WO2025070104A1 - 情報処理装置、情報処理装置の作動方法、および情報処理装置の作動プログラム - Google Patents
情報処理装置、情報処理装置の作動方法、および情報処理装置の作動プログラム Download PDFInfo
- Publication number
- WO2025070104A1 WO2025070104A1 PCT/JP2024/032771 JP2024032771W WO2025070104A1 WO 2025070104 A1 WO2025070104 A1 WO 2025070104A1 JP 2024032771 W JP2024032771 W JP 2024032771W WO 2025070104 A1 WO2025070104 A1 WO 2025070104A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- information
- amino acid
- information processing
- processing device
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B15/00—ICT specially adapted for analysing two-dimensional [2D] or three-dimensional [3D] molecular structures, e.g. structural or functional relations or structure alignment
- G16B15/20—Protein or domain folding
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B15/00—ICT specially adapted for analysing two-dimensional [2D] or three-dimensional [3D] molecular structures, e.g. structural or functional relations or structure alignment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
- G16B40/20—Supervised data analysis
Definitions
- the technology disclosed herein relates to an information processing device, an operating method for an information processing device, and an operating program for an information processing device.
- biopharmaceuticals use proteins such as interferons and antibodies as active ingredients. These proteins such as antibodies have different property information that indicates their properties, such as the hydrogen ion exponent (pH (Potential Of Hydrogen) value) at which the highest activity is obtained or the quality is most stably maintained, depending on the type.
- pH Hydrogen
- JP 2021-158973 A describes a technology for predicting the thermal stability (whether the heat resistance temperature is higher than a set temperature, etc.) of a protein mutant that has an amino acid substitution mutation in the amino acid sequence using a machine learning model.
- JP 2021-158973 A in order to improve the accuracy of predicting thermal stability, feature quantities of amino acid residues in a local region of a protein mutant centered on the part of the amino acid substitution mutation (type and number of amino acid residues in the local region, size of amino acid residues in the local region, angle of the side chain, hydrophobicity, etc.) are selectively applied to the machine learning model.
- JP 2021-158973 A is targeted at protein mutants. Therefore, it cannot be used for general proteins that are not mutants.
- One embodiment of the technology disclosed herein provides an information processing device capable of predicting property information indicating the properties of a protein with high accuracy, an operating method of the information processing device, and an operating program of the information processing device.
- the information processing device disclosed herein includes a processor, which acquires structural information of an amino acid sequence specific to a protein and feature quantities of the amino acid residues that make up the amino acid sequence, sets weights for the feature quantities based on the structural information, and applies the weighted feature quantities to a machine learning model, thereby causing the machine learning model to predict property information that indicates the properties of the protein.
- the structural information is preferably segmented domain information that indicates functional segmented domains of amino acid residues.
- the protein is an antibody
- the segmented region includes a complementarity determining region
- the processor sets a weight of the feature of an amino acid residue whose segmented region indicated by the segmented region information is a complementarity determining region higher than a weight of the feature of an amino acid residue whose segmented region indicated by the segmented region information is not a complementarity determining region.
- the structural information is preferably three-dimensional structural information of an amino acid sequence.
- the three-dimensional structure information is preferably three-dimensional shape information that indicates the three-dimensional shape of the amino acid sequence.
- the three-dimensional shape includes a loop shape and a linear shape, and it is preferable that the processor sets a weight for the feature amount of an amino acid residue whose three-dimensional shape represented in the three-dimensional shape information is a loop shape higher than a weight for the feature amount of an amino acid residue whose three-dimensional shape represented in the three-dimensional shape information is a linear shape.
- the processor preferably performs preprocessing on the amino acid residue features or the machine learning model according to the weights.
- Pre-processing is preferably a process of calculating a weighted average based on the weights of the features of amino acid residues.
- Pre-processing is preferably a process of setting initial coefficients according to weights in the machine learning model.
- the property information is preferably the optimal hydrogen ion exponent of the protein.
- the protein is preferably an antibody.
- the method of operation of the information processing device disclosed herein includes acquiring structural information of an amino acid sequence specific to a protein and feature quantities of the amino acid residues that make up the amino acid sequence, setting weights for the feature quantities based on the structural information, and applying the weighted feature quantities to a machine learning model to cause the machine learning model to predict property information that indicates the properties of the protein.
- the operating program of the information processing device disclosed herein causes a computer to execute processes including obtaining structural information about an amino acid sequence specific to a protein and feature quantities of the amino acid residues that make up the amino acid sequence, setting weights for the feature quantities based on the structural information, and applying the weighted feature quantities to a machine learning model to cause the machine learning model to predict property information that indicates the properties of the protein.
- the technology disclosed herein can provide an information processing device capable of predicting property information indicating the properties of a protein with high accuracy, an operating method of the information processing device, and an operating program of the information processing device.
- FIG. 2 is a diagram illustrating an information processing server and a user terminal.
- FIG. 1 shows the basic structure of an antibody.
- FIG. 2 is a block diagram showing a computer constituting an information processing server and a user terminal.
