Compound search device, compound search method, and compound search program

Abstract

One embodiment of the present invention provides a compound search device, a compound search method, and a compound search program capable of efficiently searching for low molecular weight compounds having activity and membrane permeability. In a compound search device according to one embodiment of the present invention, a processor performs: a process for specifying, as a main site, a part of a site in which the proportion of a ligand in the binding force with a target protein is equal to or greater than a specified value; and calculates a partial characteristic quantity derived from the main site among the characteristic quantities relating to the degree of aggregation of the amino acids with respect to the ligands. The processor calculates a similarity between the feature amount for the compound to be searched and the partial feature amount. The processor associates the similarity of the compound to be searched with information indicating the molecular weight and / or activity of the compound to be searched, and outputs the associated information to a recording device and / or a display device.

Description

Technical Field

[0001] The present invention relates to a compound search device, a compound search method and a compound search program, and in particular to a technology for searching or designing another ligand with reduced molecular weight from non-peptide natural products, peptides, proteins and other ligands. Background Art

[0002] Protein-protein interaction (PPI) is at the forefront of drug development. In fact, there are about 30,000 important and untreated PPIs. As molecules that control PPIs, antibodies and the like are being studied (can be antibodies or medium-molecular peptides; the same applies below), but even if they are active, they have low molecular weight and low membrane permeability, and the development of this frontier is limited in the case of intracellular targets or in the case of intracerebral or oral drugs for extracellular targets.

[0003] Therefore, in the development of this frontier, it is necessary to impart membrane permeability to antibodies, etc. Moreover, as a countermeasure for imparting membrane permeability, it has been proposed to reduce the molecular weight of antibodies, etc. (for example, refer to non-patent document 1). Reduced molecular weight refers to a method of searching or designing compounds that are generated as low molecular weight compounds of about 500 that are expected to have membrane permeability by discarding sites that are not necessary for the activity of antibodies, etc.

[0004] Several methods have been studied in the past to reduce the molecular weight of antibodies or the like or, more generally, to convert the skeleton of a compound into another compound. For example, in the peptide mimetic method (see Patent Documents 1 to 3), a non-peptide structure that reproduces the secondary structure (α-helix, β-sheet, etc.) of a protein is prepared in advance, and side chains important for interaction with the target in the antibody or the like are introduced into the non-peptide structure under the condition that they are spatially in the same position, thereby producing a low-molecular compound.

[0005] Furthermore, in the pharmacophore method (refer to non-patent document 2), only the important partial structures in the interaction with the target are retained in the antibody, etc., and they are abstracted according to the type of interaction (hydrogen bond donor, acceptor, hydrophobic interaction site, etc.), and search or design is performed under the condition of having the same type of interaction at the same position in space as them, thereby producing low molecular weight compounds.

[0006] Furthermore, in the SBDD method (SBDD: Structure-Based Drug Design; refer to Patent Document 4), the three-dimensional structure of the target obtained by structural analysis of the complex of the antibody or the target can be calculated through MD simulation (MD: Molecular Dynamics) based on the principle, and a search or design is performed based on the condition that the bonding energy is high, thereby producing a low molecular weight compound.

[0007] Furthermore, in the AAM method (AAM: Amino Acid Mapping; refer to Patent Document 5), for the compound, the AAM descriptor can be calculated by calculating the interaction with the amino acid based on the principle, and a search or design is performed based on the high similarity of the AAM descriptor related to the activity, thereby producing another compound.

[0008] Previous technical literature

[0009] Patent Literature

[0010] Patent Document 1: International Publication No. 2009 / 148192

[0011] Patent Document 2: International Publication No. 2010 / 044485

[0012] Patent Document 3: International Publication No. 2010 / 128685

[0013] Patent Document 4: International Publication No. 2006 / 099178

[0014] Patent Document 5: Japanese Patent No. 6826672

[0015] Non-patent literature

[0016] Non-patent document 1: “Constrained Peptides in Drug Discovery and Development”, Douglas R. Cary et al., Journal of Synthetic Organic Chemistry, Japan, Volume 75 (2017) Issue 11, [retrieved on July 29, 2022], Internet (https: / / www.jstage.jst.go.jp / article / yukigoseikyokaishi / 75 / 11 / 75_1171 / _pdf / -char / en)

[0017] Non-patent document 2: “Translating peptides into small molecules”, Gerd Hummel et al., Molecular BioSystems, 2006, 2, 499-508, [retrieved on July 29, 2022], Internet (https: / / pubs.rsc.org / en / content / articlehtml / 2006 / mb / b611791k) Summary of the invention

[0018] Technical issues to be solved by the invention

[0019] However, in the peptide mimetic method described in Patent Documents 1 to 3, the part of the antibody or the like including the residues important in the interaction with the target needs to be a secondary structure of the protein, and the structure of the applicable antibody or the like is limited. In the pharmacophore method described in Non-Patent Document 2, there is no restriction on the structure of the antibody or the like based on the peptide mimetic method, but the classification of the types of interaction that is a characteristic of this method needs to be set by humans in advance, and the classification of the types of interaction has uncertainty.

[0020] Furthermore, the SBDD method described in Patent Document 4 does not have the uncertainty of classification of the types of interactions based on the pharmacophore method, but in the calculation of the bonding energy in the MD simulation, it takes time to use the information of the target with a large molecular weight, and the chemical space of the compound to be searched and designed is limited. Furthermore, the AAM method described in Patent Document 5 does not have the limitation of the chemical space caused by the increase in calculation time derived from processing the target in the SBDD method, but there is a problem that the molecular weight is almost unchanged when the skeleton is converted.

[0021] Therefore, there is currently no method for systematically reducing the molecular weight of antibodies and the like, and development is required.

[0022] The present invention has been made in view of such circumstances, and an object of the present invention is to provide a compound search device, a compound search method, and a compound search program capable of efficiently searching for low-molecular compounds having activity and membrane permeability.

[0023] Means for solving technical problems

[0024] In order to achieve the above-mentioned purpose, the compound search device involved in the first mode of the present invention has a processor, and the processor performs the following processing: obtaining information representing the three-dimensional structure of the search object compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating the characteristic quantity, i.e., the first characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein becomes above a first threshold as the main part; calculating the part of the first characteristic quantity derived from the main part, i.e., the partial characteristic quantity; calculating the characteristic quantity, i.e., the second characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search object compound around the search object compound in three-dimensional space; calculating the similarity between the second characteristic quantity for the search object compound and the partial characteristic quantity of the ligand; and establishing a correspondence between the similarity for the search object compound and the information representing the molecular weight and / or activity of the search object compound and outputting it to a recording device and / or a display device.

[0025] According to the first method, a "partial feature quantity" is a feature quantity derived from a main part of a feature quantity derived from the entire ligand, and has a high probability of becoming a value less than the first feature quantity. Then, in the first method, the similarity between the "partial feature quantity" and the second feature quantity derived from the entire search object compound is calculated, thereby calculating a compound with a high similarity (above the second threshold value). A compound with a high similarity has a feature quantity (the second feature quantity) that "becomes a value smaller than the first feature quantity (or, has a high probability of becoming it)", and therefore has a high probability of being a lower molecule than the original ligand. In the first method, the similarity of this search object compound is associated with information representing the molecular weight and / or activity of the search object compound and output to a recording device and / or a display device.

[0026] As described above, low molecular weight is a method for imparting activity and membrane permeability, so according to the first embodiment, it is possible to effectively search for low molecular weight compounds having activity and membrane permeability. In the first embodiment, the processor can refer to the output information (information indicating the similarity of the search target compound, the molecular weight and / or activity of the search target compound) to extract compounds that meet the desired conditions from the search target compounds.

[0027] In the first embodiment and the following embodiments, the first feature value for the ligand may be described as the “overall feature value” of the ligand, and the second feature value for the search target compound may be described as the “overall feature value” of the search target compound.

[0028] In addition, in the first embodiment and the following embodiments, the processor may set the compound previously acquired and recorded in the recording medium as the "search target compound", or may search the compound recorded in the library at each processing. Furthermore, the processor may specifically specify (according to the user's operation or not according to the user's operation) the search target compound when performing the processing. Such a recording medium and library may be included in the compound search device involved in the first embodiment, or may be a recording medium and library outside the device.

[0029] Furthermore, in the first embodiment and the following embodiments, the "information indicating the stereostructure" may be the stereostructure itself, or may be information that can indirectly determine the stereostructure. For example, if a structural formula is obtained as the "information indicating the stereostructure", the stereostructure of the compound can be determined from the structural formula by various methods.

