Title of invention: Feature amount calculation method, feature amount calculation program, feature amount calculation device, screening method, screening program, screening device, compound creation method, compound creation program, and compound creation device
Technical field
[0001]
　The present invention relates to a method, a program, and a device for calculating a characteristic amount, screening a compound, and creating a three-dimensional structure of a compound, and particularly to a technique for searching a drug candidate compound.
Background technology
[0002]
　Conventionally, in the drug discovery research using a computer, a drug candidate compound (a drug candidate compound ( We have searched for "hits" below. For example, in Patent Document 1 below, the structural formula of a compound is given to predict the binding force. Further, Patent Document 2 also describes repeating generation of a structural formula and prediction of a binding force to search for a compound having a desired binding force little by little (trial and error).
[0003]
　Further, Patent Document 3 describes that a search is performed using a descriptor called a “compound fingerprint (fingerprint)”. The "descriptor" is information obtained from the structural formula of the compound, and the "compound fingerprint" indicates information such as the presence or absence of various functional groups. Such a descriptor is characterized in that "if the descriptors of a compound are similar, the skeletons of the compound are similar".
Prior art documents
Patent literature
[0004]
Patent Document 1: US Patent No. 9373059,
Patent Document 2: Patent No. 5946045,
Patent Document 3: Patent No. 4564097
Summary of the invention
Problems to be Solved by the Invention
[0005]
　Recently, target proteins that are in high demand have become complicated and difficult, and it is difficult to find hits in simple library screening. On the other hand, the theoretical number of compounds is 10 (60 to the power of 60) even if it is a small molecule with a molecular weight of 500 or less. Considering that the number of such compounds is (10 9), there is still a possibility of hit discovery. However, investigating the binding force for all such astronomical numbers of compounds is almost impossible not only by experiment but also by simulation. Even in the case of examining the binding force for some compounds, the efficiency is low when trial and error are repeated as described in Patent Documents 1 and 2. Further, in the case of a conventional descriptor (feature amount) such as Fingerprint described in Patent Document 3, even if the compounds exhibit the same drug effect, the feature amounts are not always similar, and the feature amount is the target structure. Since the chemical properties of the body were not shown accurately, the efficiency of the search using features was low.
[0006]
　As described above, in the conventional technique, the feature amount does not accurately indicate the chemical property of the target structure, and therefore, the efficiency of screening using the feature amount and creation of a three-dimensional structure was low.
[0007]
　The present invention has been made in view of such circumstances, and provides a feature amount calculation method, a feature amount calculation program, and a feature amount calculation device that can calculate a feature amount that accurately indicates the chemical properties of a target structure. With the goal. Another object of the present invention is to provide a screening method, a screening program, and a screening device that can efficiently screen a drug candidate compound by using the characteristic amount. Another object of the present invention is to provide a compound creation method, a compound creation program, and a compound creation device that can efficiently create a three-dimensional structure of a drug candidate compound by using a characteristic amount.
Means for solving the problem
[0008]
　In order to achieve the above-mentioned object, the feature amount calculation method according to the first aspect of the present invention includes a target structure designating step of designating a target structure composed of a plurality of unit structures having chemical properties, A three-dimensional structure generation process for generating a three-dimensional structure of a plurality of unit structures for a target structure, and a feature amount calculation for calculating a feature amount by quantifying the degree of accumulation of one or more types of probes around the three-dimensional structure in a three-dimensional space The probe is a structure in which a plurality of points having a real charge and generating a Van der Waals force are arranged at a distance.
[0009]
　Since the chemical properties of the target structure are expressed as a result of the interaction between the target structure and one or more kinds of probes in the periphery of the target structure, the fact that the degree of probe accumulation is similar between the target structures means that the target structures are similar to each other. It shows that the chemistry of the body is similar. That is, target structures having similar feature amounts calculated according to the first aspect exhibit similar chemical properties. Therefore, according to the first aspect, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure. In the first aspect and each of the following aspects, the probe may be "a structure in which a plurality of points having a real number charge and generating a Van der Waals force are arranged at a certain distance."
[0010]
　A feature amount calculating method according to a second aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designation step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generation step. In the amount calculating step, the first characteristic amount, which is a characteristic amount obtained by quantifying the accumulation degree of amino acids as probes in the three-dimensional space around the three-dimensional structure of the compound generated in the three-dimensional structure generating step, is calculated. In the second aspect, the “probe”, “target structure”, and “plurality of unit structures” in the first aspect are amino acids, compounds, and plural atoms, respectively. The amino acid for quantifying the degree of accumulation is not limited to one type, and may be a peptide in which two or more types of amino acids are bound.
[0011]
　Similar to the first aspect, the drug efficacy (binding power to the target protein) of the compound is locally expressed as the result of the interaction between the compound and each amino acid, and therefore the degree of accumulation of amino acids is similar between the compounds. If so, it indicates that the compounds have similar binding strength to the protein. That is, the compounds having similar feature amounts according to the second aspect show similar drug effects. Therefore, according to the second aspect, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the compound. In the second aspect, a compound such as a bioligand whose three-dimensional structure and binding to the target protein are known can be designated as the target structure.
[0012]
　The feature amount calculating method according to the third aspect further includes, in the second aspect, an invariant step of invariantizing the first feature amount with respect to rotation and translation of the compound to calculate the first invariant feature amount. In the third aspect, since the first feature amount is invariant with respect to the rotation and translation of the compound, it is easy to handle and the data capacity can be reduced. The invariantization of the first feature amount can be performed by Fourier transform, angle integration of a correlation function, or the like.
[0013]
　The feature amount calculating method according to the fourth aspect is the feature amount calculating method according to the third aspect, wherein in the feature amount calculating step, the first feature amount is calculated for two different types of amino acids, and in the invariant step, the first feature amount is calculated for two different types of amino acids. The first invariantized feature amount is calculated using the feature amount of. According to the fourth aspect, it is possible to perform the invariant while maintaining the information on the interaction between amino acids by using the first feature amount for two different types of amino acids in the calculation of the first invariant feature amount. Therefore, the compounds can be accurately compared (drug efficacy determination) based on the feature amount (first invariantized feature amount).
[0014]
　A feature amount calculating method according to a fifth aspect is the method according to the first aspect, wherein in the target structure designating step, a pocket structure that binds to a pocket that is an active site of a target protein is designated as a target structure, and a three-dimensional structure generating step is performed. Then, the three-dimensional structure of the pocket structure is generated by a plurality of virtual spheres, and in the feature amount calculation step, the accumulation degree of amino acids as probes around the three-dimensional structure of the pocket structure generated in the three-dimensional structure generation step is set to 3 A second feature amount that is a quantified feature amount in the dimensional space is calculated. In the fifth aspect, the “probe”, “target structure”, and “unit structure” in the first aspect are amino acids, pocket structures, and a plurality of virtual spheres, respectively. The "active site" of the target protein means a site where the activity of the target protein is promoted or suppressed by binding of the pocket structure, and the "virtual sphere" has chemical properties such as van der Waals radius and electric charge. Can be considered to have.
[0015]
　In the above-mentioned second aspect, the degree of amino acid accumulation for a given compound is calculated, whereas in the fifth aspect, the degree of amino acid accumulation for a pocket structure that binds to a given target protein pocket is calculated. .. Since the pocket structure having similar feature amounts according to the fourth aspect exhibits similar chemical properties as described above with respect to the second aspect, the chemical properties of the pocket structure can be accurately determined by the fifth aspect. It is possible to calculate the indicated feature amount. The pocket structure corresponds to a compound that binds to the pocket of the target protein. Further, in the fifth aspect, a simulation based on the measurement result of the three-dimensional structure of the target protein, the position information of the pocket, and the like can be used for calculating the second feature amount. The three-dimensional structure of the target protein is a measurement technique (X-ray crystal structure, NMR structure (NMR: Nuclear Magnetic Resonance), cryo-TEM structure (TEM: Transmission Electron Microscopy), etc.) is not limited.
[0016]
　A feature amount calculating method according to a sixth aspect is the method of the fifth aspect, further comprising an invariant step of invariantizing the second feature amount with respect to rotation and translation of the pocket structure to calculate a second invariant feature amount. Have. According to the sixth aspect, similar to the third aspect, it is possible to easily handle the feature amount and reduce the data capacity. The invariantization of the second feature amount can be performed by Fourier transform, angle integration of the correlation function, or the like, as in the third aspect.
[0017]
　A feature amount calculating method according to a seventh aspect is the feature amount calculating method according to the sixth aspect, wherein a second feature amount is calculated for two different types of amino acids in the feature amount calculating step, and a second feature amount is calculated for two different types of amino acids in the invariant step. The second invariantized feature amount is calculated using the feature amount of. According to the seventh aspect, by using the second feature quantity for two different types of amino acids in the calculation of the second invariant feature quantity, it is possible to perform the invariant while maintaining the information on the interaction between amino acids. Therefore, it is possible to accurately compare the compounds based on the feature amount (second invariantized feature amount) (medication determination).
[0018]
　A feature quantity calculating method according to an eighth aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designation step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generation step. In the amount calculation step, the accumulation degree of the probe around the three-dimensional structure of the compound generated in the three-dimensional structure generation step, which is one or more nucleobases, one or more lipid molecules, one or more monosaccharide molecules, A third feature amount, which is a feature amount obtained by quantifying the degree of accumulation of one or more of water and one or more types of ions as probes in a three-dimensional space, is calculated. An eighth aspect is that each of the “probe”, “target structure”, and “plurality of unit structures” in the first aspect is one or more types of nucleobases (which may be any type, number or combination), compounds , With multiple atoms.
