that the process has not reached the search iteration number for the side chain particles (No at step S1506), the Ising device 1020 proceeds to step S1507.

[0232] In step S1507 , the Ising device 1020 changes the second state variable and updates the coordinates of the side chain particles, and then returns to step S1504 .

[0233] On the other hand, if it is determined at step S1506 that the process has reached the search iteration number (YES at step S1506 ), the Ising device 1020 resets the search iteration number for the side chain particles and then proceeds to step S1508 .

[0234] At step S1508, the Ising device 1020 determines whether the process has reached the search iteration number of the main chain particles. If it is determined at S1508 that the process has not reached the search iteration number of the main chain particles (No at step S1508), the Ising device 1020 returns to step S1502.

[0235] On the other hand, if it is determined at step S1508 that the process has reached the number of search iterations for the main chain particles (yes at step S1508), the Ising device 1020 terminates the search process and returns to Fig.14 Step S1405 in .

[0236] <Determine the order of processing>

[0237] Will refer to Fig.16 Details of the determination process (step S1505 ) performed by the Ising device 1020 are described. Fig.16 is a flowchart showing details of the determination process.

[0238] exist Fig.16 At step S1601 in the embodiment, the Ising device 1020 calculates the difference between the energy value E stored as the minimum energy value (minE) at the current time point and the energy value E' calculated at step S1504. The Ising device 1020 determines whether the calculated difference is less than a predetermined thermal noise (in Fig.16 In the example of , it is -Tlog(a)). Here, a is a uniform random number and is a value that satisfies a∈[0,1].

[0239] If it is determined at step S1601 that the calculated difference is smaller than the predetermined thermal noise (−Tlog(a)) (YES at step S1601 ), the Ising device 1020 proceeds to step S1602 .

[0240] At step S1602 , the Ising device 1020 determines that the transition is acceptable, adopts the updated first state variable or the second state variable, and updates the energy value (replacing the energy value E with the energy value E′).

[0241] On the other hand, if it is determined at step S1601 that the calculated difference is equal to or greater than the predetermined thermal noise (−Tlog(a)) (NO at step S1601 ), the Ising device 1020 proceeds to step S1603 .

[0242] At step S1603 , the Ising device 1020 determines that the transition is unacceptable and discards the updated first state variable or the second state variable.

[0243] At step S1604, the Ising device 1020 determines whether the energy value E is less than the minimum energy value (minE). If it is determined at step S1604 that the energy value E is less than the minimum energy value (minE) (yes at step S1604), the Ising device 1020 proceeds to step S1605.

[0244] At step S1605, the Ising device 1020 stores the energy value E as the minimum energy value (minE) at the current time point. The Ising device 1020 stores the first state variable and the second state variable as the first state variable and the second state variable corresponding to the minimum energy value at the current time point.

[0245] On the other hand, if it is determined at step S1604 that the energy value E is equal to or greater than the minimum energy value (minE) (No at step S1604), the Ising device 1020 terminates the determination process and returns to Fig.15 Step S1506 in .

[0246] <Effect of reducing the amount of calculation>

[0247] Next, the effect of reducing the amount of calculation in the stable conformation search process performed by the stable conformation search system 1000 according to the first embodiment will be described. Fig.17 is a diagram showing an example of the effect of reducing the amount of calculation.

[0248] exist Fig.17 , as a comparative example, reference numeral 1710 indicates the amount of calculation when searching for a stable conformation when the number of amino acid residues N is 20 by the following operations: [a] using a first state variable and a second state variable represented by absolute coordinates, and [b] calculating energy without restricting the calculation object.

[0249] On the other hand, reference numeral 1720 indicates the amount of calculation when a stable structure search is performed by the stable conformation search system 1000 according to the first embodiment when the number of amino acid residues N is 20 by the following operations: [a] using the first state variable and the second state variable represented by relative coordinates, and [b] limiting the calculation target to the energy difference generated when the position of each particle in the main chain particle and the side chain particle is moved.

[0250] In the items of calculation amounts denoted by reference numerals 1710 and 1720, "Number of bits" indicates the number of bits used to represent the coordinates of main chain particles and side chain particles. "RAM capacity" indicates the capacity of a memory used for energy calculation processing.

[0251] In each of the terms, the computational amount for the corresponding cases where the lattice space is a face-centered cubic lattice and a knight's walk lattice is presented.

[0252] As is apparent from the comparison between reference numerals 1710 and 1720 , the stable conformation search system 1000 according to the first embodiment makes it possible to significantly reduce the amount of calculation in the stable conformation search process.