[0030] The compound search device according to the second embodiment is a device in which, in the first embodiment, the processor performs the following processing in determining the main part: obtaining, from a non-temporary and tangible recording medium, any one of the result of the calculation of the bonding energy of the complex structure of the ligand and the target protein, the information indicating the structural activity of the ligand and the ligand analog, and the first characteristic quantity for the local stable conformation of the ligand as the information of the bonding force; and determining the main part based on the obtained information. The second embodiment specifies one embodiment of a method for determining the main part.

[0031] As examples of the "non-transitory and tangible recording medium" in the second embodiment, various magneto-optical recording devices and semiconductor memories can be cited. In addition, the "non-transitory and tangible recording medium" does not include non-tangible recording media such as the carrier signal itself and the propagation signal itself.

[0032] The compound search device according to the third embodiment is a device in which, in the first or second embodiment, the processor performs the following processing in determining the main site: selecting one or more amino acid residues for which the ratio of the total value of the detailed bonding force caused by each amino acid residue of the ligand relative to the bonding force of the ligand is greater than a first threshold value, and determining the site of the one or more amino acid residues in the ligand as the main site. The third embodiment specifically specifies one embodiment of a method for determining the main site.

[0033] The compound search device according to the fourth aspect is characterized in that, in the third aspect, the processor selects one or more amino acid residues in consideration of the distance between the amino acid residues in specifying the main part. The fourth aspect specifies the method for specifying the main part in the third aspect in more detail.

[0034] In the compound search device according to the fifth embodiment, in the first or second embodiment, the processor performs the following processing in determining the main part: determining the part of the ligand present in the specified area as the main part. In the fifth embodiment, the processor may specify the area according to the user's operation or may not specify the area according to the user's operation.

[0035] The compound search device according to the sixth aspect is characterized in that, in the fifth aspect, the processor specifies the region based on the distribution of the first feature value around the ligand in the identification of the main part. In the sixth aspect, the processor can specify the region including the portion with the highest aggregation degree, for example.

[0036] The compound search device according to the seventh aspect is configured such that, in the fifth aspect, the processor performs the following processing in determining the main part: determining at least one of the number, position, size, and shape of the region according to the operation of the user via the input device. According to the seventh aspect, the user can specify at least one of the number, position, size, and shape of the region by himself.

[0037] The compound search device according to the eighth aspect is configured such that, in the fifth aspect, the processor performs the following processing in determining the main part: information representing the three-dimensional structure of the ligand, the distribution of the first characteristic value around the ligand, and the number, position, size, and shape of the regions are overlapped and displayed on the display device. According to the eighth aspect, the user can easily grasp the relationship between the aggregation degree and the region through the display.

[0038] The compound search device according to a ninth aspect is characterized in that in the first or second aspect, the processor causes the display device to display information indicating the ratio of the partial feature value to the first feature value. According to the ninth aspect, the user can easily understand the ratio of the partial feature value to the first feature value.

[0039] In the compound search device according to the tenth aspect, in the third aspect, the processor performs the following processing in determining the main part: in the three-dimensional structure of the ligand, the part of one or more amino acid residues determined as the main part is identified and displayed on the display device. According to the tenth aspect, the user can easily grasp the part of the amino acid residue determined as the main part.

[0040] In the compound search device according to the 11th aspect, in the 1st or 2nd aspect, the processor performs the following processing in determining the main part: calculating the details of each amino acid residue constituting the ligand for the bonding energy between the ligand and the target protein, and displaying the calculated details on the display device for each amino acid residue. According to the 11th aspect, the user can easily grasp the bonding energy based on each residue.

[0041] In the compound search device according to the twelfth embodiment, in the first or second embodiment, the processor performs the following processing: in the calculation of the partial feature quantity, the partial feature quantity of the ligand is translated and / or rotated to calculate a plurality of partial feature quantities; and in the calculation of the similarity, the similarity is calculated based on the cosine similarity or Euclidean distance between the plurality of partial feature quantities and the second feature quantity for the search target compound. The twelfth embodiment specifically specifies an embodiment of a method for calculating the similarity.

[0042] In the compound search device according to the 13th embodiment, in the 1st or 2nd embodiment, the processor performs the following processing in the output: extracting compounds with a similarity greater than or equal to the second threshold value from the search target compounds, and outputting information representing the extracted compounds to the recording device and / or the display device. In the 13th embodiment, the processor may extract a determined number of compounds in descending order of similarity, or may extract all compounds with a similarity greater than or equal to the second threshold value. The processor may determine the number of extracted compounds according to the user's operation.

[0043] In order to achieve the above-mentioned purpose, the compound search method involved in the 14th mode of the present invention is executed by a compound search device equipped with a processor, and the processor performs the following processing: obtaining information representing the three-dimensional structure of the search target compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating the characteristic quantity, i.e., the first characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein becomes above a first threshold value as the main part; calculating the part of the first characteristic quantity derived from the main part, i.e., the partial characteristic quantity; calculating the characteristic quantity, i.e., the second characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in three-dimensional space; calculating the similarity between the second characteristic quantity for the search target compound and the partial characteristic quantity of the ligand; and establishing a correspondence between the similarity for the search target compound and the information representing the molecular weight and / or activity of the search target compound and outputting it to a recording device and / or a display device.

[0044] According to the fourteenth aspect, a low-molecular compound having activity and membrane permeability can be efficiently searched for similarly to the first aspect. In addition, the compound search method according to the fourteenth aspect may have the same structure as the second to thirteenth aspects.

[0045] In order to achieve the above-mentioned purpose, the compound search program involved in the 15th mode of the present invention causes a compound search device equipped with a processor to execute a compound search method, in which the processor performs the following processing: obtaining information representing the three-dimensional structure of the search target compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating a feature quantity, i.e., a first feature quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein is greater than a first threshold value, as a main part; calculating a part of the first feature quantity derived from the main part, i.e., a partial feature quantity; calculating a feature quantity, i.e., a second feature quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in three-dimensional space; calculating the similarity between the second feature quantity for the search target compound and the partial feature quantity of the ligand; and establishing a correspondence between the similarity for the search target compound and the information representing the molecular weight and / or activity of the search target compound and outputting it to a recording device and / or a display device.

[0046] According to the fifteenth aspect, a low molecular weight compound having activity and membrane permeability can be efficiently searched for similarly to the first and fourteenth aspects. In addition, the compound search program according to the fifteenth aspect may have the same structure as the second to thirteenth aspects.

[0047] Furthermore, as the methods of the present invention, non-temporary and tangible recording media (e.g., various magneto-optical recording devices and semiconductor memories) recording computer-readable codes of the programs of these methods can also be cited. In addition, the above-mentioned "non-temporary and tangible recording media" do not include non-tangible recording media such as the carrier signal itself and the propagation signal itself.

[0048] Effects of the Invention

[0049] As described above, according to the compound search device, compound search method, and compound search program of the present invention, it is possible to efficiently search for low-molecular compounds having activity and membrane permeability. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a diagram showing the structure of a compound search device.

[0051] Figure 2 It is a diagram showing the functional structure of the processing unit.

[0052] Figure 3 This is a diagram showing information stored in the storage unit.

[0053] Figure 4 is a flowchart showing the processing of the compound search method.

[0054] Figure 5 is a flowchart showing a method for calculating a three-dimensional AAM descriptor.

[0055] Figure 6 This is a diagram showing a state where the structural formula is three-dimensionalized.

[0056] Figure 7 is a diagram showing an example of a three-dimensional AAM descriptor.

[0057] Figure 8 is another diagram showing an example of a three-dimensional AAM descriptor.

[0058] Fig. 9 This is a diagram showing a state in which main parts are determined.

[0059] Fig.10 This is a diagram showing an example of displaying the contribution rate of the interaction based on the selected residues.

[0060] Fig.11 It is a diagram showing a state of calculating a partial feature amount.

[0061] Fig.12 This is a diagram showing a state where the overall feature value of the search target compound is calculated.

[0062] Fig.13 This is a diagram showing a state in which the similarity of feature quantities is calculated.

[0063] Fig.14 This is a graph showing the names and molecular weights of search target compounds.

[0064] Fig.15 This is a graph showing the relationship between the similarity of feature quantities and activity.