[0019]
　In the present invention, DNA (Deoxyribonucleic Acid), RNA (Ribonucleic Acid), cell membranes, and polysaccharides, which are biopolymers (compounds) other than proteins, can be treated as targets of medicines. The eighth aspect defines a method for calculating the characteristic amount of these target compounds, and the probe is not an amino acid but another substance (building block of each target). Specifically, when the target is DNA, RNA, cell membrane, or polysaccharide, the probe is one or more nucleobases, one or more nucleobases, one or more lipid molecules, one or more monosaccharide molecules, respectively. .. Moreover, when quantifying the degree of accumulation using these as probes, water and one or more types of ions may be taken into consideration. Similar to the second and fifth aspects, the medicinal effect of a compound (binding force to a target such as DNA) is locally expressed as a result of the interaction between the compound and a nucleobase (probe). If the degree of accumulation of nucleic acid bases and the like between them is similar, it is indicated that those compounds have similar binding power to the target. That is, the compounds having similar feature amounts according to the eighth aspect show similar drug effects. Therefore, according to the eighth aspect, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the compound.
[0020]
　The feature amount calculating method according to the ninth aspect further includes, in the eighth aspect, an invariant step of making the third feature amount invariant with respect to the rotation and translation of the compound to calculate the third invariant feature amount. According to the ninth aspect, similar to the third and sixth aspects, it is possible to easily handle the feature amount and reduce the data capacity. The invariantization of the third feature amount can be performed by Fourier transform, angle integration of the correlation function, or the like, as in the third and sixth aspects.
[0021]
　A feature amount calculating method according to a tenth aspect is the ninth aspect, wherein in the feature amount calculating step, one or more nucleobases, one or more lipid molecules, one or more monosaccharide molecules, water, and one type. A first probe composed of one or more of the above ions, one or more nucleobases, one or more lipid molecules, one or more monosaccharide molecules, water, and one or more ions A third feature quantity is calculated for a second probe composed of one or more and different from the first probe, and in the invariant step, the third feature quantity for the first probe is calculated. A third invariantized feature amount is calculated using the feature amount and the third feature amount for the second probe. According to the tenth aspect, information on probe interaction can be obtained by using the third feature amount for two different types of probes (first and second probes) in the calculation of the third invariant feature amount. Since the invariantization can be performed while maintaining it, the comparison (drug efficacy determination) of the compounds based on the feature amount (third invariantized feature amount) can be accurately performed. In the tenth aspect, if at least one of the types, numbers, and combinations of constituent elements (one or more types of nucleobases) of the first and second probes is different, “the second probe is the first It is different from the probe”.
[0022]
　A feature amount calculating method according to an eleventh aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designating step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generating step. In the amount calculation step, the accumulation degree of the probe around the three-dimensional structure of the compound generated in the three-dimensional structure generation step, which is the first probe that is one or more kinds of amino acids, one or more kinds of nucleobases, and one kind A feature quantity quantifying the degree of accumulation in the three-dimensional space using the above lipid molecule, one or more types of monosaccharide molecules, water, and a second probe that is one or more of one or more types of ions as a probe. A certain fourth characteristic amount is calculated. In an eleventh aspect, each of the “probe”, “target structure”, and “plurality of unit structures” in the first aspect has one or more amino acids (first probe) and one or more nucleobases ( Second probe; may be of any type, number and combination)), compound, multiple atoms.
[0023]
　Similar to the second, fifth, and eighth aspects, the drug efficacy of a compound (binding force to a target such as protein or DNA) is locally expressed as a result of the interaction between the compound and an amino acid, a nucleobase, or the like. Therefore, if the degree of accumulation of amino acids, nucleic acid bases, etc. is similar among the compounds, it indicates that the compounds have similar binding power to the target. That is, the compounds having similar feature amounts according to the eleventh aspect show similar drug effects. Therefore, by using a building block such as protein and DNA as a probe, it is possible to calculate a characteristic amount that accurately indicates the chemical properties of the target structure even when the complex structure is targeted.
[0024]
　The feature quantity calculating method according to the twelfth aspect further includes, in the eleventh aspect, an invariant step of invariantizing the fourth feature quantity with respect to the rotation and translation of the compound to calculate the fourth invariant feature quantity. According to the twelfth aspect, similar to the third, sixth and ninth aspects, it is possible to easily handle the feature amount and reduce the data capacity. The invariantization of the fourth feature amount can be performed by the Fourier transform, the angle integration of the correlation function, or the like, as in the third, sixth, and ninth modes.
[0025]
　A feature quantity calculating method according to a thirteenth aspect is the twelfth aspect, wherein in the feature quantity calculating step, the fourth feature quantity is calculated for two types of probes in which at least one of the first probe and the second probe is different. Then, in the invariantization step, the fourth invariantized feature amount is calculated using the fourth feature amount for the two types of probes. According to the thirteenth aspect, by using the fourth feature amount for two types of probes in the calculation of the fourth invariant feature amount, it is possible to perform the invariant process while maintaining the information on the interaction of the probes. Therefore, it is possible to accurately compare the compounds (determination of drug efficacy) based on the feature amount (fourth invariant feature amount).
[0026]
　In the thirteenth aspect, “two types of probes in which at least one of the first probe and the second probe is different” is a probe composed of the first probe and the second probe, Of the two types of probes in which the combination of the first probe and the second probe is different, one probe is composed of lysine (one type of amino acid; an example of the first probe) and nucleobase (an example of the second probe). As in the case where the other probe is composed of lysine (an example of the first probe) and a lipid molecule (another example of the second probe), “the first probe is the same and the second probe is different”. In one case, one probe is composed of lysine (an example of the first probe) and a nucleobase (an example of the second probe), and the other probe is arginine (one of the amino acids; Example) and a nucleobase (an example of the second probe), “the first probe is different and the second probe is the same”, and one probe is lysine (the first probe). And an nucleobase (an example of a second probe), and the other probe is composed of arginine (another example of the first probe) and a lipid molecule (another example of the second probe). As in the case, the case where "both the first probe and the second probe are different" is included. Here, if at least one of the type, number, and combination of the constituent elements of the probe is different, it corresponds to "one probe is different from another probe".
[0027]
　A feature amount calculating method according to a fourteenth aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designation step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generation step. In the amount calculation step, a first point charge having an electric charge of +1 and a second point charge having an electric charge of −1, which is a degree of integration of the probes around the three-dimensional structure of the compound generated in the three-dimensional structure generating step, , A third point charge having a charge of +0.1, a fourth point charge having a charge of −0.1, a dipole in which the first point charge and the second point charge are spaced apart, and a charge A fifth feature amount, which is a feature amount obtained by quantifying the degree of accumulation in a three-dimensional space using one or more (any type, number, or combination) of the fifth point charges with zero as a probe, is calculated. To do. The fourteenth aspect defines a method of calculating a feature amount for a virtual probe, but similar to the second, fifth, eighth, and eleventh aspects, the feature according to the fourteenth aspect Compounds with similar amounts show similar efficacy. Therefore, even when a virtual probe is used, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure.
[0028]
　The feature amount calculating method according to the fifteenth aspect further comprises, in the fourteenth aspect, an invariant step of invariantizing the fifth feature amount with respect to rotation and translation of the compound to calculate a fifth invariant feature amount. According to the fifteenth aspect, similar to the third, sixth, ninth and twelfth aspects, the feature amount can be easily handled and the data capacity can be reduced. The invariantization of the fifth feature amount can be performed by the Fourier transform, the angle integration of the correlation function, or the like, as in the third, sixth, ninth, and twelfth aspects.
[0029]
　A feature quantity calculating method according to a sixteenth aspect is the method according to the fifteenth aspect, wherein in the feature quantity calculating step, the first point charge, the second point charge, the third point charge, the fourth point charge, the dipole, A first probe composed of one or more of 5 point charges (which may be any type, number, or combination), a first point charge, a second point charge, a third point charge, and a fourth point charge. A second probe different from the first probe, the second probe being composed of one or more of (5) point charge, dipole, and fifth point charge (may be any type, number, or combination). , And a fifth feature amount for the first probe and a fifth feature amount for the second probe are used in the invariantization step to obtain a fifth invariant feature amount. calculate. According to the sixteenth aspect, by using the fifth feature amount for two different types of probes (first and second probes) in the calculation of the fifth invariantized feature amount, information on probe interaction can be obtained. Since the invariantization can be performed while maintaining, the comparison (drug efficacy determination) of the compounds based on the feature amount (fifth invariantized feature amount) can be accurately performed. In the sixteenth aspect, if at least one of the types, numbers, and combinations of the constituent elements (first point charges, etc.) of the first and second probes is different, “the second probe is the first probe”. Is different from”.
[0030]
　A feature amount calculating method according to a seventeenth aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designating step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generating step. In the amount calculation step, the accumulation degree of the probe around the three-dimensional structure of the compound generated in the three-dimensional structure generation step, that is, the first probe that is one or more kinds of amino acids and the first point that the charge is +1 Electric charge, a second point charge having a charge of -1, a third point charge having a charge of +0.1, a fourth point charge having a charge of -0.1, a first point charge and a second point charge Integration using a dipole disposed apart from the point charge and a second probe that is one or more (any type, number, or combination) of the fifth point charges having zero charge as a probe A sixth feature amount, which is a feature amount obtained by quantifying the degree in a three-dimensional space, is calculated. The seventeenth aspect defines a method for calculating a feature amount when the first probe (one or more kinds of amino acids) and the second probe (point charge or the like and a combination thereof) are used as the probes. Similarly to the fifth, eighth, eleventh, and fourteenth aspects, the compounds having similar feature amounts according to the seventeenth aspect show similar pharmacological effects. Therefore, even when the first probe and the second probe are used, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure.
[0031]
　A feature quantity calculating method according to an eighteenth aspect further comprises, in the seventeenth aspect, an invariant step of making the sixth feature invariant with respect to rotation and translation of the compound to calculate a sixth invariant feature amount. According to the eighteenth aspect, similar to the third, sixth, ninth, twelfth, and fifteenth aspects, it is possible to easily handle the feature amount and reduce the data capacity. The invariantization of the sixth characteristic amount can be performed by Fourier transform, angle integration of the correlation function, or the like, as in the third, sixth, ninth, twelfth, and fifteenth modes.