[0253] For example, the stable conformation search system 1000 according to the first embodiment searches for stable conformations by [a] using the first state variable and the second state variable represented by relative coordinates, and [b] limiting the calculation target to the energy difference before and after each state transition, thereby reducing the amount of calculation compared to the case of searching for stable conformations by using absolute coordinates and not limiting the calculation target.

[0254] For example, the stable conformation search system 1000 according to the first embodiment makes it possible to significantly increase the number N of amino acid residues in the two-bead model that is a search target under current computing power.

[0255] As is clear from the above description, the stable conformation search system 1000 according to the first embodiment obtains a double-bead model in which N main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, and calculates the coordinates of the (i+1)th main chain particle in the sequence in the lattice space by using:

[0256] The coordinates of the ith backbone particle in the sequence in lattice space; and

[0257] The first state variable represented by the relative coordinates between the i-th backbone particle and the (i+1)-th backbone particle in the lattice space.

[0258] The stable conformation search system 1000 according to the first embodiment obtains a double-bead model in which N main chain particles arranged in sequence and N side chain particles correspondingly bound to the respective N main chain particles are arranged in a lattice space, and calculates the coordinates of the (i+1)th side chain particle in the sequence in the lattice space by using:

[0259] The coordinates of the (i+1)th backbone particle in the sequence in lattice space; and

[0260] The second state variable represented by the relative coordinates between the (i+1)th main chain particle and the (i+1)th side chain particle in the lattice space.

[0261] The stable conformation search system 1000 according to the first embodiment repeatedly executes the process of calculating the energy value of the two-bead model every time any one of the first state variable and the second state variable changes.

[0262] The stable conformation search system 1000 identifies a first state variable and a second state variable whose energy values ​​are local minima.

[0263] As described above, the stable conformation search system 1000 according to the first embodiment uses the double bead model to search for stable conformations in consideration of side chain particles. Therefore, compared with the case of searching for stable conformations of main chain particles using a single bead model, the stable conformation search system 1000 according to the first embodiment makes it possible to improve the search accuracy.

[0264] [Second embodiment]

[0265] Although the first embodiment described above describes a case where any one of the values ​​included in the first state variable and the second state variable in the search process is randomly selected and updated, the value may be selected by any method other than random selection. For example, the value may be selected and updated in a predetermined order such as from the first value to the (N-1)th value among the (N-1) values ​​included in the first state variable, or from the first value to the Nth value among the N values ​​included in the second state variable.

[0266] In the case of a face-centered cubic lattice, there are 12 values ​​to which the value in each of the first state variable and the second state variable can be updated, and which value the value in the first state variable or the second state variable will be updated to can be randomly determined.

[0267] In the calculation formula (reference numeral 510) used for the energy calculation process in the first embodiment described above, k θ , k τ , K endThe weights of the terms are fixed, but the weights of the terms can be changed, for example, depending on the type of amino acid.

[0268] The first embodiment is described above in the following case: whenever the first state variable of the main chain particle changes, the local minimum is searched by calculating the energy of the double bead model for various values ​​in the second state variable of the side chain particle. However, the method of searching for the local minimum is not limited to this. For example, the minimum value can be searched by calculating the energy of the double bead model for various values ​​in the first state variable of the main chain particle. After all values ​​in the first state variable of the main chain particle are fixed, the local minimum can be searched by calculating the energy of the double bead model for various values ​​in the second state variable of the side chain particle.

[0269] In the above-described first embodiment, the face-centered cubic lattice and the Knight's Walk lattice are exemplified as the lattice space, but the lattice space is not limited to these.

[0270] Although the first embodiment is described above so that the stable conformation search system 1000 includes the terminal device 1010 and the Ising device 1020, the terminal device 1010 and the Ising device 1020 may be included in an integrated device. In this case, the stable conformation search program includes a stable conformation search instruction program, and the Ising device 1020 implements the reference Fig.10 Functional units of the terminal device 1010 described and reference Fig.11 Both functional units of the Ising device 1020 are described.

[0271] Based on the disclosed technology, a mode as described in this scheme can be conceived.

[0272] The present disclosure is not limited to the configurations shown herein, but may include configurations such as any one of the configurations exemplified in the above-mentioned embodiments, in combination with other elements, etc. These aspects may be changed without departing from the gist of the present disclosure, and may be appropriately determined according to the application mode thereof.