[0065] Fig.16 This is a graph showing the similarity between the overall feature value of the ligand and the partial feature values ​​and overall feature value of the search target compound. DETAILED DESCRIPTION

[0066] Embodiments of the compound search device, compound search method, and compound search program according to the present invention will be described in detail. In the description, the drawings will be referred to as needed.

[0067] [First embodiment]

[0068] Figure 1This is a block diagram showing the structure of the compound search device 10 (compound search device) involved in the first embodiment. The compound search device 10 is a device for calculating feature quantities (first feature quantity, partial feature quantity, second feature quantity) for ligands and compounds (target structures) and extracting (screening) compounds based on the feature quantities, and can be implemented using a computer. Figure 1 As shown, the compound search device 10 includes a processing unit 100, a storage unit 200, a display unit 300 and an operating unit 400 (input device), and they are connected to each other to send and receive the required information. Regarding these components, various settings can be adopted. Each component can be set in one place (inside a frame, in a room, etc.), or it can be set in separate locations and connected via a network. In addition, the compound search device 10 is connected to an external server 500 and an external database 510 such as PDB (Protein Data Bank) via a network NW such as the Internet, and can obtain information such as the structural formula of the compound and the crystal structure of the protein as needed (for example, the results of X-ray eutectic structure analysis or NMR (Nuclear Magnetic Resonance) structure analysis).

[0069] [Structure of the processing unit]

[0070] Figure 2 1 is a diagram showing the structure of the processing unit 100. The processing unit 100 includes a processor 110 (processor), a ROM 130 (ROM: Read Only Memory), and a RAM 140 (RAM: Random Access Memory). The processor 110 has functions as a structural information acquisition unit 112, a conformational information acquisition unit 114, a main part determination unit 116, a feature quantity calculation unit 118, a partial feature quantity calculation unit 120, a similarity calculation unit 122, a compound extraction unit 124, an input / output control unit 126, and a communication control unit 128.

[0071] [Overview of the functions of the processing unit]

[0072] The structural information acquisition unit 112 acquires information representing the three-dimensional structure of the search target compound. The conformation information acquisition unit 114 acquires information representing the conformation of the ligand in the complex structure of the ligand and the target protein. The main part determination unit 116 determines a part of the ligand as the main part. The feature quantity calculation unit 118 calculates the feature quantity (first feature quantity) for the ligand and the search target compound. The partial feature quantity calculation unit 120 calculates the partial feature quantity of the ligand. The similarity calculation unit 122 calculates the similarity between the overall feature quantity (second feature quantity) for the search target compound and the partial feature quantity of the ligand. The compound extraction unit 124 extracts compounds whose similarity is above the second threshold value from the search target compound.

[0073] The input / output control unit 126 controls the input / output of information between the storage unit 200, the reception of user operations via the operation unit 400, and the output of information to the display unit 300. The communication control unit 128 cooperates with the above-mentioned units to control the communication with the external server 500 and / or the external database 510 and other devices via the network NW. The input / output control unit 126 and the communication control unit 128 can, for example, perform various calculations related to the structure and / or characteristics of the compound (based on methods such as MO, MD, MM, AlphaFold2, etc.) on a server or computer such as the external server 500, and obtain the results thereof (or only obtain the results that have been executed). In addition, the compound search device 10 itself may have the function of performing such calculations.

[0074] The functions of each part of the processing unit 100 (processor 110) can be implemented using various processors. The various processors include, for example, a general-purpose processor, namely a CPU (Central Processing Unit), which executes software (programs) to implement various functions. In addition, the various processors include programmable logic devices (PLDs) such as FPGAs (Field Programmable Gate Arrays) that can change the circuit structure after manufacturing. In addition, the various processors include dedicated circuits such as ASICs (Application Specific Integrated Circuits) that are processors with circuit structures specially designed to perform specific processing.

[0075] The functions of each part can be realized by one processor or by combining multiple processors. In addition, multiple functions can be realized by one processor. As an example of multiple functions being composed of one processor, first, there is the following method, such as represented by computers such as clients and servers: one processor is composed of a combination of one or more CPUs (Central Processing Units) and software, and multiple functions are realized by the processor. Second, there is the following method: as represented by a system on chip (System On Chip: SoC), a processor that realizes the functions of the entire system by one IC (Integrated Circuit) chip is used. In this way, various functions are constructed using one or more of the above-mentioned various processors as a hardware structure. Moreover, more specifically, the hardware structure of these various processors is a circuit (circuitry) formed by combining circuit elements such as semiconductor elements.

[0076] When the processor or circuit executes the software (program), the code that can be read by the processor (computer) of the software to be executed is pre-stored in a non-temporary and tangible recording medium such as ROM130, so that the processor refers to the software. The software stored in the non-temporary and tangible recording medium includes a program (compound search program) for executing the compound search method involved in the present invention. The code can also be recorded in a non-temporary and tangible recording medium such as various magneto-optical recording devices, semiconductor memories, etc. instead of in ROM130. When processing using software, for example, RAM140 can be used as a temporary storage area, and for example, a program and / or data stored in a non-volatile memory (non-temporary and tangible recording medium) such as EEPROM (Electrically Erasable and Programmable Read Only Memory) not shown in the figure, flash memory (Flash Memory) can be referred to.

[0077] The above-mentioned “non-transitory and tangible recording medium” does not include intangible recording media such as the carrier signal itself and the propagation signal itself.

[0078] The processing using each of these components of the processing unit 100 will be described in detail later.

[0079] [Structure of storage unit]

[0080] The storage unit 200 is composed of various non-temporary and tangible recording media such as magneto-optical recording media and semiconductor memory and its input / output control unit. In the storage unit 200 in this embodiment, for example, Figure 3As shown, structural information 202 (structural information of the search object compound), conformational information 204 (conformation information of the ligand), characteristic values ​​206 (overall characteristic value (first characteristic value) and partial characteristic values ​​of the ligand, overall characteristic value (second characteristic value) of the search object compound), search conditions 208 and search results 210 are stored in a corresponding relationship.

[0081] [Structure of display unit and operation unit]

[0082] The display unit 300 includes a display 310 (display device) and can display input information, information stored in the storage unit 200, processing results based on the processing unit 100, etc. The operation unit 400 (input device) includes a keyboard 410 and a mouse 420 as input devices and / or pointing devices, and the user can perform operations required for executing the compound search method and compound search program involved in the present invention via these devices and the screen of the display 310.

[0083] [Compound search method and compound search program processing]

[0084] Below, reference Figure 4 The compound search method and the processing of the compound search program in the compound search device 10 having the above-mentioned structure will be described with reference to flowcharts of FIG.

[0085] [Acquisition of structural information]

[0086] The structural information acquisition unit 112 (processor) acquires information representing the stereostructure of the search target compound (step S100: structural information acquisition processing, structural information acquisition step). The structural information acquisition unit 112 may acquire information representing the stereostructure from the structural information 202 (storage unit 200), the external server 500, or the external database 510, or may acquire the information through a user operation via the operation unit 400. Furthermore, the "information representing the stereostructure" may be the stereostructure itself, or may be information that can indirectly determine the stereostructure. The structural information acquisition unit 112 may, for example, acquire a structural formula as the "information representing the stereostructure" and determine the stereostructure of the compound based on the structural formula (see below). Figure 6 , 12 for example).

[0087] [Acquisition of conformational information]

[0088] The conformation information acquisition unit 114 (processor) acquires information (conformation information) indicating the conformation (bonding conformation) of the ligand in the complex structure of the ligand and the target protein (step S114: conformation information acquisition processing, conformation information acquisition step). Examples of conformation information include the following information.

[0089] (1) Results of X-ray cocrystal structure analysis or NMR (Nuclear Magnetic Resonance) structure analysis of the ligand and target protein

[0090] (2) Results of docking simulation of ligands with respect to the target protein structure alone obtained by X-ray co-crystal structure analysis or NMR structure analysis of the target protein with other ligands

[0091] (3) The results of the docking simulation of the ligand are compared with the results of the target protein structure alone obtained by X-ray crystal structure analysis, NMR structure analysis or folding calculation (based on MD, MM, AlphaFold2, etc.) of the target protein alone (MD: Molecular Dynamics, MM: Molecular Mechanics, AlphaFold2 is a protein structure calculation tool published on GitHub)

[0092] (4) Results of X-ray crystal structure analysis, NMR structure analysis, and local stable conformation search calculations (based on MD, MM, AlphaFold2, etc.) of the ligand alone

[0093] [Calculation of the overall characteristic quantity (first characteristic quantity) of the ligand]

[0094] The feature quantity calculation unit 118 (processor) calculates the first feature quantity, which is a feature quantity obtained by quantifying the degree of aggregation of one or more amino acids as probes for the ligand around the ligand in a three-dimensional space for the conformation of the ligand acquired in step S110 (step S120: first feature quantity calculation processing, first feature quantity calculation step). The first feature quantity is a feature quantity derived from the entire ligand, and is sometimes described as the "overall feature quantity" of the ligand below.