- 2 is a block diagram showing a processing unit of a CPU of the information processing server;
- FIG. 13 is a diagram showing structure information.
- 11A and 11B are diagrams illustrating the process of a feature extraction unit and a group of feature amounts.
- FIG. 13 is a diagram showing setting auxiliary information.
- FIG. 13 is a diagram illustrating a process performed by a weight setting unit.
- FIG. 13 is a diagram showing a group of weighted feature amounts.
- FIG. 13 is a diagram illustrating an aggregate feature amount.
- FIG. 1 shows the basic structure of an antibody.
- FIG. 2 is a block diagram showing a computer constituting an information processing server and a user terminal.
- 2 is a block diagram showing a processing unit of a CPU of the information
- FIG. 13 is a diagram illustrating the process of a prediction unit.
- FIG. 13 is a diagram showing a neural network constituting a property information prediction model.
- 2 is a block diagram showing a processing unit of a CPU of a user terminal;
- FIG. 13 is a diagram showing an information input screen.
- FIG. 13 is a diagram showing a prediction result display screen.
- 13 is a flowchart showing a processing procedure of the information processing server.
- FIG. 13 is a diagram showing another example of the structure information.
- FIG. 11 is a diagram showing another example of setting auxiliary information.
- FIG. 13 is a diagram showing yet another example of structure information.
- FIG. 13 is a diagram showing yet another example of setting auxiliary information.
- FIG. 11 is a diagram illustrating a process of a pre-processing unit according to the second embodiment.
- an information processing server 10 is connected to a user terminal 11 via a network 12.
- the information processing server 10 is an example of an "information processing device" according to the technology of the present disclosure.
- the user terminal 11 is installed at a pharmaceutical company developing biopharmaceuticals, or at an organization contracted by a pharmaceutical company to develop biopharmaceuticals, i.e., a contract research organization (CRO).
- the user terminal 11 is operated by a user U who is involved in the development of biopharmaceuticals at the pharmaceutical company or contract research organization (hereinafter collectively referred to as a pharmaceutical facility).
- the network 12 is, for example, a wide area network (WAN) such as the Internet or a public communication network. Note that, although only one user terminal 11 is connected to the information processing server 10 in FIG. 1, multiple user terminals 11 at multiple pharmaceutical facilities are actually connected to the information processing server 10.
- WAN wide area network
- the user terminal 11 transmits a prediction request 15 to the information processing server 10.
- the prediction request 15 is a request for the information processing server 10 to predict property information 17 indicating the properties of an antibody 16, which is an active ingredient of a biopharmaceutical.
- the prediction request 15 includes amino acid sequence information 18 of the antibody 16.
- the amino acid sequence information 18 is identified through experiments, and is input by a user U operating an input device 30B (see FIG. 13) of the user terminal 11.
- the antibody 16 is an example of a "protein” according to the technology of the present disclosure.
- the prediction request 15 also includes a terminal ID (Identification Data) for uniquely identifying the user terminal 11 that transmitted the prediction request 15.
- Amino acid sequence information 18 describes the order of peptide bonds of the amino acid residues that make up antibody 16, from the amino terminus to the carboxyl terminus, using single-letter alphabetical abbreviations that represent the amino acid residues. Since antibody 16 consists of approximately 450 amino acid residues, amino acid sequence information 18 also consists of a string of approximately 450 alphabetical letters. For example, "E” is glutamic acid, “L” is leucine, and "G” is glycine. This sequence of amino acid residues is also called the primary structure.
- the information processing server 10 predicts the property information 17 based on the amino acid sequence information 18.
- the property information 17 is the hydrogen ion exponent of the storage solution of the antibody 16 in which the quality of the antibody 16 is most stably maintained, i.e., the optimal hydrogen ion exponent of the antibody 16.
- the information processing server 10 distributes the property information 17 to the user terminal 11 that sent the prediction request 15.
- the user terminal 11 displays the property information 17 on the display 29B (see FIG. 13) of the user terminal 11, and makes the property information 17 available for viewing by the user U.
- antibody 16 basically has four polypeptide chains, namely, two identical heavy chains HC (Heavy Chains) and two identical light chains LC (Light Chains).
- Antibody 16 has a structure in which these two heavy chains HC and two light chains LC are bound by disulfide bonds DB (Disulfide Bonds), and has a symmetrical Y shape.
- the heavy chain HC is composed of a variable domain VH (Variable domain, Heavy Chain) and constant domains CH (Constant domain, Heavy Chain) 1, CH2, and CH3.
- the light chain LC is composed of a variable domain VL (Variable domain, Light Chain) and constant domain CL (Constant domain, Light Chain).
- the constant domains CH1 and CH2 of the heavy chain HC are connected by a hinge region HR (Hinge Region).
- the variable domains VH and VL are collectively referred to as the variable region.
- the constant domains CH1 to CH3 and CL are collectively referred to as the constant region.
- the variable domains VH and VL, and the constant domains CH1 to CH3, and CL are examples of "segmented regions" according to the technology of this disclosure.