[0032]
　第１９の態様に係る特徴量算出方法は第１８の態様において、特徴量算出工程では、第１のプローブと第２のプローブとのうち少なくとも一方が異なる２種類のプローブについて第６の特徴量を算出し、不変量化工程では２種類のプローブについての第６の特徴量を用いて第６の不変量化特徴量を算出する。第１９の態様によれば、第６の不変量化特徴量の算出において２種類の異なるプローブについての第６の特徴量を用いることによりプローブの相互作用の情報を維持しつつ不変量化を行うことができるので、特徴量（第６の不変量化特徴量）に基づく化合物の比較（薬効判定）を正確に行うことができる。なお第１９の態様において「第１のプローブと第２のプローブとのうち少なくとも一方が異なる２種類のプローブ」とは、第１のプローブと第２のプロ―ブからなるプローブであって、第１のプローブと第２のプローブの組み合わせが異なる２種類のプローブのうち、一方のプローブがリジン（アミノ酸の１種；第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブがリジン（第１のプローブの一例）及び第２の点電荷（第２のプローブの他の例）で構成される場合のように「第１のプローブが同一で第２のプローブが異なる」場合と、一方のプローブがリジン（第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブがアルギニン（アミノ酸の１種；第１のプローブの他の例）及び第１の点電荷（第２のプローブの一例）で構成される場合のように「第１のプローブが異なり第２のプローブが同一である」場合と、一方のプローブがリジン（第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブがアルギニン（アミノ酸の１種；第１のプローブの他の例）及び第２の点電荷（第２のプローブの他の例）で構成される場合のように「第１，第２のプローブの双方が異なる」場合とを含む。ここで、プローブの構成要素の種類、数、組合せのうち少なくとも１つが異なっていれば「一のプローブが他のプローブと異なる」に該当する。
[0033]
　A feature amount calculating method according to a twentieth aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designating step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generating step. In the amount calculation step, the degree of accumulation of probes around the three-dimensional structure of the compound generated in the three-dimensional structure generation step, which is one or more nucleobases, one or more lipid molecules, water, one or more monosaccharides A molecule, a first probe that is one or more (any type, number, or combination) of one or more types of ions, a first point charge having a charge of +1 and a second probe having a charge of -1 , The third point charge having a charge of +0.1, the fourth point charge having a charge of −0.1, the first point charge and the second point charge are spaced from each other. A characteristic of quantifying the degree of integration in a three-dimensional space using a dipole and a second probe that is one or more (any kind, number, or combination) of the fifth point charges with zero charge as a probe. A seventh characteristic amount, which is an amount, is calculated. The twentieth aspect defines a feature amount calculation method in which the first probe (one or more kinds of nucleobases and combinations thereof) and the second probe (point charges and combinations thereof) are used as probes. However, similar to the second, fifth, eighth, eleventh, fourteenth, and seventeenth aspects, the compounds having similar feature amounts according to the twentieth aspect show similar pharmacological effects. Therefore, even when the first probe and the second probe are used, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure.
[0034]
　A feature amount calculating method according to a twenty-first aspect further comprises, in the twentieth aspect, an invariant step of invariantizing the seventh feature amount with respect to rotation and translation of a compound to calculate a seventh invariant feature amount. According to the twenty-first aspect, similar to the third, sixth, ninth, twelfth, fifteenth, and eighteenth aspects, the feature amount can be easily handled and the data capacity can be reduced. The invariantization of the seventh characteristic amount can be performed by the Fourier transform, the angular integration of the correlation function, or the like, as in the third, sixth, ninth, twelfth, fifteenth, and eighteenth aspects.
[0035]
　第２２の態様に係る特徴量算出方法は第２１の態様において、特徴量算出工程では、第１のプローブと第２のプローブとのうち少なくとも一方が異なる２種類のプローブについて第７の特徴量を算出し、不変量化工程では２種類のプローブについての第７の特徴量を用いて第７の不変量化特徴量を算出する。第２２の態様によれば、第７の不変量化特徴量の算出において２種類の異なるプローブについての第７の特徴量を用いることによりプローブの相互作用の情報を維持しつつ不変量化を行うことができるので、特徴量（第７の不変量化特徴量）に基づく化合物の比較（薬効判定）を正確に行うことができる。なお第２２の態様において「第１のプローブと第２のプローブとのうち少なくとも一方が異なる２種類のプローブ」とは、第１のプローブと第２のプロ―ブからなるプローブであって、第１のプローブと第２のプローブの組み合わせが異なる２種類のプローブのうち、一方のプローブが核酸塩基（第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブが核酸塩基（第１のプローブの一例）及び第２の点電荷（第２のプローブの他の例）で構成される場合のように「第１のプローブが同一で第２のプローブが異なる」場合と、一方のプローブが核酸塩基（第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブが脂質分子（第１のプローブの他の例）及び第１の点電荷（第２のプローブの一例）で構成される場合のように「第１のプローブが異なり第２のプローブが同一である」場合と、一方のプローブが核酸塩基（第１のプローブの一例）及び第１の点電荷（第２のプローブの一例）で構成され、他方のプローブが脂質分子（第１のプローブの他の例）及び第２の点電荷（第２のプローブの他の例）で構成される場合のように「第１，第２のプローブの双方が異なる」場合とを含む。ここで、プローブの構成要素の種類、数、組合せのうち少なくとも１つが異なっていれば「一のプローブが他のプローブと異なる」に該当する。
[0036]
　A feature amount calculating method according to a twenty-third aspect is the method according to the first aspect, wherein a compound is designated as a target structure in the target structure designating step, and a three-dimensional structure of the compound is generated by a plurality of atoms in the three-dimensional structure generating step. In the amount calculation step, the accumulation degree of the probe around the three-dimensional structure of the compound generated in the three-dimensional structure generation step, which is the first probe that is one or more kinds of amino acids, one or more kinds of nucleobases, and one kind A second probe that is one or more (any type, number, or combination) of the above lipid molecules, water, one or more types of monosaccharide molecules, and one or more types of ions; A point charge of 1, a second point charge of −1, a third point charge of +0.1, a fourth point charge of −0.1, and a first point charge. A dipole disposed apart from the second point charge, and a third probe which is one or more (any kind, number, or combination) of the fifth point charges having zero charge. An eighth feature amount, which is a feature amount obtained by quantifying the accumulation degree in the three-dimensional space, is calculated. A twenty-third aspect is a first probe (one or more kinds of amino acids), a second probe (one or more kinds of nucleobases and the combination thereof), a third probe (point charge and the like and combination thereof), The feature amount calculating method is defined when the third probe is used as a probe. However, similar to the second, fifth, eighth, eleventh, fourteenth, seventeenth, and twentieth aspects, The compounds having similar feature amounts according to the twenty-third aspect show similar drug effects. Therefore, even when the first probe to the third probe are used, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure.
[0037]
　The feature quantity calculating method according to the twenty-fourth aspect further comprises, in the twenty-third aspect, an invariant step of invariantizing the eighth feature quantity with respect to the rotation and translation of the compound to calculate the eighth invariant feature quantity. According to the twenty-fourth aspect, like the third, sixth, ninth, twelfth, fifteenth, eighteenth and twenty-first aspects, the feature amount can be easily handled and the data capacity can be reduced. The invariantization of the eighth feature amount can be performed by the Fourier transform, the angular integration of the correlation function, or the like as in the third, sixth, ninth, twelfth, fifteenth, eighteenth, and twenty-first modes.
[0038]
　A feature amount calculating method according to a twenty-fifth aspect is the feature amount calculating method according to the twenty-fourth aspect, wherein in the feature amount calculating step, at least one of the first probe, the second probe, and the third probe is different from each other. And the invariantization step calculates the eighth invariantized feature amount using the eighth feature amounts for the two types of probes. According to the twenty-fifth aspect, by using the eighth feature amount for two different types of probes in the calculation of the eighth invariant feature amount, it is possible to perform the invariant process while maintaining the information on the interaction of the probes. Therefore, the compound comparison (drug efficacy determination) based on the feature amount (eighth invariantized feature amount) can be accurately performed. In addition, in the twenty-fifth aspect, “two kinds of probes in which at least one of the first probe, the second probe, and the third probe is different” is a probe including the first, second, and third probes. And between one probe and the other of the two types of probes in which the combination of the first, second, and third probes is different, "one of the first, second, and third probes is It includes a case of “different”, a case of “two of the first, second, and third probes are different” and a case of “all of the first, second, and third probes are different”. Here, if at least one of the type, number, and combination of the constituent elements of the probe is different, it corresponds to "one probe is different from another probe".
[0039]
　In order to achieve the above-mentioned object, a feature amount calculation program according to a twenty-sixth aspect of the present invention causes a computer to execute the feature amount calculation method according to any one of the first to twenty-fifth aspects. The “computer” in the twenty-sixth aspect can be realized by using one or more various processors such as a CPU (Central Processing Unit). A non-transitory recording medium in which a computer-readable code of the feature amount calculation program according to the twenty-sixth aspect is recorded can also be cited as an aspect of the present invention.
[0040]
　In order to achieve the above-mentioned object, a feature amount calculation apparatus according to a twenty-seventh aspect of the present invention includes a target structure designating unit that designates a target structure composed of a plurality of unit structures having chemical properties, and a target structure designating unit. A three-dimensional structure generation unit that generates a three-dimensional structure of a plurality of unit structures for a structure, and a feature amount calculation unit that calculates a feature amount that quantifies the accumulation degree of one or more types of probes around the three-dimensional structure in a three-dimensional space. And a probe is a structure in which a plurality of points having a real number charge and generating a Van der Waals force are arranged at a distance. According to the twenty-seventh aspect, it is possible to calculate the characteristic amount that accurately indicates the chemical property of the target structure, as in the first aspect. In the twenty-seventh aspect, as described above regarding the first to twenty-sixth aspects, the target compound is a protein, DNA or the like, the probe is an amino acid, a nucleic acid base or the like, a virtual charge or the like, and the first to fifth characteristic amounts are set. It can be calculated. In the twenty-seventh aspect, the probe may be “a structure in which a plurality of points having a real number charge and generating a Van der Waals force are arranged at a certain distance”.