[0095] [Calculation method of the first feature value (three-dimensional AAM descriptor)]

[0096] The first feature quantity is the "three-dimensional AAM descriptor" (AAM: Amino Acid Mapping) described in the above-mentioned Patent Document 5, and the feature quantity calculation unit 118 (processor) can calculate the three-dimensional AAM descriptor by the method described in Patent Document 5. The calculation method of the three-dimensional AAM descriptor is described in detail below. In addition, in the present invention, including the matters described below, the entire contents of Patent Document 5 are cited for reference.

[0097] (Method 1)

[0098] The feature quantity calculation method of method 1 is executed by a feature quantity calculation device equipped with a processor, and the processor executes the following steps: an object structure specifying step, specifying an object structure composed of multiple unit structures with chemical properties; a three-dimensional structure generating step, generating a three-dimensional structure based on multiple unit structures for the object structure; and a feature quantity calculation step, calculating a feature quantity obtained by quantifying the degree of aggregation of more than one probe in the periphery of the three-dimensional structure in three-dimensional space, wherein the probe is a structure having a real charge and having multiple points where van der Waals forces are generated and separated.

[0099] (Method 2)

[0100] In method 2 in the above method 1, the processor performs the following processing: in the object structure designation step, a compound is designated as an object structure, in the stereostructure generation step, a stereostructure of the compound based on multiple atoms is generated, and in the feature quantity calculation step, a feature quantity, i.e., a three-dimensional AAM descriptor, is calculated in three-dimensional space by quantifying the degree of aggregation of amino acids serving as probes around the stereostructure of the compound generated in the stereostructure generation step.

[0101] In the above-mentioned method 2, the feature quantity calculation unit 118 (processor) can calculate the three-dimensional AAM descriptor of the ligand by specifying the ligand as the "object structure", and this "three-dimensional AAM descriptor" is equivalent to the "first feature quantity" in the present invention. The calculation method of the three-dimensional AAM descriptor is described in detail below. In addition, the feature quantity calculated by method 2 is sometimes described as a "three-dimensional AAM descriptor" or simply "AAM descriptor".

[0102] The compound search device 10 can calculate the three-dimensional AAM descriptor according to the instruction of the user via the operation unit 400 . Figure 5 3D AAM descriptor calculation process for a compound (target structure). Figure 5 In step S200, the structural information acquisition unit 112 inputs the structural formula (an example of information indicating a three-dimensional structure) of the compound according to the user's operation. Thus, the compound represented by the input chemical formula is designated as the target structure (target structure designation processing, target structure designation process).

[0103] The feature quantity calculation unit 118 converts the input structural formula into three-dimensional form and generates a three-dimensional structure of the compound based on multiple atoms (multiple unit structures having chemical properties) (step S210: three-dimensional structure generation processing, three-dimensional structure generation process). Various methods are known for converting the structural formula into three-dimensional form, and the method used in step S210 is not particularly limited. Figure 6 is an example of a three-dimensional structural formula, where part (a) represents the input structural formula and part (b) represents the three-dimensional structural formula (in addition, Figure 6 Part (a) and part (b) show the state when viewed from opposite sides).

[0104] The feature quantity calculation unit 118 calculates the spatial distribution ΔG of the free energy felt by each atom “μ” of the amino acid “a” (a is a number indicating the type of amino acid; 1 to 20) aμ (r) (Step S220; feature quantity calculation processing, feature quantity calculation step). aμ The calculation method of (r) can adopt the molecular dynamics method (MD: Molecular Dynamics), but is not limited to this. The amino acid for which the characteristic quantity is calculated can be a predetermined type, or it can be determined according to the instructions of the user via the operation unit 400 (as long as it is more than one type, it can also be a plurality of types). In addition, the amino acid for calculating the characteristic quantity can be selected by considering the classification. In addition, there are various classifications of amino acids, for example, the classification of "hydrophilic, hydrophobic, special" can be used, and other classifications such as "polar, non-polar, negatively charged, positively charged" or "alkaline, acidic, neutral" can also be used.

[0105] The feature quantity calculation unit 118 calculates the feature quantity based on ΔG. aμ (r) Calculate the distribution function g of each atom "μ" of amino acid "a" aμ (r) (Step S230: feature quantity calculation processing, feature quantity calculation step) If T is set to room temperature and K B Let it be the Boltzmann constant, then g aμ (r) is represented by the following formula (1).

[0106] [Formula 1]

[0107] g aμ (r) = exp(-ΔG aμ (r) / K B T) …(1)

[0108] The feature quantity calculation unit 118 calculates the distribution function g aμ (r) Calculate the distribution function g of the center of gravity of amino acids a (r) (Step S240: feature quantity calculation processing, feature quantity calculation step) In order to calculate, g aμ (r) Take the geometric mean of each atom "μ". The distribution function g a(r) is a three-dimensional AAM descriptor (first feature quantity) obtained by quantifying the degree of aggregation of one or more amino acids "a" around the three-dimensional structure of the compound in three-dimensional space. The feature quantity calculation unit 118 can establish a corresponding association between the calculated three-dimensional AAM descriptor and the structural information and / or structural information (information indicating the three-dimensional structure) of the compound and store it in the storage unit 200 (refer to Figure 3 ). In addition, ΔG aμ (r), g aμ (r) and g a "r" in (r) is a three-dimensional position vector (r = (x, y, z)).

[0109] Figure 7 It is for Figure 6 Examples of three-dimensional AAM descriptors for the compounds shown. Figure 7 Part (a) of represents the three-dimensional AAM descriptor for alanine, Figure 7 Part (b) of shows the three-dimensional AAM descriptor for valine. Figure 7 In the graph, darker regions are regions where the degree of aggregation (existence probability) of amino acids is high. Figure 8 It means for Figure 7 The table shows another example of the three-dimensional AAM descriptor of the compound shown in FIG. 1 , and shows the three-dimensional AAM descriptors for different directions 1, 2, and 3 (the first, second, and third sections of the table, respectively). The left column of the table shows the three-dimensional AAM descriptors (the contour surface relative to the threshold), and the right column shows the three-dimensional AAM descriptors (the contour surface relative to the threshold) and the three-dimensional structure of the compound.

[0110] [Determination of main parts]

[0111] The main site identification unit 116 (processor) identifies a part of the ligand, that is, a part of the ligand whose proportion of the bonding force with the target protein is greater than a first threshold value, as the main site ( Figure 4 Step S120: main part determination processing, main part determination step).

[0112] The main part determination unit 116 can determine the main part by, for example, any one of the following methods (1) to (3). The main part determination unit 116 can determine the main part without using the "three-dimensional AAM descriptor" as in the method (1) or the method (2), or can determine the main part by using the "three-dimensional AAM descriptor" as in the method (3).

[0113] (1) Calculation results of bonding energy using the complex structure of the ligand and the target protein (based on FMO, MO, MD, MM, etc.) (FMO method: Fragment Molecular Orbital method, MO method: Molecular Orbital method)

[0114] (2) Calculation results using structure-activity correlation of ligands and ligand analogs

[0115] (3) Determining the main site based on the three-dimensional AAM descriptor relative to the local stable conformation of the ligand alone (for example, detecting the point where the intensity of the three-dimensional AAM descriptor is the largest, and determining the site of the ligand near the point as the main site)

[0116] The calculation method (feature quantity calculation method) of the "three-dimensional AAM descriptor" in method (3) is as described above. In addition, a specific example of the main part (a method in which the bonding force caused by the amino acid residue is considered in method (1)) is described in detail in the "Examples" section.

[0117] In addition, the main part identification unit 116 may identify the main part by any method through the operation of the user via the operation unit 400, or may identify the main part without the operation of the user. Furthermore, the main part identification unit 116 may determine at least one of the number, position, size, and shape of the region for identifying the main part based on the operation of the user via the operation unit 400 (input device). In this case, the main part identification unit 116 can identify the part of the ligand present in the specified region as the main part.

[0118] Furthermore, the main part identification unit 116 may determine the first threshold value through a user operation via the operation unit 400 , or may determine the first threshold value without a user operation (for example, using a predetermined value).