- the reception unit 40 receives various requests from the user terminal 11.
- the reception unit 40 receives a prediction request 15 from the user terminal 11.
- the prediction request 15 includes amino acid sequence information 18. Therefore, by receiving the prediction request 15, the reception unit 40 acquires the amino acid sequence information 18.
- the reception unit 40 outputs the amino acid sequence information 18 to the RW control unit 41.
- the reception unit 40 also outputs the terminal ID of the user terminal 11 (not shown) to the screen delivery control unit 47.
- the screen distribution control unit 47 controls the distribution of various screens to the user terminal 11. Specifically, the screen distribution control unit 47 distributes and outputs various screens to the user terminal 11 that is the sender of the various requests in the form of screen data for web distribution created by a markup language such as XML (Extensible Markup Language). At this time, the screen distribution control unit 47 identifies the user terminal 11 that is the sender of the various requests based on the terminal ID from the reception unit 40.
- the various screens include an information input screen 75 (see FIG. 14) for inputting amino acid sequence information 18, and a prediction result display screen 80 (see FIG. 15) for displaying property information 17. Note that other data description languages such as JSON (Javascript (registered trademark) Object Notation) may be used instead of XML.
- the segmented regions are either the variable domains VH and VL, the constant domains CH1 to CH3 and CL, or the complementarity determining regions CDR-H1 to CDR-H3 and CDR-L1 to CDR-L3 shown in Figure 2.
- the three-dimensional shape is either a linear shape or a loop shape.
- the structural information derivation unit 42 identifies the segmental region of the antibody 16 whose segmental region is unknown, for example, by multiple alignment with a collection of antibodies 16 whose segmental regions are known.
- the segmental region can be obtained by analyzing the antibody 16 using well-known techniques such as X-ray crystal structure analysis, or can be obtained from information in papers or databases such as the Protein Data Bank.
- the structural information derivation unit 42 also inputs the amino acid sequence information 18 into, for example, "AlphaFold2," a protein three-dimensional structure prediction program developed by DeepMind. It then causes "AlphaFold2" to output the three-dimensional coordinates of each amino acid residue, and identifies the three-dimensional shape based on the three-dimensional coordinates.
- "AlphaFold2” a protein three-dimensional structure prediction program developed by DeepMind. It then causes "AlphaFold2" to output the three-dimensional coordinates of each amino acid residue, and identifies the three-dimensional shape based on the three-dimensional coordinates.
- the feature extraction unit 43 inputs the amino acid sequence information 18 to the feature extraction model 36, which then outputs a group of features 51.
- the feature extraction model 36 is, for example, a machine learning model such as an encoder unit of a neural network or a support vector machine (SVM).
- SVM support vector machine
- the feature extraction model 36 may be trained in the information processing server 10 or in a device other than the information processing server 10. Furthermore, the feature extraction model 36 may be trained continuously even after operation.
- feature A there are multiple types of features: feature A, feature B, feature C, ..., feature ZZZ.
- Multiple types of features are registered for each amino acid residue constituting the amino acid sequence of antibody 16.
- ZA1 is registered as feature A
- ZB1 is registered as feature B
- ZC1 as feature C
- ZZZZ1 as feature ZZZ are registered.
- each amino acid residue can be represented by multiple types of features.
- Multiple types of features are, so to speak, multidimensional vector data.
- the types of features (the number of dimensions of the feature vector) range from several hundred to several thousand.
- a molecular dynamics (MD) method may be used to extract features.
- MD molecular dynamics
- the setting auxiliary information 37 is information in which the weight to be set for each combination of segmented region and three-dimensional shape is registered.
- the segmented regions are divided into two groups: complementarity determining regions CDR and regions other than complementarity determining regions CDR.
- the two groups are further divided according to whether the three-dimensional shape is a straight line or a loop.
- the weight is 2, and when the three-dimensional shape is a loop, the weight is 2.5.
- the weight is 1, and when the three-dimensional shape is a loop, the weight is 1.5.
- the weight is set higher than when the segmented region is a region other than complementarity determining regions CDR. Also, when the three-dimensional shape is a loop, the weight is set higher than when the three-dimensional shape is a straight line.
- the weight setting unit 44 refers to the setting auxiliary information 37 and sets a weight for each amino acid residue according to the combination of the segmented region and the three-dimensional shape of the structural information 50. More specifically, the weight setting unit 44 sets a higher weight for the feature amount of an amino acid residue whose segmented region is the complementarity determining region CDR than for the feature amount of an amino acid residue whose segmented region is not the complementarity determining region CDR. The weight setting unit 44 also sets a higher weight for the feature amount of an amino acid residue whose three-dimensional shape is a loop shape than for the feature amount of an amino acid residue whose three-dimensional shape is a linear shape. For example, the weight setting unit 44 sets a weight of 1 for the amino acid residue "Q" at No.
- the weight setting unit 44 sets a weight of 2.5 for the amino acid residue "S" at No. 100, since the segmented region is the complementarity determining region CDR-L1 and the three-dimensional shape is a loop shape.