[0041]
　In order to achieve the above-mentioned object, a screening method according to a twenty-eighth aspect of the present invention is a screening method for extracting a target compound that binds to a target protein from a plurality of compounds, wherein a plurality of atoms are included in each of the plurality of compounds. The storage step of associating and storing the three-dimensional structure of the compound by the method and the first characteristic amount calculated using the characteristic amount calculation method according to the second aspect for the three-dimensional structure of the compound, and the binding to the target protein is confirmed. Amount calculating step of calculating a first characteristic amount of a ligand which is a compound being analyzed, and a similarity degree of calculating a similarity between the first characteristic amount of a plurality of compounds and the first characteristic amount of a ligand It has a calculation process and a compound extraction process which extracts a target compound from a plurality of compounds based on a degree of similarity. As described above regarding the second aspect, if the first feature amount (three-dimensionally quantified feature amount) of the ligand and the target compound are similar, the drug effects of both are similar. Therefore, according to the twenty-eighth aspect, it is possible to efficiently screen the drug candidate compound by extracting the target compound having similar drug efficacy to the ligand based on the first feature amount.
[0042]
　In order to achieve the above-mentioned object, a screening method according to a twenty-ninth aspect of the present invention is a screening method for extracting a target compound that binds to a target protein from a plurality of compounds, wherein a plurality of atoms are included in each of the plurality of compounds. A third aspect of the storage step of storing the three-dimensional structure of the compound according to 1 and the first invariant characteristic amount of the three-dimensional structure of the compound in association with each other, and the ligand that is a compound whose binding to the target protein is confirmed. A feature quantity calculating step of calculating a first invariant feature quantity using the feature quantity calculating method according to the first aspect, and a first invariant feature quantity for a plurality of compounds and a first invariant feature quantity for a ligand. The method includes a similarity calculation step of calculating a similarity and a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. The twenty-ninth aspect is common to the twenty-eighth aspect in that the feature amount of the ligand is calculated, but in the twenty-ninth aspect, the drug efficacy is similar to that of the ligand based on the similarity of the first invariantized feature amount. The target compound can be extracted and the drug candidate compound can be efficiently screened.
[0043]
　The screening method according to the thirtieth aspect is the method according to the twenty-eighth aspect or the twenty-ninth aspect, wherein in the compound extraction step, compounds having a similarity of not less than a threshold value are extracted. The thirtieth aspect defines a specific standard for extracting a target compound based on the degree of similarity. By extracting a compound having a degree of similarity equal to or higher than a threshold value, it is possible to efficiently screen a drug candidate compound. The threshold value can be set based on conditions such as the purpose of screening and accuracy, and may be set based on a value specified by the user.
[0044]
　The screening method according to the thirty-first aspect is the method according to any one of the twenty-eighth to thirtieth aspects, in which the compounds are extracted in descending order of similarity in the compound extraction step. The thirty-first aspect defines specific criteria of target compound extraction based on similarity, and by screening compounds in descending order of similarity, drug candidate compounds can be efficiently screened.
[0045]
　In order to achieve the above-mentioned object, a screening method according to a thirty-second aspect of the present invention is a screening method for extracting a target compound that binds to a target protein from a plurality of compounds, and a plurality of atoms are included in each of the plurality of compounds. According to a fifth aspect of the present invention, a storage step of associating and storing the three-dimensional structure of the compound according to 1) and the first feature amount calculated using the feature amount calculating method according to the second aspect, and the pocket structure of the target protein. And a similarity between a first feature amount for a plurality of compounds and a second feature amount for a pocket structure. And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above regarding the fifth aspect, if the second feature amount of the pocket structure and the target compound are similar, the chemical properties of the two are similar. Therefore, according to the thirty-second aspect, it is possible to efficiently screen the drug candidate compound by extracting the target compound having similar chemical properties to the pocket structure. Since the pocket structure corresponds to the compound that binds to the target protein, it is possible to compare and contrast the feature amount of the pocket structure (second feature amount) and the feature amount of the compound (first feature amount). Yes Similarity can be calculated.
[0046]
　In order to achieve the above-mentioned object, the screening method according to the thirty-third aspect of the present invention is a screening method for extracting a target compound that binds to a target protein from a plurality of compounds, wherein each of the plurality of compounds has a plurality of atoms. The storage step of storing the three-dimensional structure of the compound according to 3 and the first invariantized feature amount calculated by using the feature amount calculation method according to the third aspect in association with each other, and the pocket structure of the target protein, A feature amount calculating step of calculating a second invariant feature amount using the feature amount calculating method according to the first aspect, a first invariant feature amount for a plurality of compounds, and a second invariant amount for a pocket structure. A similarity calculation step of calculating a similarity with the feature amount, and a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity are provided. In the thirty-third aspect, it is possible to efficiently screen a drug candidate compound by extracting a target compound having a chemical property similar to that of the pocket structure using the first and second invariantized feature quantities. Similar to the 32nd aspect described above, the feature amount (the second invariant feature amount) for the pocket structure and the feature amount (the first invariant feature amount) for the compound can be compared and compared. And the degree of similarity can be calculated.
[0047]
　The screening method according to the thirty-fourth aspect is the thirty-second or thirty-third aspect, wherein in the compound extraction step, compounds having a similarity of not less than a threshold value are extracted. The thirty-fourth aspect defines specific criteria for target compound extraction based on the degree of similarity. By extracting compounds having a degree of similarity equal to or higher than a threshold value, screening of drug candidate compounds can be efficiently performed. The threshold value can be set based on conditions such as the purpose of screening and accuracy, and may be set based on a value specified by the user.
[0048]
　The screening method according to the thirty-fifth aspect is the method according to any one of the thirty-second to thirty-fourth aspects, wherein in the compound extraction step, compounds are extracted in descending order of similarity. The thirty-fifth aspect defines specific criteria for target compound extraction based on similarity, and by screening compounds in descending order of similarity, drug candidate compounds can be efficiently screened.
[0049]
　In order to achieve the above-mentioned object, the screening method according to the 36th aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer other than a protein from a plurality of compounds, each of the plurality of compounds Regarding a three-dimensional structure of a compound having a plurality of atoms and a third characteristic amount calculated using the characteristic amount calculation method according to the eighth aspect for the three-dimensional structure of a compound, a storage step of storing them in association with each other A feature amount calculating step of calculating a third feature amount of a binding compound that is a compound that is confirmed to bind to the target biopolymer, and a third feature amount of a plurality of compounds and a third feature amount of the binding compound. 3 includes a similarity calculation step of calculating the similarity with the feature amount of 3, and a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the eighth aspect, the present invention can handle DNA, which is a target biopolymer other than protein, and the third characteristic amount of the binding compound that binds to the target biopolymer and the target compound is If similar, the medicinal effects of both are similar. Therefore, according to the thirty-sixth aspect, it is possible to efficiently screen the drug candidate compound by extracting the target compound having similar drug efficacy to the binding compound based on the third feature amount.
[0050]
　In order to achieve the above-mentioned object, the screening method according to the 37th aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, A storage step of associating and storing the three-dimensional structure of the compound with the atom of the fourth feature amount and the fourth feature amount of the three-dimensional structure of the compound calculated by using the feature amount calculating method according to the eleventh aspect; Of a fourth feature amount of a plurality of compounds and a fourth feature amount of a plurality of compounds, and a similarity between the fourth feature amount of a plurality of compounds And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the eleventh aspect, compounds having similar fourth characteristic amounts show similar drug effects. Therefore, according to the thirty-seventh aspect, even when targeting a complex structure, it is possible to efficiently screen a drug candidate compound by extracting a target compound having similar drug efficacy to the binding compound based on the fourth characteristic amount. You can
[0051]
　In order to achieve the above-mentioned object, a screening method according to the 38th aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, A storage step of associating and storing the three-dimensional structure of the compound by the atom and the fifth characteristic amount calculated for the three-dimensional structure of the compound using the characteristic amount calculation method according to the fourteenth aspect; Of a fifth feature amount of a binding compound, which is a compound whose binding is confirmed, and a similarity between the fifth feature amount of a plurality of compounds and the fifth feature amount of a binding compound And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the fourteenth aspect, compounds having similar fifth characteristic amounts show similar drug effects. Therefore, according to the thirty-eighth aspect, even when a virtual probe is used, it is possible to efficiently screen a drug candidate compound by extracting a target compound having a drug efficacy similar to that of the binding compound based on the fifth feature amount. ..
[0052]
　In order to achieve the above-mentioned object, the screening method according to the thirty-ninth aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, A storage step of associating and storing the three-dimensional structure of the compound by the atoms of the compound and the sixth characteristic amount calculated by using the characteristic amount calculation method according to the seventeenth aspect with respect to the three-dimensional structure of the compound; Of a sixth feature amount of a binding compound that is a compound whose binding is confirmed, and similarity between the sixth feature amount of a plurality of compounds and the sixth feature amount of a binding compound And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the seventeenth aspect, compounds having similar sixth feature amounts show similar drug effects. Therefore, according to the thirty-ninth aspect, even when one or more kinds of amino acids (first probe) and virtual charges and the like (second probe) are used, the binding compound and the medicinal effect are based on the sixth characteristic amount. It is possible to efficiently screen a drug candidate compound by extracting a similar target compound.