Claims

1. A stable conformation search system, comprising: A first computing unit is configured to: A model is obtained in which N main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, wherein N is an integer of 2 or greater, and calculating the coordinates of the i+1th main-chain particle in the sequence in the lattice space by using the coordinates of the i-th main-chain particle in the sequence in the lattice space and a first state variable represented by the relative coordinates between the i-th main-chain particle and the i+1th main-chain particle in the lattice space; a second calculation unit configured to calculate the coordinates of the i+1th side chain particle corresponding to the i+1th main chain particle in the lattice space by using the coordinates of the i+1th main chain particle in the lattice space and a second state variable, the second state variable being represented by the relative coordinates between the i+1th main chain particle and the i+1th side chain particle; a calculation unit configured to calculate a value of energy of the model in a case where any one of the first state variable and the second state variable changes; and An identification unit is configured to identify the first state variable and the second state variable whose value of the energy is a local minimum.

2. The stable conformation search system according to claim 1, further comprising: A searching unit configured to change any one of N-1 values ​​included in the first state variable and any one of N values ​​included in the second state variable.

3. The stable conformation search system according to claim 1, wherein: The position of the i+1th side chain particle in the lattice space is determined not to coincide with the positions of the main chain particle and the side chain particles other than the i+1th side chain particle.

4. The stable conformation search system according to claim 1, wherein: The calculation unit is configured to calculate a value of energy corresponding to an interaction between the main chain particle and the side chain particle based on coordinates of each of the main chain particle and the side chain particle in the lattice space.

5. The stable conformation search system according to claim 4, wherein: The energy corresponding to the interaction includes any one of the following: energy depending on the angle between the main chain particles; energy depending on the dihedral angle between the main chain particles; energy depending on the repulsive or attractive force between the main chain particles; energy depending on the distance between the main chain particles at both ends when the model has a ring structure; or energy depending on the characteristics of the side chain particles.

6. The stable conformation search system according to claim 5, wherein: The energy depending on the characteristics of the side chain particles includes any one of the following: a penalty in the case where the coordinates of the i+1th side chain particle in the lattice space coincide with the coordinates of any particle among the main chain particles other than the i+1th main chain particle and the side chain particles in the lattice space, a penalty in the case where the coordinates of the i+1th side chain particle in the lattice space are adjacent to the coordinates of any main chain particle among the main chain particles except the i+1th main chain particle in the lattice space, A penalty when the distance between the i+1th side chain particle and the side chain particle forming an SS bond with the i+1th side chain particle is equal to or greater than a target value, or The derived energy depends on the type of amino acid.

7. The stable conformation search system according to claim 2, wherein: The calculation unit is configured to calculate a difference between energies after the first state variable and the second state variable are updated and energies before the first state variable and the second state variable are updated.

8. The stable conformation search system according to claim 7, wherein: The search unit is configured to determine that changes in the first state variable and the second state variable are acceptable if the difference calculated by the calculation unit satisfies a predetermined condition.

9. A computer-implemented stable conformation search method, the stable conformation search method comprising: Obtaining a model in which N main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, wherein N is an integer of 2 or greater; calculating the coordinates of the i+1th main-chain particle in the sequence in the lattice space by using the coordinates of the i-th main-chain particle in the sequence in the lattice space and a first state variable represented by the relative coordinates between the i-th main-chain particle and the i+1th main-chain particle in the lattice space; calculating the coordinates of the i+1th side chain particle corresponding to the i+1th main chain particle in the lattice space by using the coordinates of the i+1th main chain particle in the lattice space and the second state variable, wherein the second state variable is represented by the relative coordinates between the i+1th main chain particle and the i+1th side chain particle; calculating a value of energy of the model when any one of the first state variable and the second state variable changes; and The first state variable and the second state variable whose values ​​of the energy are local minima are identified.

10. A non-transitory computer-readable recording medium storing a stable conformation search program, the stable conformation search program comprising instructions which, when executed by a computer, cause the computer to perform a process comprising: Obtaining a model in which N main chain particles arranged in sequence and N side chain particles correspondingly linked to the respective N main chain particles are arranged in a lattice space, wherein N is an integer of 2 or greater; calculating the coordinates of the i+1th main-chain particle in the sequence in the lattice space by using the coordinates of the i-th main-chain particle in the sequence in the lattice space and a first state variable represented by the relative coordinates between the i-th main-chain particle and the i+1th main-chain particle in the lattice space; calculating the coordinates of the i+1th side chain particle corresponding to the i+1th main chain particle in the lattice space by using the coordinates of the i+1th main chain particle in the lattice space and the second state variable, wherein the second state variable is represented by the relative coordinates between the i+1th main chain particle and the i+1th side chain particle; calculating a value of energy of the model when any one of the first state variable and the second state variable changes; and The first state variable and the second state variable whose values ​​of the energy are local minima are identified.

Images