- weighted feature set 51W is a feature set 51 to which a weight term has been added.
- the preprocessing unit 45 calculates a weighted average of the weights of each feature as a preprocessing before applying the feature to the property information prediction model 38, and sets the weighted average as the aggregated feature 52. Specifically, the preprocessing unit 45 calculates a weighted average ZA of feature A based on weights ZA1, ZA2, ZA3, ..., a weighted average ZB of feature B based on weights ZB1, ZB2, ZB3, ..., a weighted average ZC of feature C based on weights ZC1, ZC2, ZC3, ..., and a weighted average ZZZZ of feature ZZZ based on weights ZZZZ1, ZZZZ2, ZZZZ3, .... Then, the set of weighted averages ZA, ZB, ZC, ..., ZZZZ is output as the aggregated feature 52.
- the property information prediction model 38 is constructed by a neural network 60.
- the neural network 60 has an input layer 61, an intermediate layer (also called a hidden layer) 62, and an output layer 63.
- the input layer 61, intermediate layer 62, and output layer 63 each have a plurality of nodes ND.
- Coefficients indicating the strength of connection of each node ND are set between the node ND of the input layer 61 and the node ND of the intermediate layer 62, between the nodes ND in the intermediate layer 62, and between the node ND of the intermediate layer 62 and the node ND of the output layer 63.
- An appropriate activation function such as a linear function or a ReLu (Rectified Linear Unit) function, is set for the node ND of the output layer 63.
- the browser control unit 72 After the user U inputs the desired amino acid sequence information 18 in the input box 76, the user U selects the prediction button 77. When the prediction button 77 is selected, the browser control unit 72 generates a prediction request 15 including the amino acid sequence information 18 input in the input box 76, and transmits the generated prediction request 15 to the information processing server 10.
- a prediction result display screen 80 as shown in FIG. 15 is displayed on the display 29B under the control of the browser control unit 72.
- the prediction result display screen 80 displays the property information 17, or more precisely, a message representing the property information 17. In this way, the property information 17 is presented to the user U in the form of screen data distribution.
- An amino acid sequence information display button 81 is provided at the top of the prediction result display screen 80. When the amino acid sequence information display button 81 is selected, a pop-up display screen for the amino acid sequence information 18 is displayed.
- a save button 82 and an OK button 83 are provided at the bottom of the prediction result display screen 80.
- the save button 82 is selected, the property information 17 and amino acid sequence information 18 are stored in association with the storage 25B of the user terminal 11.
- the OK button 83 is selected, the display of the prediction result display screen 80 is cleared.
- the CPU 27A of the information processing server 10 functions as a reception unit 40, a RW control unit 41, a structural information derivation unit 42, a feature extraction unit 43, a weight setting unit 44, a preprocessing unit 45, a prediction unit 46, and a screen delivery control unit 47, as shown in FIG. 4.
- the CPU 27B of the user terminal 11 functions as a browser control unit 72, as shown in FIG. 13.
- the information input screen 75 shown in FIG. 14 is displayed on the display 29B of the user terminal 11.
- a prediction request 15 is sent from the browser control unit 72 to the information processing server 10.
- the prediction request 15 includes the amino acid sequence information 18 and the terminal ID of the user terminal 11, etc.
- the RW control unit 41 reads the amino acid sequence information 18 from the storage 25A (step ST120).
- the amino acid sequence information 18 is output from the RW control unit 41 to the structure information derivation unit 42 and the feature extraction unit 43.
- the structural information derivation unit 42 derives the structural information 50 shown in FIG. 5 based on the amino acid sequence information 18 (step ST130).
- the structural information 50 is output from the structural information derivation unit 42 to the weight setting unit 44.
- the amino acid sequence information 18 is input to the feature extraction model 36, and as a result, the feature extraction model 36 outputs a feature group 51 (step ST140).
- the feature extraction unit 43 outputs the feature group 51 to the weight setting unit 44.
- the weight setting unit 44 sets the weights of the features based on the structural information 50 while referring to the auxiliary setting information 37 (step ST150).
- the weight setting unit 44 generates a weighted feature set 51W including the set weight terms.
- the weighted feature set 51W is output from the weight setting unit 44 to the preprocessing unit 45.
- the CPU 27A of the information processing server 10 includes a structural information derivation unit 42, a feature extraction unit 43, a weight setting unit 44, and a prediction unit 46.
- the structural information derivation unit 42 obtains structural information 50 of the amino acid sequence unique to the antibody 16 by deriving it.
- the feature extraction unit 43 obtains structural information by extracting the features of the amino acid residues that make up the amino acid sequence of the antibody 16.
- the weight setting unit 44 sets weights for the features based on the structural information 50.
- the prediction unit 46 applies the weighted features to the property information prediction model 38, causing the property information prediction model 38 to predict property information 17 indicating the properties of the antibody 16.
- the structural information 50 is information that can be derived regardless of whether the antibody 16 is a mutant or not. For this reason, the technology disclosed herein is not limited to use only with mutants, as with the technology described in JP 2021-158973 A, but can be used generally not only with mutants, but also with general antibodies 16 that are not mutants.