[0053]
　In order to achieve the above-mentioned object, the screening method according to the 40th aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, A storage step of associating and storing the three-dimensional structure of the compound by the atoms of the compound and the seventh characteristic amount calculated for the three-dimensional structure of the compound using the characteristic amount calculation method according to the twentieth aspect; Of a seventh feature amount of a plurality of compounds and a seventh feature amount of a plurality of compounds and a feature amount calculating step of calculating a seventh feature amount of a binding compound that is a compound whose binding is confirmed. And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the twentieth aspect, the compounds having similar seventh characteristic amounts show similar drug effects. Therefore, according to the 40th aspect, even when one or more kinds of nucleic acid bases and the like (first probe) and virtual charges and the like (second probe) are used, a binding compound is determined based on the seventh characteristic amount. It is possible to efficiently screen a drug candidate compound by extracting a target compound having similar drug efficacy.
[0054]
　In order to achieve the above-mentioned object, the screening method according to the 41st aspect of the present invention is a screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, A storage step of associating and storing the three-dimensional structure of the compound by the atom of 8 and the eighth characteristic amount calculated by using the characteristic amount calculation method according to the twenty-third aspect with respect to the three-dimensional structure of the compound; and a target biopolymer. Of the eighth feature amount of the plurality of compounds and the eighth feature amount of the plurality of compounds, and the similarity between the eighth feature amount of the plurality of compounds and the eighth feature amount of the plurality of compounds. And a compound extraction step of extracting a target compound from a plurality of compounds based on the similarity. As described above with respect to the twenty-third aspect, the compounds having similar eighth feature amounts show similar drug effects. Therefore, according to the forty-first aspect, one or more kinds of amino acids (first probe), one or more kinds of nucleobase etc. (second probe), and virtual charges etc. (third probe) are used. Even in such a case, a target compound having similar drug efficacy to the binding compound can be extracted based on the eighth characteristic amount, and the drug candidate compound can be efficiently screened.
[0055]
　In order to achieve the above-mentioned object, the screening program according to the 42nd aspect of the present invention causes a computer to execute the screening method according to any one of the 28th to 41st aspects. The “computer” in the forty-second aspect can be realized by using one or more various processors such as a CPU (Central Processing Unit). A non-transitory recording medium in which the computer-readable code of the screening program according to the forty-second aspect is recorded can also be cited as an aspect of the present invention.
[0056]
　In order to achieve the above-mentioned object, a screening device according to a forty-third aspect of the present invention is a screening device for extracting a target compound that binds to a target protein from a plurality of compounds, wherein a plurality of atoms are included in each of the plurality of compounds. The binding between the target protein and the storage unit that stores the three-dimensional structure of the compound according to 1 and the first feature amount calculated by using the feature amount calculation method according to the second aspect for the three-dimensional structure of the compound and the binding to the target protein are confirmed. Amount calculating unit that calculates a first characteristic amount for a ligand that is a compound being analyzed, and a similarity degree that calculates a degree of similarity between the first characteristic amount for a plurality of compounds and the first characteristic amount for a ligand A calculation unit and a compound extraction unit that extracts a target compound from a plurality of compounds based on the degree of similarity are provided.
[0057]
　As described above regarding the second aspect, if the first feature amount (three-dimensionally quantified feature amount) of the ligand and the target compound are similar, the drug effects of both are similar. Therefore, according to the forty-third aspect, it is possible to efficiently screen the drug candidate compound by extracting the target compound having similar drug efficacy to the ligand based on the first feature amount.
[0058]
　In order to achieve the above-mentioned object, a screening apparatus according to a forty-fourth aspect of the present invention is a screening apparatus for extracting a target compound that binds to a target protein from a plurality of compounds, and a plurality of atoms are included in each of the plurality of compounds. A third aspect of a ligand that is a compound whose binding to a target protein has been confirmed, and a storage unit that stores the three-dimensional structure of the compound according to 1 and the first invariant feature amount for the three-dimensional structure of the compound in association with each other. Of the first invariantized feature amount using the feature amount calculation method according to the first aspect, and a first invariantized feature amount for a plurality of compounds and a first invariantized feature amount for a ligand. A similarity calculation unit that calculates a similarity and a compound extraction unit that extracts a target compound from a plurality of compounds based on the similarity are provided.
[0059]
　The forty-fourth aspect is similar to the forty-third aspect in that the feature amount for the ligand is calculated, but in the forty-fourth aspect, the drug efficacy is similar to that of the ligand based on the similarity of the first invariantized feature amount. The target compound can be extracted and the drug candidate compound can be efficiently screened.
[0060]
　The screening apparatus according to the 45th aspect is the 43rd or 44th aspect, wherein the compound extraction unit extracts a compound whose similarity is equal to or greater than a threshold value. The forty-fifth aspect defines a specific standard for extracting a target compound based on the degree of similarity. By extracting a compound having a degree of similarity equal to or higher than a threshold value, it is possible to efficiently screen a drug candidate compound. The threshold value can be set based on conditions such as the purpose of screening and accuracy, and may be set based on a value specified by the user.

claims
[Claim 1]
　A target structure designating step of designating a target structure composed of a plurality of unit structures having chemical properties, and a
　three-dimensional structure generating step of generating a three-dimensional structure by the plurality of unit structures for the target structure,
　A feature amount calculation step of calculating a feature amount by quantifying the accumulation degree of one or more kinds of probes in the periphery of the three-dimensional structure in a three-dimensional space
　,
　wherein the probe has a real charge and a van der Waals force. A method of calculating a feature amount, which is a structure in which a plurality of points to be generated are arranged apart from each other.
[Claim 2]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The feature amount calculation according to claim 1, wherein a first feature amount that is a feature amount obtained by quantifying the accumulation degree of amino acids as the probe in the three-dimensional space around the three-dimensional structure of the compound is calculated. Method.
[Claim 3]
　The feature quantity calculating method according to claim 2, further comprising an invariant step of making the first feature quantity invariant with respect to rotation and translation of the compound to calculate the first invariant feature quantity.
[Claim 4]
　In the characteristic amount calculation step, the first characteristic amount is calculated for
　two different kinds of amino acids, and in the invariantization step, the first invariantization is performed by using the first characteristic quantities of the two different kinds of amino acids. The feature amount calculating method according to claim 3, wherein the feature amount is calculated.
[Claim 5]
　In the target structure specifying step, a pocket structure that binds to a pocket that is an active site of a target protein is specified as the target structure, and in the
　three-dimensional structure generating step, the three-dimensional structure of the pocket structure by a plurality of virtual spheres. In the
　feature amount calculation step, the accumulation degree of amino acids as the probe around the three-dimensional structure of the pocket structure generated in the three-dimensional structure generation step is quantified in the three-dimensional space. The feature amount calculation method according to claim 1, wherein the second feature amount, which is an amount, is calculated.
[Claim 6]
　The feature quantity calculation method according to claim 5, further comprising an invariant step of making the second feature quantity invariant with respect to rotation and translation of the pocket structure to calculate a second invariant feature quantity.
[Claim 7]
　In the characteristic amount calculation step, the second characteristic amount is calculated for two different kinds of amino acids, and in the
　invariantization step, the second invariantization is performed by using the second characteristic quantities of the two different kinds of amino acids. The feature amount calculating method according to claim 6, wherein the feature amount is calculated.
[Claim 8]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of accumulation of the probe around the three-dimensional structure of the compound, including one or more nucleobases, one or more lipid molecules, one or more monosaccharide molecules, water, and one or more ions. The feature amount calculating method according to claim 1, wherein a third feature amount, which is a feature amount obtained by quantifying the degree of integration in which one or more of the probes is used as the probe in the three-dimensional space, is calculated.
[Claim 9]
　The feature amount calculation method according to claim 8, further comprising an invariant step of invariantizing the third feature amount with respect to rotation and translation of the compound to calculate a third invariant feature amount.
[Claim 10]
　In the characteristic amount calculation step, one or more of the one or more kinds of nucleobases, the one or more kinds of lipid molecules, the one or more kinds of monosaccharide molecules, the water, and the one or more kinds of ions are formed. A first probe, and one or more of the one or more nucleobases, the one or more lipid molecules, the one or more monosaccharide molecules, the water, and the one or more ions. And a second probe different from the first probe, the third characteristic amount is calculated, and in the
　invariant step, the third characteristic of the first probe is calculated. The feature amount calculation method according to claim 9, wherein the third invariantized feature amount is calculated using an amount and the third feature amount for the second probe.
[Claim 11]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of accumulation of the probe around the three-dimensional structure of the compound, the first probe being one or more amino acids, one or more nucleobases, one or more lipid molecules, water, 1 A fourth feature amount, which is a feature amount obtained by quantifying the degree of accumulation in the three-dimensional space with the second probe, which is one or more kinds of monosaccharide molecules and one or more types of ions, being quantified in the three-dimensional space. The feature amount calculation method according to claim 1, wherein the feature amount calculation is performed.
[Claim 12]
　The feature amount calculation method according to claim 11, further comprising an invariant step of invariantizing the fourth feature amount with respect to rotation and translation of the compound to calculate a fourth invariant feature amount.
[Claim 13]
　In the characteristic amount calculation step, the fourth characteristic amount is calculated for two types of probes in which at least one of the first probe and the second probe is different, and in the
　invariantization step, the two types of 13. The feature amount calculation method according to claim 12, wherein the fourth invariantized feature amount is calculated using the fourth feature amount of the probe.
[Claim 14]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of integration of the probe around the three-dimensional structure of the compound, the first point charge having a charge of +1, the second point charge having a charge of −1, and the charge of +0.1. A third point charge, a fourth point charge having a charge of −0.1, a dipole in which the first point charge and the second point charge are spaced apart from each other, and a fifth point having a zero charge. The feature amount calculation method according to claim 1, wherein a fifth feature amount, which is a feature amount obtained by quantifying the degree of accumulation in which one or more of the point charges of (1) is used as the probe in the three-dimensional space, is calculated.
[Claim 15]
　15. The feature quantity calculation method according to claim 14, further comprising an invariant step of making the fifth feature quantity invariant with respect to rotation and translation of the compound to calculate a fifth invariant feature quantity.