[0119] [Calculation of some feature quantities]

[0120] The partial feature value calculation unit 120 (processor) calculates the partial feature value ( Figure 4 Step S140: partial feature amount calculation processing, partial feature amount calculation step). The partial feature amount calculation unit 120 can calculate the partial feature amount by, for example, the following method.

[0121] (1) For the point of the center of gravity of the main part of the ligand, a three-dimensional AAM descriptor cut out from a sphere centered at the point is used as a partial feature quantity.

[0122] (2) For the point of the center of gravity of the main part of the ligand, a three-dimensional AAM descriptor cut out from an arbitrary region (which can be divided into multiple regions) including the point is set as a partial feature amount.

[0123] (3) For any point near the main part of the ligand, a three-dimensional AAM descriptor cut out from a sphere centered at the point is used as a partial feature value.

[0124] (4) For any point near the main part of the ligand, a three-dimensional AAM descriptor cut out from any region (which can be divided into multiple regions) including the point is set as a partial feature value.

[0125] (5) The difference between the three-dimensional AAM descriptor of the ligand removed and the three-dimensional AAM descriptor of the virtual compound with the main part of the ligand removed is taken, and the point with the maximum intensity of the difference is detected. The three-dimensional AAM descriptor of the ligand cut out from a sphere centered at the point is set as a partial feature value.

[0126] (6) Take the difference between the three-dimensional AAM descriptor of the ligand removed and the three-dimensional AAM descriptor of the virtual compound with the main part of the ligand removed, detect the point with the largest intensity of the difference, and set the three-dimensional AAM descriptor of the ligand cut out from any area (which can be divided into multiple areas) including the point as a partial feature value.

[0127] In addition, the partial feature amount calculation unit 120 may determine to calculate the partial feature amount by any method through the operation of the user via the operation unit 400, or may determine it without the operation of the user. Furthermore, the “arbitrary point near the main part” and the like may be determined through the operation of the user via the operation unit 400, or may determine it without the operation of the user.

[0128] [Calculation of the overall feature quantity (second feature quantity) of the search candidate compound]

[0129] The feature quantity calculation unit 118 (processor) calculates a feature quantity, i.e., a second feature quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in a three-dimensional space (step S150: second feature quantity calculation processing, second feature quantity calculation step). The second feature quantity is a feature quantity derived from the entire search target compound, and is sometimes described as the "overall feature quantity" of the search target compound hereinafter.

[0130] In the feature quantity calculation unit 118, the overall feature quantity (second feature quantity) of the search candidate compound is calculated by (method 1) and (method 2) in the same manner as the overall feature quantity (first feature quantity) for the ligand as described above. That is, the overall feature quantity (second feature quantity) of the search candidate compound is a three-dimensional AAM descriptor of the search compound. In addition, the feature quantity calculation unit 118 preferably makes the number and combination of amino acids the same by calculating the overall feature quantity (first feature quantity) of the ligand and calculating the overall feature quantity (second feature quantity) of the search candidate compound. In addition, the feature quantity calculation unit 118 can calculate the first feature quantity and the second feature quantity in parallel.

[0131] [Calculation of similarity]

[0132] The similarity calculation unit 122 (processor) calculates the similarity between the overall feature value (second feature value) of the search target compound and the partial feature value of the ligand (step S160: similarity calculation processing, similarity calculation step). The similarity calculation unit 122 can calculate the similarity by, for example, the following method.

[0133] (1) For the most stable conformation obtained by the search calculation (MD, MM, AlphaFold2, etc.) of the most stable conformation of the implemented molecule (herein, the search target compound, the same below), the three-dimensional AAM descriptor (the overall feature value (the second feature value) of the search target compound) is calculated, and multiple partial feature values ​​of the ligand are calculated by translating the partial feature value of the ligand by an arbitrary distance and rotating it at an arbitrary angle, thereby calculating all cosine similarities (cosine similarity, the same below) between them, and calculating the maximum value of their cos similarities as "similarity".

[0134] (2) For the most stable conformation obtained by searching for the most stable conformation of the molecule (MD, MM, AlphaFold2, etc.), a three-dimensional AAM descriptor (the overall characteristic value (the second characteristic value) of the search target compound) is calculated, and multiple partial characteristic values ​​of the ligand produced by translating the partial characteristic value of the ligand by an arbitrary distance and rotating it by an arbitrary angle are calculated, thereby calculating all Euclid distances between them, and calculating the minimum value of their Euclid distances (Euclidean distance, the same below) as "similarity".

[0135] (3) For multiple local stable conformations obtained by performing search calculations (MD, MM, AlphaFold2, etc.) on the local stable conformations of the molecule, a three-dimensional AAM descriptor (the overall characteristic value (the second characteristic value) of the search object compound) is calculated, and multiple partial characteristic values ​​of the ligand produced by translating the partial characteristic value of the ligand at an arbitrary distance and rotating it at an arbitrary angle are calculated, thereby calculating all cos similarities between them, and calculating the maximum value of their cos similarities as the "similarity".

[0136] (4) For multiple local stable conformations obtained by performing search calculations (MD, MM, AlphaFold2, etc.) on the local stable conformations of the molecule, a three-dimensional AAM descriptor (the overall characteristic value (the second characteristic value) of the search object compound) is calculated, and multiple partial characteristic values ​​of the ligand produced by translating the partial characteristic value of the ligand by an arbitrary distance and rotating it by an arbitrary angle are calculated, thereby calculating all Euclid distances between them, and calculating the minimum value of their Euclid distances as "similarity".

[0137] In addition, the similarity calculation unit 122 may determine to calculate the similarity by any method according to the operation of the user via the operation unit 400 , or may determine it without the operation of the user.

[0138] [Output of the relationship between similarity and molecular weight / activity]

[0139] The input / output control unit 126 (processor) associates the similarity with the molecular weight and / or activity of the search target compound and outputs it to the recording device (storage unit 200, etc.) and / or the display device (display 310, etc.) (step S170: output process, output processing). The processor 110 can obtain the calculated information on the molecular weight and activity, or calculate it from the structural information of the search target compound.

[0140] Thus, according to the compound search device 10 (compound search device), the compound search method, and the compound search program involved in the first embodiment, it is possible to provide information for effectively extracting compounds that have been reduced in molecular weight. As described above, reduction in molecular weight is a scheme for imparting activity and membrane permeability, so according to the first embodiment, it is possible to effectively search for low molecular weight compounds that have activity and membrane permeability. Specific output examples will be described later in the examples (refer to Fig.15 , 16 wait).

[0141] [Extraction of Compounds]

[0142] Furthermore, the compound extraction unit 124 (processor) can refer to the calculated information (similarity, molecular weight, activity) to extract compounds that meet the desired conditions. For example, the compound extraction unit 124 can extract compounds with a similarity of more than the second threshold value from the search target compound (step S180: compound extraction processing, compound extraction process). The compound extraction unit 124 can extract a determined number of compounds from the compounds with a similarity of more than the second threshold value in descending order of similarity, or can extract all compounds with a similarity of more than the second threshold value. The compound extraction unit 124 can determine the number of compounds to be extracted according to the user's operation.

[0143] In addition, the compound extraction unit 124 may extract compounds by referring to information other than similarity (molecular weight, activity), and the user may extract desired compounds by referring to the information output to the storage unit 200 and / or the display 310 .

[0144] The search target compound having a characteristic quantity (second characteristic quantity) with high similarity to a partial characteristic quantity of the ligand (a value smaller than the value of the overall characteristic quantity (first characteristic quantity) or a high probability thereof) has a high probability of being a lower molecular weight than the original ligand. Therefore, according to the compound search device 10 (compound search device), compound search method and compound search program involved in the first embodiment, it is possible to effectively extract low-molecular-weight compounds. And, as described above, low-molecular-weighting is a scheme for imparting activity and membrane permeability, so according to the first embodiment, it is possible to effectively search for low-molecular-weight compounds having activity and membrane permeability.

[0145] [Example]

[0146] A specific example of compound search is described in detail.

[0147] [Acquisition of structural and conformational information, determination of key sites]

[0148] First, the conformation (= bonding conformation) and main parts of the ligand when forming a complex with the target protein are determined. In this example, the X-ray cocrystal structure (based on the structural information acquisition unit 112 and the conformational information acquisition unit 114; structural information acquisition processing / step, conformational information acquisition processing / step) of the peptide p53-TAD1 (an example of the ligand; hereinafter, sometimes simply described as "p53-TAD1") and the target protein MDMX (an example of the target protein; hereinafter, sometimes simply described as "MDMX") was used. Fig. 9 Part (a) is a diagram showing a complex of the target protein 600 (MDMX) and the ligand 610 (p53-TAD1).