- the structural information 50 is segmented region information that indicates functional segmented regions of amino acid residues.
- the protein is an antibody 16, and the segmented region includes a complementarity determining region (CDR).
- the weight setting unit 44 sets the weight of the feature amount of an amino acid residue whose segmented region is a complementarity determining region (CDR) higher than the weight of the feature amount of an amino acid residue whose segmented region is not a complementarity determining region (CDR).
- the complementarity determining region CDR is the binding site with the antigen, and is the region that determines the function of the antibody 16. Furthermore, the complementarity determining region CDR is generally exposed to a large amount of solvent, such as the storage solution of the antibody 16. For this reason, if the structural information 50 is set as segmented region information indicating a segmented region including the complementarity determining region CDR, and the weight of the feature amount of the amino acid residues in the segmented region that is the complementarity determining region CDR is set higher than the weight of the feature amount of the amino acid residues in the segmented region that is not the complementarity determining region CDR, it becomes possible to predict the property information 17 with even higher accuracy. This effect is particularly enhanced when the property information 17 is related to the storage solution of the antibody 16, such as the optimal hydrogen ion exponent of the antibody 16.
- the structural information 50 is three-dimensional structural information of the amino acid sequence
- the three-dimensional structural information is three-dimensional shape information that indicates the three-dimensional shape of the amino acid sequence.
- the three-dimensional shape includes a loop shape and a linear shape.
- the weight setting unit 44 sets the weight of the feature amount of an amino acid residue whose three-dimensional shape is a loop shape higher than the weight of the feature amount of an amino acid residue whose three-dimensional shape is a linear shape.
- the structural information 50 is three-dimensional shape information that indicates the three-dimensional shape of the amino acid sequence including loop and linear shapes, and the weight of the feature amount of the amino acid residue whose three-dimensional shape is a loop shape is set higher than the weight of the feature amount of the amino acid residue whose three-dimensional shape is a linear shape, it becomes possible to predict the property information 17 with even higher accuracy. This effect is particularly enhanced when the property information 17 is related to the storage solution of the antibody 16, such as the optimal hydrogen ion exponent of the antibody 16.
- the preprocessing unit 45 performs preprocessing on the features of the amino acid residues according to the weights. Specifically, as shown in FIG. 10, the preprocessing is a process of calculating a weighted average of the weights of the features of the amino acid residues. Therefore, the features reflecting the weights can be applied to the property information prediction model 38.
- the optimal hydrogen ion exponent is extremely important for stably maintaining the quality of the antibody 16. Therefore, if the property information 17 is set to the optimal hydrogen ion exponent of the antibody 16, it will contribute to stably maintaining the quality of the antibody 16, and as a result, it will be possible for the efficacy of a biopharmaceutical containing the antibody 16 as an active ingredient to be fully exerted.
- Biopharmaceuticals that contain antibody 16 as a protein are called antibody drugs, and are widely used to treat chronic diseases such as cancer, diabetes, and rheumatoid arthritis, as well as rare diseases such as hemophilia and Crohn's disease. Therefore, this example, in which the protein is antibody 16, can contribute to the stable supply of antibody drugs that are widely used to treat a variety of diseases.
- the three-dimensional structure information is not limited to the three-dimensional shape information shown in the example. As an example, it may be information on the three-dimensional coordinates of each amino acid residue and the centroid coordinates of the antibody 16, as shown in the structural information 90 in FIG. 17.
- the three-dimensional coordinates of each amino acid residue and the centroid coordinates of the antibody 16 are obtained by inputting the amino acid sequence information 18 into "AlphaFold2".
- the weight setting unit 44 sets the weights by referring to the setting auxiliary information 92 shown in FIG. 18 as an example.
- the setting auxiliary information 92 has a term for the distance from the center of gravity instead of the term for the three-dimensional shape of the setting auxiliary information 37.
- the distance from the center of gravity can be calculated from the three-dimensional coordinates of each amino acid residue and the coordinates of the center of gravity of the antibody 16.
- the segmented regions are divided into two groups, namely, the complementarity determining region CDR and the region other than the complementarity determining region CDR. The two groups are further divided according to whether the distance from the center of gravity is less than the threshold or not less than the threshold.
- the weight When the segmented region is the complementarity determining region CDR and the distance from the center of gravity is less than the threshold, the weight is 2, and when the distance from the center of gravity is not less than the threshold, the weight is 2.5.
- the weight When the segmented region is a region other than the complementarity determining region CDR and the distance from the center of gravity is less than the threshold, the weight is 1, and when the distance is not less than the threshold, the weight is 1.5.
- the weight is set higher than when the distance from the center of gravity is less than the threshold.
- the three-dimensional structure information is information on the three-dimensional coordinates of each amino acid residue and the center of gravity coordinates of antibody 16, and the weight of the feature amount of amino acid residues whose distance from the center of gravity is equal to or greater than a threshold value is set higher than the weight of the feature amount of amino acid residues whose distance from the center of gravity is less than the threshold value, it becomes possible to predict property information 17 with even higher accuracy. This effect is particularly enhanced when property information 17 is related to a storage solution for antibody 16, such as the optimal hydrogen ion exponent for antibody 16.