[Claim 16]
　In the feature amount calculation step, at least one of the first point charge, the second point charge, the third point charge, the fourth point charge, the dipole, and the fifth point charge is used. One of a first probe configured and one of the first point charge, the second point charge, the third point charge, the fourth point charge, the dipole, and the fifth point charge The fifth feature amount is calculated for the second probe configured as above and different from the first probe, and in the
　invariant step, the fifth feature amount is calculated for the first probe. 16. The feature quantity calculating method according to claim 15, wherein the fifth invariant feature quantity is calculated using a fifth feature quantity and the fifth feature quantity for the second probe.
[Claim 17]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of accumulation of the probe around the three-dimensional structure of the compound, the first probe being one or more kinds of amino acids, the first point charge having a charge of +1 and the charge of -1 A second point charge, a third point charge having a charge of +0.1, a fourth point charge having a charge of −0.1, the first point charge and the second point charge are separated from each other. A sixth feature amount which is a feature amount obtained by quantifying the degree of integration in which the dipoles are arranged as a unit and a second probe that is one or more of a fifth point charge with zero charge is the three-dimensional space. The feature amount calculation method according to claim 1, wherein the feature amount is calculated.
[Claim 18]
　18. The feature amount calculation method according to claim 17, further comprising an invariant step of invariantizing the sixth feature amount with respect to rotation and translation of the compound to calculate a sixth invariant feature amount.
[Claim 19]
　In the feature amount calculation step, the sixth feature amount is calculated for two types of probes in which at least one of the first probe and the second probe is different, and in the
　invariantization step, the two types of The feature amount calculation method according to claim 18, wherein the sixth invariantized feature amount is calculated using the sixth feature amount of the probe.
[Claim 20]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of accumulation of the probe around the three-dimensional structure of the compound, including one or more nucleobases, one or more lipid molecules, water, one or more monosaccharide molecules, and one or more ions. One or more of the first probe, a first point charge having a charge of +1, a second point charge having a charge of −1, a third point charge having a charge of +0.1, and a charge One or more of a fourth point charge of −0.1, a dipole in which the first point charge and the second point charge are spaced apart, and a fifth point charge of zero charge 2. The feature quantity calculation method according to claim 1, wherein a seventh feature quantity, which is a feature quantity obtained by quantifying the degree of integration of the second probe as the probe in the three-dimensional space, is calculated.
[Claim 21]
　21. The feature amount calculation method according to claim 20, further comprising an invariant step of invariantizing the seventh feature amount with respect to rotation and translation of the compound to calculate a seventh invariant feature amount.
[Claim 22]
　In the feature amount calculation step, the seventh feature amount is calculated for two types of probes in which at least one of the first probe and the second probe is different, and in the
　invariantization step, the two types of 22. The feature quantity calculation method according to claim 21, wherein the seventh invariantized feature quantity is calculated using the seventh feature quantity of the probe.
[Claim 23]
　In the target structure designating step, a compound is designated as the target structure, in the
　three-dimensional structure generating step, a three-dimensional structure of the compound is generated by a plurality of atoms, and in the
　feature amount calculating step, the three-dimensional structure generating step occurs. The degree of accumulation of the probe around the three-dimensional structure of the compound, the first probe being one or more kinds of amino acids, one or more kinds of nucleobases, one or more kinds of lipid molecules, water, 1 More than one type of monosaccharide molecule, a second probe that is at least one of one or more types of ions, a first point charge with an electric charge of +1 and a second point charge with an electric charge of -1, A third point charge of +0.1, a fourth point charge of −0.1, a dipole in which the first point charge and the second point charge are spaced apart,
8. An eighth feature amount, which is a feature amount obtained by quantifying the degree of integration of the third probe, which is one or more of the zero fifth point charges, as the probe in the three-dimensional space. 1. The feature amount calculation method described in 1.
[Claim 24]
　24. The feature quantity calculation method according to claim 23, further comprising an invariant step of making the eighth feature quantity invariant with respect to rotation and translation of the compound to calculate an eighth invariant feature quantity.
[Claim 25]
　In the feature amount calculation step calculates said first probe, said second probe, the feature value of the eighth for at least one of two different types of the probe of said third probe,
　the 25. The feature quantity calculation method according to claim 24, wherein in the invariantization step, the eighth invariant feature quantity is calculated using the eighth feature quantity for the two types of probes.
[Claim 26]
　A characteristic amount calculation program that causes a computer to execute the characteristic amount calculation method according to any one of claims 1 to 25.
[Claim 27]
　A target structure designating section for designating a target structure composed of a plurality of unit structures having chemical properties, and a
　three-dimensional structure generating section for generating a three-dimensional structure by the plurality of unit structures for the target structure,
　a feature quantity calculation unit that calculates a quantified characteristic amounts in a three-dimensional space integration degree of 1 or more probes in the vicinity of the three-dimensional structure
　provided with,
　the probe generates a van der Waals forces have a real charge A feature amount calculation device that is a structure in which a plurality of points to be separated are arranged.
[Claim 28]
　A screening method for extracting a target compound that binds to a target protein from a plurality of compounds,
　wherein, for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms and the three-dimensional structure of the compound are described in claim 2. The first feature amount calculated by using the feature amount calculation method described above, and a storage step of storing the first feature amount in association with each other, and
　the first feature amount for the ligand that is a compound whose binding to the target protein is confirmed. a feature amount calculation step of calculating,
　a similarity calculation step of calculating a similarity between the first feature amount for said ligand and said first feature value for said plurality of compounds,
　wherein from the plurality of compound And a compound extraction step of extracting the target compound based on the degree of similarity
　.
[Claim 29]
　A screening method for extracting a target compound that binds to a target protein from a plurality of compounds,
　comprising: for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms, and the first invariant for the three-dimensional structure of the compound. A storage step of storing a quantified feature quantity in association with each
　other, and a ligand that is a compound whose binding to the target protein has been confirmed, using the feature quantity calculation method according to claim 3 for the first invariant A feature amount calculating step of calculating a feature amount, and
　a similarity degree calculating step of calculating a degree of similarity between the first invariantized feature amount for the plurality of compounds and the first invariantized feature amount for the ligand And
　a compound extraction step of extracting the target compound from the plurality of compounds based on the similarity
　.
[Claim 30]
　30. The screening method according to claim 28 or 29, wherein in the compound extracting step, compounds having the similarity not less than a threshold value are extracted.
[Claim 31]
　31. The screening method according to claim 28, wherein in the compound extraction step, compounds are extracted in descending order of similarity.
[Claim 32]
　A screening method for extracting a target compound that binds to a target protein from a plurality of compounds,
　comprising: using a three-dimensional structure of a compound having a plurality of atoms for each of the plurality of compounds; and the feature amount calculating method according to claim 2. The storage step of storing the first feature amount calculated in association with the storage step and the
　pocket structure of the target protein using the feature amount calculating method according to claim 5, A characteristic amount calculating step
　of calculating, a similarity degree calculating step of calculating a degree of similarity between the first characteristic amount of the
　plurality of compounds and the second characteristic amount of the pocket structure, and the plurality of compounds And a compound extraction step of extracting the target compound based on the similarity
　.
[Claim 33]
　A screening method for extracting a target compound that binds to a target protein from a plurality of compounds,
　wherein, for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms and the feature amount calculating method according to claim 3 are used. The storage step of storing the first invariantized feature quantity calculated in association with the storage step and
　the pocket structure of the target protein using the feature amount calculation method according to claim 6, A characteristic amount calculation step of calculating a characteristic
　amount, and a similarity degree calculation of calculating a degree of similarity between the first invariantized characteristic amount for the plurality of compounds and the second invariantized characteristic amount for the pocket structure. a step,
　a compound extraction step of extracting the target compound based on the similarity of the plurality of compound
　screening method having.
[Claim 34]
　34. The screening method according to claim 32 or 33, wherein in the compound extracting step, compounds having the similarity of not less than a threshold value are extracted.
[Claim 35]
　The screening method according to any one of claims 32 to 34, wherein in the compound extraction step, compounds are extracted in descending order of similarity.
[Claim 36]
　A screening method for extracting a target compound that binds to a target biopolymer other than a protein from a
　plurality of compounds, for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms, and the three-dimensional structure of the compound
　It has been confirmed that the storage step of associating and storing the third feature amount calculated using the feature amount calculating method according to claim 8 and binding to the target biopolymer other than the protein. A feature amount calculating step of calculating
　the third feature amount of the binding compound which is a compound, and a similarity between the third feature amount of the plurality of compounds and the third feature amount of the binding compound. 　A screening method
　comprising: a similarity calculation step of calculating; and a compound extraction step of extracting the target compound from the plurality of compounds based on the similarity
.
[Claim 37]
　A screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　comprising: for each of the plurality of compounds, a three-dimensional structure of the compound with a plurality of atoms and the three-dimensional structure of the compound. The storage step of associating and storing the fourth feature amount calculated by using the feature amount calculating method according to claim 4,
　and the binding compound that is a compound whose binding with the target biopolymer has been confirmed. A feature amount calculating step of calculating a feature amount of
　4, and a similarity degree calculating step of calculating a similarity degree between the fourth feature amount of the plurality of compounds and the fourth feature amount of the binding compound,
　A compound extraction step of extracting the target compound from the plurality of compounds based on the degree of similarity
　.
[Claim 38]
　A screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The storage step of storing the fifth feature quantity calculated using the feature quantity calculation method according to claim 1 in association with each other,
　and the binding compound which is a compound whose binding to the target biopolymer has been confirmed. A feature amount calculating step of calculating a feature amount of
　5, and a similarity degree calculating step of calculating a degree of similarity between the fifth feature amount of the plurality of compounds and the fifth feature amount of the binding compound,
　A compound extraction step of extracting the target compound from the plurality of compounds based on the degree of similarity
　.