[0149] For this X-ray eutectic structure, the bonding energy (bonding force) between p53-TAD1 and MDMX was calculated by FMO, and the details of each amino acid residue of p53-TAD1 relative to this bonding energy were obtained (based on the main site determination unit 116; main site determination processing, main site determination step). By selecting amino acid residues so that the ratio of the total value of the details accounts for 60% (first threshold), amino acid residues 19, 22, 23, and 26 were determined as the main sites (based on the main site determination unit 116; main site determination processing, main site determination step). Fig. 9 Part (b) of FIG. 1 is a diagram showing the details of the bonding energy (bonding force) for each amino acid residue. The main part determination unit 116 and the input / output control unit 126 can display such a diagram on the display 310 (display device) when determining the main part, so that the user can easily understand the details of the bonding energy for each amino acid residue. Fig. 9 In the horizontal axis of part (b), 19, 23, etc. are the numbers of amino acid residues, and "PHE", "TRP", etc. represent the original amino acids (PHE: Phenylal anine, TRP: Tryptophan or Tryptophane).

[0150] [Identification and display of main parts]

[0151] Fig. 9 Part (c) is a diagram showing the three-dimensional structure of the ligand 610, which shows the state in which the main part determination unit 116 and the input / output control unit 126 identify and display (specifically, change the color) the amino acid residues No. 19, 22, 23, and 26 (one or more amino acid residues) determined as the main part on the display 310 (display device). Through this identification display, the user of the compound search device 10 can easily grasp the position of the amino acid residue determined as the main part. In addition, the identification display can be performed not only by a change in color, but also by a combination of characters, numbers, graphics, symbols, or a change in color.

[0152] Fig.10 It means in Fig. 9 The diagram shown in part (b) of FIG. 1 is a diagram showing the details of the bonding energy in the amino acid residues identified as the main parts. Fig.10In the figure, the contribution rate of the bonding energy based on the selected residue is displayed in a color different from that of other residues. Similar to the case of the main part, the identification and display of the residue can be performed not only by a change in color, but also by characters, numbers, graphics, symbols, or a combination of these and color changes. The main part determination unit 116 and the input / output control unit 126 display such a graph on the display 310, so that the user can easily understand the relationship between the bonding energy of the residues constituting the main part and the bonding energy of the residues constituting the parts other than the main part. In addition, amino acid residues such as No. 28 and No. 18 also contribute greatly to the bonding energy, but in this embodiment, the main part determination unit 116 selects amino acid residues No. 19, 22, 23, and 26 (more than one amino acid residue) in consideration of the distance between the amino acid residues. In the case where the discreteness of the selected residues is high (in Fig. 9 , Fig.10 In the example, amino acid residues 28, 18, etc. are also selected). If the bonding energy (interaction) derived from these residues is to be realized through a single molecule (compound), the molecular weight of such a molecule may become larger, so it is preferred to select amino acid residues in consideration of the distance.

[0153] [Calculation of some feature quantities]

[0154] In the feature quantity calculation unit 118, the three-dimensional AAM descriptor (first feature quantity, overall feature quantity) is calculated for the bonding conformation of p53-TAD1 (ligand) by the above-mentioned method (first feature quantity calculation processing, first feature quantity calculation step), and the partial feature quantity calculation unit 120 determines the part (partial feature quantity) of the calculated three-dimensional AAM descriptor originating from the main part of p53-TAD1 (partial feature quantity calculation step, partial feature quantity calculation processing). Fig.11 It is a diagram showing a state of calculating a partial feature amount. Fig.11 Part (a) shows that in p53-TAD1 (ligand 610), the center of gravity of amino acid residues 19, 22, 23, and 26 identified as the main site is detected, and the radius is generated from the point The state of the ball 620 of (7 angstroms). The main part identification unit 116 and the partial feature quantity calculation unit 120 can calculate the partial feature quantity by cutting out the part included in the ball 620 in the three-dimensional AAM descriptor (the overall feature quantity, the first feature quantity). That is, the ball 620 represents the number, position, size and shape of the area for calculating the partial feature quantity.

[0155] In addition, the area from which the partial feature quantity is cut out is not limited to one, and the shape is not limited to a sphere. Also, the size of the area is not limited to The partial feature calculation unit 120 may determine at least one of the number, position, size, and shape of the region from which the partial feature is cut out according to a user operation via the operation unit 400 (input device), or may set the number, position, size, and shape without a user operation.

[0156] Fig.11 Part (b) shows a state where the feature quantity calculation unit 118, the partial feature quantity calculation unit 120 and the input / output control unit 126 (processor) overlap and display the three-dimensional structure of p53-TAD1 (ligand), the distribution of the three-dimensional AAM descriptor 630 (first feature quantity) around the ligand and the sphere 620 (the area where the partial feature quantity is calculated or cut out) on the display 310 (display device). And, Fig.11 (c) is a diagram showing the calculated partial feature quantity 640. Through these displays, the user can easily understand the relationship between the overall feature quantity of the ligand and the partial feature quantity calculation region and the distribution of the calculated partial feature quantities.

[0157] In addition, the feature quantity calculation unit 118, the partial feature quantity calculation unit 120 and the input / output control unit 126 can display information representing the ratio of the partial feature quantity to the overall feature quantity (the first feature quantity) on the display 310 (display device) and / or store it in the storage unit 200.

[0158] The partial feature quantity calculation unit 120 can calculate the distribution function g of the three-dimensional AAM descriptor by a The product of (r) and θa(r) representing the definition range of the partial feature quantity (for example, the sphere 620) is g a (r)θa(r) is the radius of the defined range. A function that is 1 inside the sphere and 0 outside the sphere. In addition, the subscript a represents the type of amino acid, and the independent variable r = (x, y, z) represents the position in three-dimensional space. When the number of types of amino acids is N, the subscript a is a = 1, 2, ..., N. In addition, as described above, the number of amino acids used as probes only needs to be 1 or more, but in this embodiment, all amino acids (20 types) are used as probes to calculate the overall feature value and partial feature value.

[0159] [Calculation of overall characteristic quantities of search target compounds]

[0160] Next, a three-dimensional AAM descriptor (overall feature quantity, second feature quantity) is calculated for the conformation of the molecule to be searched or designed, and the similarity with the partial feature quantity of the ligand is determined. Fig.12 is a diagram showing the state of calculating the overall characteristic value of the search target compound. Fig.12The structural formula of the search target compound (hereinafter sometimes referred to as "Compound A") exemplified in part (a) of Fig.12 The three-dimensional structure shown in part (b) of FIG. 1 is used to calculate its most stable conformation by MO calculation, and the three-dimensional AAM descriptor ( Fig.12 As described above, these processes can be performed by the structure information acquisition unit 112, the feature amount calculation unit 118, etc.

[0161] [Calculation of similarity]

[0162] The partial feature quantity calculation unit 120 translates and rotates the partial feature quantity of p53-TAD1 to generate a plurality of partial feature quantities of p53-TAD1. The partial feature quantity calculation unit 120 calculates the cos similarity between the overall feature quantity of compound A and multiple partial feature quantities of p53-TAD1, and adopts the largest cos similarity among their cos similarities as the final "similarity". Regarding the calculation of cos similarity, within the definition range of the partial feature quantity, i.e., the radius is carried out inside the ball. Fig.13 It is a diagram showing the state of similarity calculation. Fig.13 Part (a) shows that the cos similarity between the partial feature value 640 of p53-TAD1 and the overall feature value 650 of compound A is zero. Fig.13 Part (b) shows the state where the cos similarity is maximum (0.71).

[0163] In addition, the overall feature quantity of compound A and the partial feature quantity produced by translation and rotation were respectively calculated by G a (r), g a (Rr+s)θ a (Rr+s). Here, R is a rotation matrix represented by Euler angles α, β, γ, and s is a translation vector represented by X, Y, and Z (R = R (α, β, γ), s = (X, Y, Z)). In addition, the cosine similarity between the overall feature value of compound A and the partial feature value produced by translation and rotation can be expressed as a function of R and s by the following equations (2) to (6).