- the three-dimensional structure information may be, for example, information on the solvent accessible surface area (SASA) of each amino acid residue, as shown in structural information 95 in FIG. 19.
- SASA solvent accessible surface area
- the solvent accessible surface area is derived based on the three-dimensional coordinates of each amino acid residue.
- the weight setting unit 44 sets the weights by referring to the auxiliary setting information 97 shown in FIG. 20 as an example.
- the auxiliary setting information 97 sets the weight to 0 when the solvent-exposed surface area is less than a threshold value, and the weight to 1 when the solvent-exposed surface area is equal to or greater than the threshold value.
- the feature quantities of amino acid residues for which the weight is set to 0 are not added when calculating the weighted average in preprocessing.
- the weights set by the weight setting unit 44 may include 0.
- the preprocessing unit 100 of the second embodiment performs preprocessing according to the weight on the property information prediction model 101 before learning, and sets the property information prediction model 101 as a preprocessed property information prediction model 101 A.
- each feature of the aggregated feature 52 is input to each node ND of the input layer 61, but in the property information prediction model 101 of the second embodiment, each feature of the feature group 51 is input to each node ND of the input layer 61.
- the preprocessing unit 100 sets initial coefficients between each node ND of the pre-learned property information prediction model 101 according to the weights. Specifically, in order to give more importance to features with relatively high weights, the preprocessing unit 100 sets the initial coefficients between nodes ND related to features with relatively high weights higher than the coefficients between nodes ND related to features with relatively low weights.
- the preprocessed property information prediction model 101A with the initial coefficients set in this way is trained to obtain the property information prediction model 101X used in the prediction unit 46.
- the preprocessing is a process of setting initial coefficients according to the weights in the property information prediction model 101. Therefore, the learning of the preprocessed property information prediction model 101A can be converged with a small amount of training data, and a property information prediction model 101X with a relatively high prediction accuracy can be easily obtained.
- Feature quantities of amino acid residues may include spatial aggregation propensity (SAP) and spatial charge map (SCM), etc.
- SAP spatial aggregation propensity
- SCM spatial charge map
- the property information 17 is not limited to the optimal hydrogen ion exponent shown as an example. It may be the viscosity of a storage solution of the antibody 16 in which the quality of the antibody 16 is most stably maintained, the type of additive added to the storage solution such as L-arginine hydrochloride or refined sucrose, or the concentration of the additive. It may also be an index indicating the tendency of the antibody 16 to aggregate, an index indicating the hydrophobicity of the antibody 16, the presence or absence of toxicity of the antibody 16, etc.
- the protein is not limited to the illustrated antibody 16. It may be a peptide, a nucleic acid, etc. It may also be a cytokine (interferon, interleukin, etc.), a hormone (insulin, glucagon, follicle-stimulating hormone, erythropoietin, etc.), a growth factor (IGF (Insulin-Like Growth Factor)-1, bFGF (Basic Fibroblast Growth Factor), etc.), a blood coagulation factor (Factor 7, Factor 8, Factor 9, etc.), an enzyme (lysosomal enzyme, DNA (Deoxyribonucleic Acid) degrading enzyme, etc.), an Fc (Fragment Crystallizable) fusion protein, a receptor, albumin, or a protein vaccine. Antibodies also include bispecific antibodies, antibody-drug conjugates, low molecular weight antibodies, and glycosylated antibodies.
- the property information 17 itself may be delivered from the information processing server 10 to the user terminal 11.
- the manner in which the property information 17 is made available for viewing by the user U is not limited to the exemplary prediction result display screen 80.
- a printed copy of the property information 17 may be provided to the user U, or an email with the property information 17 attached may be sent to the user U's mobile device.
- the information processing server 10 may be installed in each pharmaceutical facility, or in a data center independent of the pharmaceutical facility.
- the user terminal 11 may assume some or all of the functions of each of the processing units 40 to 47 of the information processing server 10.
- the hardware configuration of the computer constituting the information processing server 10 can be modified in various ways.
- the information processing server 10 can be composed of multiple computers separated as hardware in order to improve processing power and reliability.
- the functions of the reception unit 40, RW control unit 41, structural information derivation unit 42, and feature extraction unit 43, and the functions of the weight setting unit 44, preprocessing unit 45, prediction unit 46, and screen distribution control unit 47 are distributed and assigned to two computers.
- the information processing server 10 is composed of two computers.
- the hardware configuration of the computer of the information processing server 10 can be changed as appropriate according to the required performance such as processing power, safety, and reliability. Furthermore, it is of course possible to duplicate not only the hardware but also APs such as the operating program 35 and the prediction AP 70, or to store them in a distributed manner across multiple storage devices, in order to ensure safety and reliability.