[Claim 39]
　A screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The storage step of storing the sixth feature quantity calculated using the feature quantity calculation method according to claim 1 in association with each other,
　and the binding compound which is a compound whose binding to the target biopolymer has been confirmed. A feature quantity calculating step of calculating a feature quantity of
　6, and a similarity calculating step of calculating a similarity between the sixth feature quantity of the plurality of compounds and the sixth feature quantity of the binding compound,
　A compound extraction step of extracting the target compound from the plurality of compounds based on the similarity
　.
[Claim 40]
　A screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The storage step of storing the seventh feature quantity calculated using the feature quantity calculating method according to claim 1 in association with each other,
　and the binding compound which is a compound whose binding to the target biopolymer has been confirmed. A characteristic amount calculating step of calculating a characteristic amount of
　No. 7, and a similarity degree calculating step of calculating a degree of similarity between the seventh characteristic amount of the plurality of compounds and the seventh characteristic amount of the binding compound;
　A compound extraction step of extracting the target compound from the plurality of compounds based on the similarity
　.
[Claim 41]

　24. 　A screening method for extracting a target compound that binds to a target biopolymer from a plurality of compounds, comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The storage step of associating and storing the eighth feature quantity calculated using the feature quantity calculating method according to claim 7,
　and the binding compound which is a compound whose binding to the target biopolymer has been confirmed. A feature quantity calculating step of calculating a feature quantity of
　8, and a similarity calculating step of calculating a similarity between the eighth feature quantity of the plurality of compounds and the eighth feature quantity of the binding compound,
　A compound extraction step of extracting the target compound from the plurality of compounds based on the similarity
　.
[Claim 42]
　A screening program for causing a computer to execute the screening method according to any one of claims 28 to 41.
[Claim 43]
　It is a screening device which extracts the target compound which couple|bonds with a target protein from
　a some compound, Comprising: About each of the said some compound, the three-dimensional structure of the compound by a plurality of atoms, and the said three-dimensional structure of the said compound. The first feature amount is calculated for a storage unit that stores the first feature amount calculated using the feature amount calculation method in association with
　each other, and a ligand that is a compound whose binding to a target protein is confirmed. a feature amount calculation unit, for
　a similarity calculation unit which calculates the similarity between the first feature amount for said ligand and said first feature value for said plurality of compounds,
　the similarity from the plurality of compound And a compound extraction unit that extracts the target compound based on the degree
　.
[Claim 44]
　A screening device for extracting a target compound that binds to a target protein from a plurality of compounds,
　comprising: for each of the plurality of compounds, the three-dimensional structure of the compound by a plurality of atoms, and the first invariant of the three-dimensional structure of the compound.
　The first invariant of the storage unit that stores a quantified feature amount in association with the ligand that is a compound whose binding to the target protein has been confirmed, using the feature amount calculation method according to claim 3. A feature amount calculation unit that calculates a feature amount, and
　a similarity degree calculation unit that calculates a degree of similarity between the first invariantized feature amount for the plurality of compounds and the first invariantized feature amount for the ligand. ,
　a compound extraction unit configured to extract the target compound, based on the similarity from the plurality of compound
　screening apparatus comprising a.
[Claim 45]
　The screening device according to claim 43 or 44, wherein the compound extraction unit extracts a compound whose similarity is equal to or higher than a threshold value.
[Claim 46]
　The screening device according to any one of claims 43 to 45, wherein the compound extraction unit extracts compounds in descending order of similarity.
[Claim 47]
　A screening device for extracting a target compound that binds to a target protein from a plurality of compounds,
　wherein the compound three-dimensional structure of a plurality of atoms and the feature amount calculating method according to claim 2 are used for each of the plurality of compounds. The storage unit that stores the first feature amount calculated in association with the storage unit and the
　pocket structure of the target protein by using the feature amount calculating method according to claim 5. A feature amount calculating section
　for calculating, a similarity degree calculating part for calculating a degree of similarity between the first feature amount for the plurality of compounds and the second feature amount for the pocket structure, and the
　plurality of compounds To a compound extraction unit that extracts the target compound based on the similarity
　.
[Claim 48]
　A screening device for extracting a target compound that binds to a target protein from a plurality of compounds,
　wherein the compound has a three-dimensional structure of a plurality of atoms for each of the plurality of compounds, and the feature amount calculating method according to claim 3 is used. The storage unit that stores the first invariant characteristic amount calculated in association with the storage unit and
　the pocket structure of the target protein using the feature amount calculation method according to claim 6, feature amount calculation unit for calculating the quantity and,
　similarity calculation unit for calculating a similarity between the second invariant feature value for said first invariant feature value and the pocket structure for the plurality of compound And
　a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 49]
　The screening device according to claim 47 or 48, wherein the compound extraction unit extracts a compound whose similarity is equal to or higher than a threshold value.
[Claim 50]
　The screening device according to any one of claims 47 to 49, wherein the compound extraction unit extracts compounds in descending order of similarity.
[Claim 51]
　A screening device for extracting a target compound that binds to a target compound that binds to a target biopolymer other than a protein from a plurality of compounds,
　wherein for each of the plurality of compounds, a three-dimensional structure of the compound with a plurality of atoms, and the compound The storage unit that stores the three-dimensional structure of the third feature amount calculated by using the feature amount calculation method according to claim 8 in association with the
　target biopolymer other than the protein. A characteristic amount calculation unit that calculates the third characteristic amount of the binding compound that is a confirmed compound; and a third characteristic amount of the
　plurality of compounds and the third characteristic amount of the binding compound. 　A screening device
　comprising: a similarity calculation unit that calculates a similarity; and a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
.
[Claim 52]
　A screening device for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　wherein each of the plurality of compounds has a three-dimensional structure of the compound with a plurality of atoms and the three-dimensional structure of the compound. The storage unit that stores the fourth feature amount calculated by using the feature amount calculation method according to claim 4,
　and the binding compound that is a compound whose binding to the target biopolymer has been confirmed. A feature quantity calculating unit that calculates a feature quantity of
　4; and a similarity degree calculating unit that calculates a degree of similarity between the fourth feature quantity of the plurality of compounds and the fourth feature quantity of the binding compound,
　And a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 53]
　A screening device for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　wherein the compound has a three-dimensional structure of a plurality of atoms and the three-dimensional structure of the compound for each of the plurality of compounds. The storage unit that stores the fifth feature amount calculated by using the feature amount calculation method described in (1)
　and the binding compound, which is a compound whose binding to the target biopolymer has been confirmed. A feature amount calculation unit that calculates the feature amount of
　5, and a similarity degree calculation unit that calculates the degree of similarity between the fifth feature amount of the plurality of compounds and the fifth feature amount of the binding compound,
　And a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 54]
　A screening device for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The sixth feature amount calculated using the feature amount calculation method according to claim 6, and a storage unit that stores the sixth feature amount in association with each other,
　and the binding compound that is a compound whose binding to the target biopolymer has been confirmed. A feature amount calculation unit that calculates the feature amount of
　No. 6, and a similarity degree calculation unit that calculates the degree of similarity between the sixth feature amount of the plurality of compounds and the sixth feature amount of the binding compound;
　And a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 55]
　21. A screening device for extracting a target compound that binds to a target biopolymer from a plurality of compounds,
　wherein, for each of the plurality of compounds, a three-dimensional structure of the compound with a plurality of atoms and the three-dimensional structure of the compound. The storage unit that stores the seventh feature amount calculated by using the feature amount calculation method described in
　(1) and the binding compound that is a compound whose binding to the target biopolymer has been confirmed. A feature amount calculation unit that calculates the feature amount of
　No. 7, and a similarity degree calculation unit that calculates the degree of similarity between the seventh feature amount of the plurality of compounds and the seventh feature amount of the binding compound;
　And a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 56]

　24. 　A screening device for extracting a target compound that binds to a target biopolymer from a plurality of compounds, comprising: for each of the plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms and the three-dimensional structure of the compound. The storage unit that stores the eighth feature amount calculated by using the feature amount calculation method according to claim 8 in association with the storage unit,
　and the binding compound that is a compound whose binding to the target biopolymer has been confirmed. A feature quantity calculating unit that calculates a feature quantity of
　8, and a similarity degree calculating unit that calculates a degree of similarity between the eighth feature quantity of the plurality of compounds and the eighth feature quantity of the binding compound,
　And a compound extraction unit that extracts the target compound from the plurality of compounds based on the similarity
　.
[Claim 57]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target protein from a
　plurality of compounds, comprising: for each of a plurality of compounds, a three-dimensional structure of the compound with a plurality of atoms and the first characteristic amount.
　A feature amount calculation for calculating the first feature amount by using the feature amount calculation method according to claim 2 for a storage step of storing in association with each other and a ligand that is a compound whose binding to the target protein has been confirmed. A step, a
　generator construction step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teaching data, and the first feature amount as an explanatory variable; and the
　ligand using the generator. And a compound three-dimensional structure generation step of generating a three-dimensional structure of the target compound from the first characteristic amount of
　.
[Claim 58]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds, the three-dimensional structure of a compound having
　a plurality of atoms, and the feature amount calculation according to claim 3, for each of the plurality of compounds. A storage step of associating and storing the first invariantized feature amount calculated by using a method, and the first invariantized feature amount
　for a ligand that is a compound whose binding to the target protein is confirmed. a feature amount calculation step of calculating,
　the three-dimensional structure of said plurality of compounds as teacher data, and generator constructing step of constructing the generator by machine learning as explanatory variables the first invariant feature value,
　said generated And a compound three-dimensional structure generation step of generating a three-dimensional structure of the target compound from the first invariantized characteristic amount of the ligand using a container
　.
[Claim 59]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds,
　wherein the first value is calculated for each of the plurality of compounds using the feature amount calculation method according to claim 2. And
　a feature amount calculation step of calculating the second feature amount using the feature amount calculation method according to claim 5 for the pocket structure of the target protein. And a
　generator construction step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the first feature quantity as an explanatory variable, and the
　pocket structure using the generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the second characteristic amount
　.