[0164] [Formula 2]

[0165]

[0166] [Formula 3]

[0167]

[0168] [Formula 4]

[0169]

[0170] [Formula 5]

[0171]

[0172] [Formula 6]

[0173]

[0174] Then, the final similarity (AAM similarity) is calculated by the following formula (7).

[0175] [Formula 7]

[0176]

[0177] In this example, the cos similarity was calculated, but as described above, the similarity based on the Euclid distance may also be calculated. Also, in this example, the similarity based on the most stable conformation of compound A was calculated, but the similarity based on the local stable conformation may also be calculated.

[0178] [result]

[0179] Calculate 8 compounds (equivalent to the search target compound of "Compound A" above; the name and molecular weight are shown in Fig.14 The three-dimensional AAM descriptors (overall feature quantity, second feature quantity) in the table (in the table) were compared with the AAM similarity of the partial feature quantity of p53-TAD1 (the above formula (7)), and the inhibitory activity IC of MD MX in the experiment was compared with the inhibitory activity IC of MD MX in the experiment. 50 The results are shown in Fig.15 In the figure, pIC 50 =-logIC 50 .like Fig.15 As shown in the figure, the compounds with higher AAM similarity have higher inhibitory activity (pIC 50 Since the molecular weight of a highly active compound is about 500 to 600, the molecular weight can be reduced by about 900 to 1000 compared to the molecular weight of p53-TAD1 of about 1500.

[0180] Based on this result, the compound extraction unit 124 can extract compounds (compound extraction process, compound extraction processing) whose similarity is greater than or equal to the second threshold value (for example, greater than or equal to 0.6, greater than or equal to 0.65, etc.), and the input / output control unit 126 can output information indicating the extracted compounds to the storage unit 200 (recording device) and / or the display 310 (display device). The compound extraction unit 124 can set the "second threshold value" according to the user's operation, or it may not be set by the user's operation.

[0181] Furthermore, the compound extraction unit 124 may extract compounds by referring to information other than similarity (molecular weight, activity). Furthermore, the user may extract desired compounds by referring to the information output to the storage unit 200 and / or the display 310 .

[0182] [Comparative Example]

[0183] In the above embodiment, the similarity between the "partial feature value" of the ligand and the overall feature value (second feature value) of the search target compound is calculated. However, the comparison result when the similarity between the "overall feature value (first feature value)" of the ligand and the overall feature value of the search target compound is taken into account is shown in FIG. Fig.16 . Fig.16 Part (a) is a table showing the name of the target compound (the same as in the above example), the activity value, the similarity with the partial characteristic quantity of the ligand, and the similarity with the overall characteristic quantity of the ligand, Fig.16 Part (b) shows the similarity of the characteristic quantity and the inhibitory activity (pIC 50 )’s relationship diagram.

[0184] like Fig.16 As shown, the compounds with high similarity to the partial characteristic quantity of the ligand (SJ212 to SJ298) also have high activity values, but the compounds with high similarity to the overall characteristic quantity (especially CHEMBL3650050, CHEMBL3950021) do not necessarily have high activity values. Therefore, from the results of the comparative example, it can be seen that in order to extract compounds with high activity values, it is effective to pay attention to the similarity between the overall characteristic quantity of the search object compound and the partial characteristic quantity of the ligand.

[0185] [About applicable objects of the present invention]

[0186] According to the compound search device, compound search method and compound search program of the present invention described above, low molecular weight compounds having activity and membrane permeability can be efficiently searched for. Furthermore, such low molecular weight compounds can be used as candidate compounds for new drugs.

[0187] In addition, the compound search device, compound search method and compound search program of the invention can also be applied to the composition of culture medium such as cells. In the field of cell culture medium, as a culture medium additive, protein FGF (Fibroblast Growth Factor (s): fibroblast growth factor) and the like are used. FGF promotes the growth of cells by bonding with FGFR (Fibroblast Growth Factor Receptor: fibroblast growth factor receptor) which is a receptor of this protein. However, FGF is expensive, so it is necessary to reduce costs, so the use of cheaper peptides or low molecular compounds to bond with FGFR to reproduce the cell growth promoting effect is being studied to replace the expensive FGF. However, it has not yet been found that such peptides or low molecular compounds can play a sufficient effect. By applying the low molecular weight technology of the present invention to FGF, there is the possibility of discovering peptides or low molecular compounds that bond with FGFR and reproduce the cell growth promoting effect.

[0188] As mentioned above, although embodiment of this invention was described, this invention is not limited to the said form, Various deformation|transformation is possible.

[0189] [Note]

[0190] The present invention includes the aspects described in the claims and also includes the following aspects.

[0191] (Note 1)

[0192] The compound search device involved in Appendix 1 has a processor, which performs the following processing: obtaining information representing the three-dimensional structure of the search object compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating the characteristic quantity, i.e., the first characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein becomes above a first threshold value as the main part; calculating the part of the first characteristic quantity derived from the main part, i.e., the partial characteristic quantity; calculating the characteristic quantity, i.e., the second characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search object compound around the search object compound in three-dimensional space; calculating the similarity between the second characteristic quantity for the search object compound and the partial characteristic quantity of the ligand; and establishing a correspondence between the similarity of the search object compound and the information representing the molecular weight and / or activity of the search object compound and outputting it to a recording device and / or a display device.

[0193] (Note 2)

[0194] In the compound search device involved in Appendix 2 in Appendix 1, the processor performs the following processing in determining the main part: obtaining the result of the bonding energy calculation using the complex structure of the ligand and the target protein, the information related to the structure activity of the ligand and the ligand analog, and any one of the first characteristic quantities for the local stable conformation of the ligand as the bonding force from a non-temporary and tangible recording medium; and determining the main part based on the acquired information.

[0195] (Note 3)

[0196] In the compound search device involved in Note 3, in Note 1 or Note 2, the processor performs the following processing in determining the main site: selects one or more amino acid residues for which the ratio of the total value of the details of the bonding force caused by each amino acid residue of the ligand relative to the bonding force of the ligand is greater than a specified value, and determines the site of one or more amino acid residues in the ligand as the main site.

[0197] (Note 4)

[0198] The compound search device according to Supplement 4 is characterized in that, in Supplement 3, the processor performs a process of selecting one or more amino acid residues in consideration of the distance between the amino acid residues in specifying the main part.

[0199] (Note 5)

[0200] The compound search device according to Supplement 5 is characterized in that, in any one of Supplements 1 to 4, the processor performs a process of specifying a site of a ligand present in a designated region as the main site in the identification of the main site.

[0201] (Note 6)

[0202] The compound search device according to Supplementary Note 6 is configured such that, in specifying the main part, the processor performs a process of specifying a region based on a distribution of the first feature value around the ligand.

[0203] (Note 7)

[0204] The compound search device according to Supplement 7 is characterized in that in Supplement 5, the processor performs the following process in specifying the main part: specifying at least one of the number, position, size and shape of the region according to the user's operation via the input device.

[0205] (Note 8)

[0206] In the compound search device involved in Note 8, in any one of Notes 5 to 7, the processor performs the following processing in determining the main part: information representing the three-dimensional structure of the ligand, the distribution of the first characteristic value around the ligand, and the number, position, size and shape of the regions are overlapped and displayed on the display device.

[0207] (Note 9)

[0208] The compound search device according to Supplement 9 is characterized in that, in any one of Supplements 1 to 8, the processor causes the display device to display information indicating a ratio of the partial feature amount to the first feature amount.

[0209] (Note 10)

[0210] The compound search device according to Supplement 10 is one of Supplements 3 to 9, wherein the processor performs the following processing in the identification of the main part: in the three-dimensional structure of the ligand, the part identification of one or more amino acid residues identified as the main part is displayed on the display device.

[0211] (Note 11)

[0212] In the compound search device involved in Note 11, in any one of Notes 1 to 10, the processor performs the following processing in determining the main part: calculating the details of each amino acid residue constituting the ligand that determines the bonding energy between the ligand and the target protein, and displaying the calculated details on the display device for each amino acid residue.

[0213] (Note 12)

[0214] In the compound search device involved in Note 12, in any one of Notes 1 to 11, the processor performs the following processing: in the calculation of partial feature quantities, the partial feature quantities of the ligand are translated and / or rotated to calculate multiple partial feature quantities; and in the calculation of similarity, the similarity is calculated based on the cosine similarity or Euclidean distance between the multiple partial feature quantities and the second feature quantity for the search object compound.

[0215] (Note 13)

[0216] In the compound search device according to Appendix 13, in any one of Appendixes 1 to 12, the processor performs the following processing in the output: extracting compounds having a similarity greater than a second threshold value from the search target compounds, and outputting information representing the extracted compounds to the recording device and / or the display device.

[0217] (Note 14)

[0218] The compound search method involved in Note 14 is executed by a compound search device equipped with a processor, and the processor performs the following processing: obtaining information representing the three-dimensional structure of the search object compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating the characteristic quantity, i.e., the first characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein is above a first threshold value as the main part; calculating the part of the first characteristic quantity derived from the main part, i.e., the partial characteristic quantity; calculating the characteristic quantity, i.e., the second characteristic quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search object compound around the search object compound in three-dimensional space; calculating the similarity between the second characteristic quantity for the search object compound and the partial characteristic quantity of the ligand; and establishing a correspondence between the similarity of the search object compound and the information representing the molecular weight and / or activity of the search object compound and outputting it to a recording device and / or a display device.

[0219] The compound search method according to Supplement 14 may have the same configuration as that of Supplements 2 to 13.

[0220] (Note 15)

[0221] The compound search program involved in Appendix 15 enables a compound search device equipped with a processor to execute a compound search method, in which the processor performs the following processing: obtaining information representing the three-dimensional structure of the search object compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; for the conformation of the ligand, calculating a feature quantity, i.e., a first feature quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in three-dimensional space; determining a part of the ligand, i.e., a part of the ligand whose proportion in the bonding force with the target protein is above a first threshold value, as a main part; calculating a part of the first feature quantity derived from the main part, i.e., a partial feature quantity; calculating a feature quantity, i.e., a second feature quantity, obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search object compound around the search object compound in three-dimensional space; calculating the similarity between the second feature quantity for the search object compound and the partial feature quantity of the ligand; and establishing a correspondence between the similarity of the search object compound and the information representing the molecular weight and / or activity of the search object compound and outputting it to a recording device and / or a display device.

[0222] The compound search program according to Supplement 15 may have the same configuration as that of Supplements 2 to 13 .

[0223] Furthermore, as the methods of the present invention, non-temporary and tangible recording media (e.g., various magneto-optical recording devices and semiconductor memories) recording computer-readable codes of the programs of these methods can also be cited. In addition, the above-mentioned "non-temporary and tangible recording media" do not include non-tangible recording media such as the carrier signal itself and the propagation signal itself.

[0224] Explanation of symbols

[0225] 10-compound search device, 100-processing unit, 110-processor, 112-structure information acquisition unit, 114-conformation information acquisition unit, 116-main part determination unit, 118-feature quantity calculation unit, 120-partial feature quantity calculation unit, 122-similarity calculation unit, 124-compound extraction unit, 126-input / output control unit, 128-communication control unit, 130-ROM, 140-RAM, 200-storage unit, 202-structure information, 204-conformation information, 20 6-feature quantity, 208-search condition, 210-search result, 300-display unit, 310-display, 400-operation unit, 410-keyboard, 420-mouse, 500-external server, 510-external database, 600-target protein, 610-ligand, 620-sphere, 630-three-dimensional AAM descriptor, 640-partial feature quantity, 650-overall feature quantity, MDMX-target protein, NW-network, S100 to S240-steps of the compound search method.

Claims

1. A compound search device comprising a processor, wherein: The processor performs the following processing: Acquiring information indicating the three-dimensional structure of the search target compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; For the conformation of the ligand, calculating a first feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in a three-dimensional space; determining a portion of the ligand, that is, a portion of the ligand whose proportion in the bonding force with the target protein is greater than a first threshold, as a main portion; calculating a partial feature quantity, which is a portion of the first feature quantity derived from the main part; calculating a second feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in a three-dimensional space; calculating the similarity between the second feature quantity of the search target compound and the partial feature quantity of the ligand; and The similarity with respect to the search target compound is associated with information indicating the molecular weight and / or activity of the search target compound and is output to a recording device and / or a display device.

2. The compound search device according to claim 1, wherein: The processor performs the following processing in determining the main part: acquiring, as the information on the bonding force, from a non-transitory and tangible recording medium any one of a result of bonding energy calculation using the complex structure of the ligand and the target protein, information indicating the structural activity relationship of the ligand and the ligand analog, and the first characteristic quantity for the local stable conformation of the ligand; and The main part is determined according to the acquired information.

3. The compound search device according to claim 1 or 2, wherein: The processor performs the following processing in determining the main part: Select one or more amino acid residues for which the ratio of the detailed total value of the bonding force caused by each amino acid residue of the ligand relative to the bonding force of the ligand is above the first threshold, and determine the site of the one or more amino acid residues in the ligand as the main site.

4. The compound search device according to claim 3, wherein: The processor performs the following processing in determining the main part: The one or more amino acid residues are selected in consideration of the distance between the amino acid residues.

5. The compound search device according to claim 1 or 2, wherein: The processor performs the following processing in determining the main part: The site of the ligand present in the designated region is determined as the main site.

6. The compound search device according to claim 5, wherein: The processor performs the following processing in determining the main part: The region is specified based on the distribution of the first feature value around the ligand.

7. The compound search device according to claim 5, wherein: The processor performs the following processing in determining the main part: At least one of the number, position, size, and shape of the regions is determined according to a user's operation via an input device.

8. The compound search device according to claim 5, wherein: The processor performs the following processing in determining the main part: Information indicating the three-dimensional structure of the ligand, distribution of the first characteristic value around the ligand, and the number, position, size, and shape of the regions are displayed in a superimposed manner on a display device.

9. The compound search device according to claim 1 or 2, wherein: The processor causes a display device to display information indicating a ratio of the partial feature amount to the first feature amount.

10. The compound search device according to claim 3, wherein: The processor performs the following processing in determining the main part: In the three-dimensional structure of the ligand, the site of the one or more amino acid residues determined as the main site is identified and displayed on a display device.

11. The compound search device according to claim 1 or 2, wherein: The processor performs the following processing in determining the main part: The details of each amino acid residue constituting the ligand of the bonding energy between the ligand and the target protein are calculated, and the calculated details are displayed on a display device for each amino acid residue.

12. The compound search device according to claim 1 or 2, wherein: The processor performs the following processing: In the calculation of the partial characteristic quantities, the partial characteristic quantities of the ligand are translated and / or rotated to calculate a plurality of partial characteristic quantities; and In the calculation of the similarity, the similarity is calculated based on cosine similarity or Euclidean distance between the plurality of partial feature quantities and the second feature quantity for the search target compound.

13. The compound search device according to claim 1 or 2, wherein: The processor performs the following processing on the output: Compounds having the similarity greater than or equal to a second threshold value are extracted from the search target compounds, and information indicating the extracted compounds is output to a recording device and / or a display device.

14. A compound search method, which is performed by a compound search device having a processor, wherein: The processor performs the following processing: Acquiring information indicating the three-dimensional structure of the search target compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; For the conformation of the ligand, calculating a first feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in a three-dimensional space; determining a portion of the ligand, that is, a portion of the ligand whose proportion in the bonding force with the target protein is greater than a first threshold, as a main portion; calculating a partial feature quantity, which is a portion of the first feature quantity derived from the main part; calculating a second feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in a three-dimensional space; calculating the similarity between the second feature quantity of the search target compound and the partial feature quantity of the ligand; and The similarity with respect to the search target compound is associated with information indicating the molecular weight and / or activity of the search target compound and is output to a recording device and / or a display device.

15. A compound search program, which causes a compound search device having a processor to execute a compound search method, wherein the processor performs the following processing: Acquiring information indicating the three-dimensional structure of the search target compound; obtaining information representing the conformation of the ligand in the complex structure of the ligand and the target protein; For the conformation of the ligand, calculating a first feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the ligand around the ligand in a three-dimensional space; determining a portion of the ligand, that is, a portion of the ligand whose proportion in the bonding force with the target protein is greater than a first threshold, as a main portion; calculating a partial feature quantity, which is a portion of the first feature quantity derived from the main part; calculating a second feature value obtained by quantifying the degree of aggregation of one or more amino acids serving as probes for the search target compound around the search target compound in a three-dimensional space; calculating the similarity between the second feature quantity of the search target compound and the partial feature quantity of the ligand; and The similarity with respect to the search target compound is associated with information indicating the molecular weight and / or activity of the search target compound and is output to a recording device and / or a display device. 16 . A recording medium which is non-transitory and computer-readable and has the program according to claim 15 recorded thereon.

Images