- the hardware structure of the processing unit that performs various processes, such as the reception unit 40, RW control unit 41, structural information derivation unit 42, feature extraction unit 43, weight setting unit 44, preprocessing unit 45 and 100, prediction unit 46, screen delivery control unit 47, and browser control unit 72, can use the various processors shown below.
- the various processors include the CPUs 27A and 27B, which are general-purpose processors that execute software (operation program 35 and prediction AP 70) and function as various processing units, as well as programmable logic devices (PLDs) such as FPGAs (Field Programmable Gate Arrays) that are processors whose circuit configuration can be changed after manufacture, and dedicated electrical circuits such as ASICs (Application Specific Integrated Circuits), which are processors with circuit configurations designed specifically to execute specific processes.
- PLDs programmable logic devices
- FPGAs Field Programmable Gate Arrays
- ASICs Application Specific Integrated Circuits
- Examples of configuring multiple processing units with a single processor include, first, a form in which one processor is configured with a combination of one or more CPUs and software, as typified by client and server computers, and this processor functions as multiple processing units. Second, a form in which a processor is used to realize the functions of the entire system, including multiple processing units, with a single IC (Integrated Circuit) chip, as typified by system-on-chip (SoC). In this way, the various processing units are configured as a hardware structure using one or more of the various processors listed above.
- SoC system-on-chip
- the hardware structure of these various processors can be an electrical circuit that combines circuit elements such as semiconductor elements.
- the information processing device is three-dimensional shape information that indicates a three-dimensional shape of the amino acid sequence.
- the three-dimensional shape includes a loop shape and a linear shape
- the processor An information processing device according to claim 5, further comprising setting a weight of the feature of an amino acid residue whose three-dimensional shape represented in the three-dimensional shape information is the loop shape higher than a weight of the feature of an amino acid residue whose three-dimensional shape represented in the three-dimensional shape information is the linear shape.
- the processor, 7. The information processing device according to any one of claims 1 to 6, wherein preprocessing is performed on the features of the amino acid residues or the machine learning model according to the weights.
- the preprocessing is a process of calculating a weighted average of the feature amounts of the amino acid residues using the weights.
- the preprocessing is a process of setting initial coefficients according to the weights to the machine learning model.
- the property information is an optimal hydrogen ion exponent of the protein.
- the protein is an antibody.
- the technology disclosed herein can be appropriately combined with the various embodiments and/or various modified examples described above. Furthermore, it is not limited to the above-mentioned embodiments, and various configurations can be adopted without departing from the gist of the technology. Furthermore, the technology disclosed herein extends to not only programs, but also storage media that non-temporarily store programs, and computer program products that include programs.
- a and/or B is synonymous with “at least one of A and B.”
- a and/or B means that it may be only A, only B, or a combination of A and B.
- the same idea as “A and/or B” is also applied when three or more things are expressed by connecting them with “and/or.”
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biotechnology (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Crystallography & Structural Chemistry (AREA)
- Bioethics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Peptides Or Proteins (AREA)
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202480055463.XA CN121794754A (zh) | 2023-09-28 | 2024-09-12 | 信息处理装置、信息处理装置的运行方法及信息处理装置的运行程序 |
| JP2025548767A JPWO2025070104A1 (cg-RX-API-DMAC7.html) | 2023-09-28 | 2024-09-12 | |
| EP24871899.1A EP4760727A1 (en) | 2023-09-28 | 2024-09-12 | Information processing device, operation method for information processing device, and operation program for information processing device |
| US19/551,423 US20260196292A1 (en) | 2023-09-28 | 2026-02-26 | Information processing apparatus, operation method of information processing apparatus, and operation program of information processing apparatus |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2023-167582 | 2023-09-28 | ||
| JP2023167582 | 2023-09-28 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US19/551,423 Continuation US20260196292A1 (en) | 2023-09-28 | 2026-02-26 | Information processing apparatus, operation method of information processing apparatus, and operation program of information processing apparatus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2025070104A1 true WO2025070104A1 (ja) | 2025-04-03 |
Family
ID=95204085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2024/032771 Ceased WO2025070104A1 (ja) | 2023-09-28 | 2024-09-12 | 情報処理装置、情報処理装置の作動方法、および情報処理装置の作動プログラム |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20260196292A1 (cg-RX-API-DMAC7.html) |