[Claim 60]
　It is a compound creation method of creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds,
　wherein each of the plurality of compounds is calculated using the feature amount calculation method according to claim 3.
　The second invariant characteristic amount is calculated by using the characteristic amount calculation method according to claim 6 for a storage step of storing the invariant characteristic amount of the target protein and the pocket structure of the target protein. a feature quantity calculation step,
　the conformation of the plurality of compounds as teacher data, and said first generator constructing step of constructing the generator by machine learning as explanatory variables invariants feature value,
　using the generator And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the second invariant characteristic amount of the pocket structure
　.
[Claim 61]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer other than a protein from a
　plurality of compounds, comprising: for each of a plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms; A storage step of storing a feature quantity in association with each
　other, and a binding compound that is a compound whose binding to the target biopolymer other than the protein has been confirmed, using the feature quantity calculation method according to claim 8. A feature amount calculation step of calculating
　the third feature amount, and a generator construction for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the third feature amount as an explanatory variable. A
　method for producing a compound,
　comprising: a step; and a step of generating a three-dimensional structure of the target compound from the third characteristic amount of the binding compound using the generator .
[Claim 62]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a plurality of compounds, the three-dimensional structure of the compound having
　a plurality of atoms, and the fourth characteristic amount for each of the plurality of compounds. The storage step of associating and storing,
　and a binding compound that is a compound whose binding with the target biopolymer is confirmed, the fourth characteristic amount is calculated using the characteristic amount calculation method according to claim 11. A feature amount calculating step of calculating, a
　generator constructing step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the fourth feature amount as an explanatory variable; and the
　generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the fourth characteristic amount of the binding compound
　.
[Claim 63]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a plurality of compounds, the three-dimensional structure of the compound having
　a plurality of atoms, and the fifth characteristic amount for each of the plurality of compounds. And a storage step of storing the associated
　biomolecule, and a binding compound that is a compound whose binding to the target biopolymer has been confirmed, by using the feature amount calculation method according to claim 14. A characteristic amount calculating step of calculating; a
　generator constructing step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the fifth characteristic amount as an explanatory variable; and the
　generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the fifth characteristic amount of the binding compound
　.
[Claim 64]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a
　plurality of compounds, comprising: for each of a plurality of compounds, a three-dimensional structure of the compound having a plurality of atoms; For the storage step of storing the,
　and the binding compound that is a compound whose binding to the target biopolymer is confirmed, the sixth characteristic amount is calculated using the characteristic amount calculation method according to claim 17. A characteristic amount calculating step of calculating, a
　generator constructing step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the sixth characteristic amount as an explanatory variable; and the
　generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the sixth characteristic amount of the binding compound
　.
[Claim 65]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a plurality of compounds, the three-dimensional structure of the compound having
　a plurality of atoms, and the seventh characteristic amount for each of the plurality of compounds. The storage step of storing the,
　and the binding compound, which is a compound whose binding to the target biopolymer has been confirmed, by using the feature amount calculation method according to claim 20. A characteristic amount calculating step of calculating, a
　generator constructing step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the seventh characteristic amount as an explanatory variable; and the
　generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the seventh characteristic amount of the binding compound
　.
[Claim 66]
　A compound creation method for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a plurality of compounds, the three-dimensional structure of the compound having
　a plurality of atoms, and the eighth characteristic amount for each of the plurality of compounds. The storage step of associating and storing,
　and a binding compound that is a compound whose binding to the target biopolymer has been confirmed, by using the feature amount calculation method according to claim 23. A characteristic amount calculating step of calculating, a
　generator constructing step of constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the eighth characteristic amount as an explanatory variable; and the
　generator. And a compound three-dimensional structure generating step of generating a three-dimensional structure of the target compound from the eighth characteristic amount of the binding compound
　.
[Claim 67]
　A compound creation program for causing a computer to execute the compound creation method according to any one of claims 57 to 66.
[Claim 68]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target protein from a
　plurality of compounds, wherein, for each of a plurality of compounds, a three-dimensional structure of the compound by a plurality of atoms and the first characteristic amount are provided.
　A feature amount calculation for calculating the first feature amount using the feature amount calculation method according to claim 2, for a storage unit that is stored in association with each other and a ligand that is a compound whose binding to the target protein has been confirmed. Section, a
　generator construction unit that constructs a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the first feature amount as an explanatory variable, and the
　ligand using the generator. And a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the first characteristic amount of
　.
[Claim 69]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds, the three-dimensional structure of a compound having
　a plurality of atoms, and the feature amount calculation according to claim 3, for each of the plurality of compounds. A storage unit that stores the first invariant characteristic amount calculated by using a method in association with each other, and
　the first invariant characteristic amount for a ligand that is a compound whose binding to the target protein is confirmed. a feature quantity calculating unit that calculates,
　with the three-dimensional structure as teacher data, the first generator construction unit for constructing a generator by machine learning as explanatory variables invariants feature value of the plurality of compounds,
　the product And a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the first invariantized characteristic amount of the ligand using a container
　.
[Claim 70]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds,
　wherein the first amount calculated using the feature amount calculation method according to claim 2 for each of the plurality of compounds. The feature amount
　calculation unit that calculates the second feature amount using the feature amount calculation method according to claim 5, for a storage unit that stores the feature amount of the target protein and the pocket structure of the target protein. And a
　generator construction unit that constructs a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the first feature amount as an explanatory variable, and the
　pocket structure using the generator. And a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the second characteristic amount of
　.
[Claim 71]
　A compound creation apparatus for creating a three-dimensional structure of a target compound that binds to a target protein from a plurality of compounds,
　wherein the first amount calculated using the feature amount calculation method according to claim 3 for each of the plurality of compounds.
　The second invariantized feature amount is calculated using the feature amount calculation method according to claim 6, with respect to a storage unit that stores the invariantized feature amount of the target protein and the pocket structure of the target protein. A feature amount calculation unit, a
　generator construction unit that constructs a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the first invariant feature amount as an explanatory variable, and the
　generator are used. And a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the second invariant characteristic amount of the pocket structure
　.
[Claim 72]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target biopolymer other than a protein from a
　plurality of compounds, wherein the three-dimensional structure of a compound having a plurality of atoms and the third structure
　The feature amount calculation method according to claim 8 is used for a binding compound that is a compound in which the binding between the storage unit that stores the feature amount and the target biopolymer other than the protein is confirmed. A feature amount calculation unit that calculates
　the third feature amount, and a generator construction that constructs a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the third feature amount as an explanatory variable. And
　a compound three-dimensional structure generation unit that uses the generator to generate a three-dimensional structure of the target compound from the third characteristic amount of the binding compound
　.
[Claim 73]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a
　plurality of compounds, wherein, for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms, and the fourth characteristic amount.
　For a binding compound that is a compound whose binding to the target biopolymer has been confirmed and a storage unit that stores, in association with each other, the fourth feature amount using the feature amount calculating method according to claim 11. A feature amount calculating unit for calculating, a
　generator constructing unit for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the fourth feature amount as an explanatory variable, and the
　generator. A
　compound creation device comprising: a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the fourth characteristic amount of the binding compound .
[Claim 74]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a
　plurality of compounds, wherein, for each of the plurality of compounds, the three-dimensional structure of the compound with a plurality of atoms, and the fifth characteristic amount. With
　respect to a binding compound that is a compound whose binding to the target biopolymer has been confirmed and a storage unit that stores the fifth feature amount using the feature amount calculating method according to claim 14. A feature amount calculating unit for calculating, a
　generator constructing unit for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the fifth feature amount as an explanatory variable, and the
　generator. A
　compound creation device comprising: a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the fifth characteristic amount of the binding compound .
[Claim 75]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a
　plurality of compounds, wherein, for each of the plurality of compounds, the three-dimensional structure of the compound by a plurality of atoms, and the sixth characteristic amount.
　For a binding compound that is a compound whose binding to the target biopolymer has been confirmed and a storage unit that stores the sixth feature amount using the feature amount calculating method according to claim 17. A feature amount calculating unit for calculating, a
　generator constructing unit for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the sixth feature amount as an explanatory variable, and the
　generator. A
　compound creation device comprising: a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the sixth characteristic amount of the binding compound .
[Claim 76]
　A compound creation device for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a plurality of compounds, the three-dimensional structure of a compound having
　a plurality of atoms, and the seventh characteristic amount for each of the plurality of compounds. For
　the binding compound, which is a compound whose binding to the target biopolymer has been confirmed , and a storage unit that stores, in association with each other, the seventh feature amount using the feature amount calculating method according to claim 20. A feature amount calculating unit for calculating, a
　generator constructing unit for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data, and the seventh feature amount as an explanatory variable, and the
　generator. A
　compound creation device , comprising: a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the seventh characteristic amount of the binding compound .
[Claim 77]
　A compound creation apparatus for creating a three-dimensional structure of a target compound that binds to a target biopolymer from a
　plurality of compounds, wherein, for each of the plurality of compounds, the three-dimensional structure of the compound by a plurality of atoms, and the eighth characteristic amount. With
　respect to a binding compound that is a compound whose binding to the target biopolymer has been confirmed and a storage unit that stores, in association with each other, the eighth feature amount using the feature amount calculating method according to claim 23. A feature amount calculating unit for calculating, a
　generator constructing unit for constructing a generator by machine learning using the three-dimensional structures of the plurality of compounds as teacher data and the eighth feature amount as an explanatory variable, and the
　generator. A
　compound creation device comprising: a compound three-dimensional structure generation unit that generates a three-dimensional structure of the target compound from the eighth characteristic amount of the binding compound .
[Claim 78]
　A non-transitory computer-readable recording medium that stores the feature amount calculation program according to claim 26.
[Claim 79]
　A non-transitory computer-readable recording medium storing the screening program according to claim 42.
[Claim 80]
　A non-transitory computer-readable recording medium that stores the compound creation program according to claim 67.