| EP (1) | EP4760727A1 (cg-RX-API-DMAC7.html) |
| JP (1) | JPWO2025070104A1 (cg-RX-API-DMAC7.html) |
| CN (1) | CN121794754A (cg-RX-API-DMAC7.html) |
| WO (1) | WO2025070104A1 (cg-RX-API-DMAC7.html) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2020077206A (ja) * | 2018-11-07 | 2020-05-21 | 合同会社みらか中央研究所 | 情報処理方法および学習モデル |
| JP2021158973A (ja) | 2020-03-31 | 2021-10-11 | 株式会社豊田中央研究所 | 変異体の熱的安定性を予測するためのモデル及びその利用 |
| JP2022527381A (ja) * | 2019-04-09 | 2022-06-01 | エーテーハー チューリッヒ | 抗体を分類するためのシステムおよび方法 |
-
2024
- 2024-09-12 JP JP2025548767A patent/JPWO2025070104A1/ja active Pending
- 2024-09-12 EP EP24871899.1A patent/EP4760727A1/en active Pending
- 2024-09-12 CN CN202480055463.XA patent/CN121794754A/zh active Pending
- 2024-09-12 WO PCT/JP2024/032771 patent/WO2025070104A1/ja not_active Ceased
-
2026
- 2026-02-26 US US19/551,423 patent/US20260196292A1/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2020077206A (ja) * | 2018-11-07 | 2020-05-21 | 合同会社みらか中央研究所 | 情報処理方法および学習モデル |
| JP2022527381A (ja) * | 2019-04-09 | 2022-06-01 | エーテーハー チューリッヒ | 抗体を分類するためのシステムおよび方法 |
| JP2021158973A (ja) | 2020-03-31 | 2021-10-11 | 株式会社豊田中央研究所 | 変異体の熱的安定性を予測するためのモデル及びその利用 |
Non-Patent Citations (1)
| Title |
|---|
| SAITO YUTAKA: "Protein function prediction and design by deep learning", SEIBUTSU KOGAKU KAISHI = JOURNAL OF THE SOCIETY FOR BIOSCIENCE AND BIOENGINEERING, JAPAN, NIPPON SEIBUTSU KOGAKKAI, SUITA,, JP, vol. 100, no. 11, 1 January 2022 (2022-01-01), JP , pages 589 - 592, XP093296845, ISSN: 0919-3758, DOI: 10.34565/seibutsukogaku.100.11_589 * |
Also Published As
| Publication number | Publication date |
|---|---|
| JPWO2025070104A1 (cg-RX-API-DMAC7.html) | 2025-04-03 |
| US20260196292A1 (en) | 2026-07-09 |
| EP4760727A1 (en) | 2026-06-17 |
| CN121794754A (zh) | 2026-04-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Heo et al. | Experimental accuracy in protein structure refinement via molecular dynamics simulations | |
| Fersht | On the simulation of protein folding by short time scale molecular dynamics and distributed computing | |
| Bublil et al. | Stepwise prediction of conformational discontinuous B‐cell epitopes using the Mapitope algorithm | |
| JP5589163B2 (ja) | モバイル通信設備へのコンテンツの提供 | |
| King et al. | Removing T-cell epitopes with computational protein design | |
| Shakhnovich et al. | Immunogenicity in clinical practice and drug development: when is it significant? | |
| Thrift et al. | Graph-pMHC: graph neural network approach to MHC class II peptide presentation and antibody immunogenicity | |
| Barnett et al. | Seroprevalence of measles, rubella, and varicella in refugees | |
| Doki et al. | Assessing potential drug–drug interactions between dabigatran etexilate and a P‐glycoprotein inhibitor in renal impairment populations using physiologically based pharmacokinetic modeling | |
| Dillon et al. | Predicting the long-term stability of biologics with short-term data | |
| Coutant et al. | Therapeutic protein drug interactions: a white paper from the international consortium for innovation and quality in pharmaceutical development | |
| Sheridan | Pharma consolidates its grip on post-antibody landscape. | |
| Waibel et al. | Bayesian optimization for efficient multiobjective formulation development of biologics | |
| McGovern et al. | The pandemic and resilience for the future: AccBio 2021 | |
| Wang et al. | COVID-19 and preparing planetary health for future ecological crises: Hopes from glycomics for vaccine innovation | |
| Słyk et al. | Minimal coarse-grained model for immunoglobulin G: Diffusion and binding under crowding | |
| Bertrand et al. | Antibody response to SARS-CoV-2 mRNA BNT162b2 vaccine in kidney transplant recipients and in-centre and satellite centre haemodialysis patients | |
| Carrasco-Triguero | Insights on the immunogenicity of antibody–drug conjugates | |
| Ruzmetov et al. | A Binding Kinetics Toolkit for Analyzing Transient Molecular Conformations and Computing Free Energy Landscapes | |
| EP4524975A1 (en) | Pharmaceutical assistance device, method of operating pharmaceutical assistance device, and program for operating pharmaceutical assistance device | |
| Leach | Regulatory Forum Opinion Piece* Differences between Protein-based Biologic Products (Biotherapeutics) and Chemical Entities (Small Molecules) of Relevance to the Toxicologic Pathologist | |
| Jarab et al. | Biosimilars and immunogenicity: a matter of concern? | |
| Chamberlain et al. | Immunogenicity assessment of biosimilars: a multidisciplinary perspective | |
| Goncalves et al. | SB5 shows cross-immunogenicity to adalimumab but not infliximab: results in patients with inflammatory bowel disease or rheumatoid arthritis | |
| Jiang et al. | Technical decision making with higher order structure data: higher order structure characterization during protein therapeutic candidate screening |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 24871899 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2025548767 Country of ref document: JP Kind code of ref document: A |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2025548767 Country of ref document: JP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2024871899 Country of ref document: EP |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |