KR20000004909A - Computational method designing chemical structures having funtional characteristics - Google Patents

Abstract

PURPOSE: Computational method is to design chemical structures having common functional proprieties based on specific combinations of steric and binding affinity. CONSTITUTION: Computational process produces stimulated receptors which resembles biological receptors. The stimulated receptors are designed to exhibit optimized selective affinity for chemical structures are then generated and evolved to exhibit selective affinity for the stimulated receptors.

Description

Computer designed design of chemical structures with common functional characteristics

Bioreceptors are linear polymers of amino acids or nucleotides that form a three-dimensional envelope structure for binding to a substrate. The specific three-dimensional structure of such linear arrays and the placement of charges on the envelope surface are the result of evolutionary selection based on functional efficiency.

The selectivity of the bioreceptor depends on the difference in intensity of attraction and repulsion produced between the receptor and the substrate. The magnitude of this attraction and repulsion depends in part on the size and proximity of the charged sites on the substrate surface and the receptor surface.

Since the substrates differ both in terms of the number and size of charge sites present or induced on these surfaces, as well as in the spatial arrangement of such positions, the binding affinity can be varied depending on the structure of the substrate.

Substrates that exhibit similar binding affinity for the same receptor are likely to share a common spatial arrangement of at least some of the induced and fixed charge positions.

If the function of the receptor correlates with the binding affinity, the substrates with similar binding affinity will be functionally similar in their efficiency.

In this sense, it can be said that the receptors recognize or quantify the similarity between substrates.

Conventional molecular identification methods for identifying or discovering new chemical compounds or substrates for selective binding affinity to receptors are based on the discovery of molecular common subgraphs of active substrates and using them to predict new analogous compounds. .

The drawback of this method is that the substrates exhibiting similar binding efficiencies have similar structures.

In many cases, structurally dissimilar substrates may exhibit similar binding affinity for the same receptor.

More recent techniques based on quantitative structure-activity relationships (QSAR) are only suitable for developing new compounds within the same structural type and are mostly unsuitable for developing new molecular structures that exhibit the desired affinity for selection. [See, eg, Dean, Philip M. "Molecular Identification: Determination and Study of Molecular Similarity of Ligand-Receptor Interactions", Concepts and Applications of Molecular Similarity, Ed Mark A. Jason and Gerald M. Maggiora, pp. 211-238 (1990).

Recently, a great deal of effort has been devoted to designing atomic models of minireceptors composed of pseudoreceptors, which are connected to atoms and functional groups, or non-bonded sets of atoms or functional groups. [See: Snyder, J.P. (1993), 3D QSAR in Drug Preparation: Theory, Methods, and Applications; Kirbinai, H. Ed; Escum, Leiden, p.336).

As a related method, there is a method of enclosing a known target ligand with a plurality of model atoms and calculating an intermolecular attraction generated between a ligand and a receptor model. This model shows a high correlation between the calculated binding energy and the biological activity, but Walters, D.E. And Hinds, R.M. (1994) J. Medic. Chem. 37: 2527], which has not evolved to allow the production of new chemical structures with selective affinity for receptor models.

Therefore, it would be beneficial to provide a method for identifying unusual similarities that are necessary and sufficient (hereinafter also referred to as "necessity / sufficiency") to cause common properties between different chemical structures. Such a method can be used as a basis for creating a novel chemical structure of useful functional properties based on specific combinations of steric atom alignment and binding affinity.

[Summary of invention]

The present invention provides a method for identifying unusual / sufficient similarities necessary for causing common functionality between different chemical structures. The present invention also provides a method for producing novel chemical structures exhibiting similar functional properties.

The basic idea of the present invention is to use a two-step computer process to design or discover chemical structures of useful functional properties based on specific combinations of conformation and binding affinity.

In the first step of this process, emulation of antibody formation is used to generate a population of computer simulated receptors similar to bioreceptors with optimal binding affinity for a selected target substrate.

In the second step of the process, simulation or virtual receptors are used to evaluate the binding affinity of existing compounds or to design new substrates with optimal binding power.

The method of the present invention provides a simulated receptor that mimics selected features of the bioreceptor, including evolutionary processes that optimize binding selectivity.

Simulated or simulated receptors produced by the present invention can be used to confirm specific similarities between molecules. Like antibodies and other bioreceptors, simulated receptors produced in accordance with the present invention feature an excerpting mechanism. That is, they can be used to identify or identify common or similar structural features of target substrates.

The binding affinity between the receptor and the target substrate is used as a measure of characterization. Target substrates can be quantitatively classified based on binding affinity with specific simulated receptors. Compounds that share specific structural features will share similar binding affinity for the same virtual receptor.

The binding affinity between the bioreceptor and the substrate depends on the degree of conformation that fits well between the proximity receptor and the substrate surface (hereinafter referred to as "bonding"), the degree of water rejection between the nonpolar regions of the two surfaces, and the strength of static electricity generated between adjacent charge sites. Is measured. In some cases, covalent formation between the substrate and the receptor may contribute to binding affinity.

Simulation receptors generated according to this method mimic the binding mechanisms of the bioreceptors relative to them. The binding affinity value is calculated using the strength of the electrostatic attraction formed between the average proximity of the receptor and the target surface and the charged positions of both surfaces. The resulting binding affinity can be used to evaluate the similarity of substrate molecules.

Binding affinity can be measured collectively. That is, the binding affinity may depend on the interaction between the entire surface of the substrate and the closed receptor or receptor envelope that completely surrounds the substrate. In this case, the overall similarity analysis between substrates is suitable as a basis for developing a useful quantitative structure-activity relationship.

However, in most biological systems, affinity is measured locally rather than overall. The interaction between substrate molecules and bioreceptors is usually limited to the contact between the substrate surface and fragments of the receptor. In this situation, since only fragments of substrates are directly involved in binding affinity generation, it is not appropriate to analyze the overall similarity between the substrates as a way to develop a structure-activity relationship.

Locally spoken structures share similar structural fragments in similar relative positions and orientations. Locally similar structures are not always totally similar. Sampling of molecular properties includes full sampling, including assessment of total similarity; Fragment sampling, including local similarity assessment; And multiple fragment sampling strategies that evaluate both local and global similarities.

Local similarity analysis relies on sampling the discrete regions of the substrate for similar structures and charge distributions. In the case of bioreceptors, local sampling is caused by irregular irregularities on adjacent substrates and receptor surfaces.

In binding affinity measurements, the interactions between closely facing surfaces will prevail over the interactions between more distant sites. Proximity between adjacent surfaces also determines the strength of hydrophobic bonds. Effective simulation receptors produced according to the present method should utilize separate local sampling of the target substrate (molecule) to assess functionally related similarities between compounds.

Local similarity analysis is complicated by the following factors. That is, 1) the number, location, and identity of relevant fragments necessary / sufficient for specific binding affinity are not known by simply inferring from the chemical structure of the substrate; 2) The location and orientation of the sampled fragments depend on the underlying structure of the overall structure.

One feature of the method is the creation of a simulated receptor that has the ability to distinguish between similarities between chemical substrates, essentially finding a receptor that samples relevant fragments of substrates at relevant locations in space. The optimization process depends on four features of the simulation receptor:

1) generality: the receptor can bind to one or more substrates;

2) specificity: the binding affinity of the receptor depends on the substrate structure;

3) Phosphoricity: the receptor can distinguish substrates based on a few local structural features;

4) Mutability: Changes in receptor structure can alter binding affinity for specific substrates.

By encoding the receptor phenotype into a linear genotype that is represented by a string of strings, it is possible to facilitate mutation, recombination and genetic processing of the structural features of the simulated receptor.

Simulation receptors that meet these fundamental criteria can be optimized to have specific binding affinity for locally similar substrates using evolutionary selection strategies. This can be accomplished by encoding the spatial atomic arrangement and charge position distribution of the receptor in a genetic form that can undergo changes or mutations.

Like the bioreceptors, the simulation receptors produced according to the present method occupy a three-dimensional monopoly space. Such a three-dimensional space can be outlined at any exploded view by a one-dimensional path of sufficient length and twist. Proteins formed from linear polymers of amino acids are examples of such structures.

Similarly, the three-dimensional structure of the simulation receptor can be encoded into a linear array with turning instructions. The encoded one-dimensional form of the receptor is the genotype of this receptor. The translated form used for binding affinity evaluation is the phenotype of this receptor.

Changes (mutations) are made in the receptor genotype during the optimization process. The impact of these changes on the binding affinity of the phenotype is then assessed. After selecting genotypes that produce the desired binding affinity phenotype, additional changes are introduced by repeating the mutation and selection process until the degree of phenotypic selected optimization is achieved.

Various evolutionary strategies, including conventional genetic algorithms, can be used to generate populations of simulated receptors with optimal binding features.

Receptors produced according to the present methods are used to generate or identify novel chemical structures (compounds) that share useful specific properties of the target molecular species used as selection criteria in the preparation of simulation receptors.

By using the interaction with the receptor as a selection criterion, new chemical structures can be created that optimally conjugate with the receptor. Because these structures must meet the needs / sufficient requirements for receptor selectivity, they are likely to exhibit biological activity similar to the original molecular targets.

Simulated receptor populations of enhanced selectivity can also be used to screen high-affinity compounds from existing chemical structures that can share these useful properties. This method can also be used for the selection of compounds with selected toxic or immunogenic properties.

According to one aspect of the invention, to design a computer-based chemical structure having a preselected functional feature

(A) preparing a physical model of a simulated receptor phenotype encoded by a linear coding sequence and providing a series of target molecules that share one or more quantifiable functional features;

(B) for each target molecule

(i) calculating the affinity between the target molecule and the receptor for each of the various orientations using the effective affinity calculation;

(ii) adding up the calculated affinity to calculate the total affinity;

(iii) confirming maximum affinity;

(C) using the calculated total affinity and maximum affinity;

(i) calculating the maximum affinity correlation coefficient between the maximum affinity and the quantifiable functional feature;

(ii) calculating a total affinity correlation coefficient between the total affinity and the quantifiable functional feature;

(D) calculating a joint coefficient using the maximum correlation coefficient and the total sum correlation coefficient;

(E) changing the receptor structure and repeating steps (b) to (d) until a receptor population having a preselected conjugation coefficient is obtained;

(F) providing a physical model of a chemical structure encoded by a linear linear sequence of molecules, using the above effective affinity calculations, calculating the affinity between the chemical structure and each receptor for each of the various orientations, and calculating the calculated affinity. Calculating an affinity junction score using;

(G) changing the chemical structure to generate chemical structural variants and repeating step (f);

(H) A method comprising the step of selecting a chemical structural variant whose affinity score is close to a preselected affinity score and adding an additional change is provided.

According to another feature of the invention

(A) preparing a simulated receptor genotype by generating receptor linear coding sequences encoding space occupancy and charge;

(B) deciphering the genotype to generate a receptor phenotype, providing one or more target molecules with the selected functional feature, and using affinity calculations to calculate the affinity between each target molecule and receptor for various orientations, Calculate the total and maximum affinity between each target molecule and the receptor, and calculate the total affinity correlation coefficient for the functional feature versus the total affinity of the target molecule and the maximum affinity correlation coefficient for the functional feature versus the maximum affinity. Calculating a joint coefficient based on the sum and the maximum affinity correlation coefficient;

(C) mutating the receptor genotype and repeating step (b), selecting a receptor that exhibits an improved conjugation coefficient and mutating it until a group of receptors having a preselected conjugation coefficient is obtained;

(D) Calculate the affinity between each receptor for the various orientations and the chemical structure to be selected for the various orientations using the effective affinity calculation, calculate the total and maximum affinity between the compound and each receptor, and then determine one or more of the total and maximum affinity. Computing affinity junction scores by comparing the total and maximal affinity between one or more targets and the receptor population, and using the comparisons as a guide for the level of functional activity of the chemical structure for the one or more target molecules. A method for screening chemical structures with functional features is provided.

In accordance with another aspect of the present invention for designing a simulated receptor similar to a bioreceptor that exhibits selective affinity for compounds with similar functional features.

(A) preparing a simulated receptor genotype by generating receptor linear coding sequences encoding space occupancy and charge;

(B) deciphering the genotype to generate a phenotype of the receptor, providing a series of target molecules that share similar functional features, and using affinity calculations to calculate the affinity between each target molecule and receptor for various orientations; Calculate the total affinity and maximum affinity between each target molecule and the receptor, calculate the total affinity correlation coefficient for the functional trait versus the total affinity for each target molecule, and calculate the functional trait versus the affinity for each target molecule. Calculating a maximum affinity correlation coefficient for the maximum affinity and calculating a joint coefficient for each target molecule based on the sum and the maximum affinity correlation coefficient;

(C) mutagenizing the genotype and repeating step (b), and then selecting a receptor that exhibits an improved conjugation coefficient and mutating it until a population of receptors having a preselected conjugation coefficient is obtained. do.

In accordance with another aspect of the invention, a computer-based design for a chemical structure having preselected functional properties

(A) providing a series of target molecules that share a physical model of the receptor and one or more quantifiable functional features;

(B) for each target molecule

(i) calculating the affinity between the target molecule and the receptor for each of the various orientations using the effective affinity calculation;

(ii) adding up the calculated affinity to calculate the total affinity;

(iii) confirming maximum affinity;

(C) using the calculated total affinity and maximum affinity;

(i) calculating the maximum affinity correlation coefficient between the maximum affinity and the quantifiable functional feature;

(ii) calculating a total affinity correlation coefficient between the total affinity and the quantifiable functional feature;

(D) calculating a joint coefficient using the maximum correlation coefficient and the total sum correlation coefficient;

(E) changing the receptor structure and repeating steps (b) to (d) until a receptor population having a preselected conjugation coefficient is obtained;

(F) Providing a physical model of the chemical structure, calculating the affinity between the chemical structure and each receptor for each of the various orientations using the effective affinity calculation, and calculating the affinity junction score using the calculated affinity. Making a step;

(G) changing the chemical structure to generate chemical structural variants and repeating step (f);

(H) A method comprising the step of selecting a chemical structural variant whose affinity score is close to a preselected affinity score and adding an additional change is provided.

In accordance with another aspect of the present invention, a method of encrypting a chemical structure comprising atomic components is provided. This method includes providing a linear sequence of coding for space occupancy and charge for each atom of the chemical structure.

The present invention relates to a method for computer designing chemical structures that share common useful functional properties based on specific combinations of steric atomic arrangements and binding affinity. More specifically, the present invention relates to methods of making computer-simulated receptors that are similar in function to biological receptors (hereinafter abbreviated as "bioreceptors").

The simulated receptors are designed to have an optimized affinity for known target molecules and then chemical structures with selective affinity for these simulated receptors are generated according to evolutionary processes.

The invention will now be described with reference to the following drawings:

1 is a process diagram illustrating the relationship between genotyping and translation (genetic information translation) for the production of a corresponding phenotype that forms part of the present invention;

2 is a process diagram schematically illustrating the step of optimizing a receptor to selectively bind to a series of substrates using point mutations that form part of the present invention;

FIG. 3 is a process diagram schematically illustrating the steps of preparing a relevant population of receptors with selective binding affinity optimized for a series of chemical substrates and for producing a series of new chemical substrates with common functional features using such optimized receptors. ;

4A is a table showing some chemical compounds used in Examples relating to Ligand Preparation Example;

FIG. 4B is a diagram showing ligands 1.1-1.4 produced by the method of the present invention in the ligand production example, wherein each ligand has one or more orientations similar in structure to benzaldehyde;

FIG. 4C is a diagram showing ligands 2.1-2.4 produced by the method of the present invention in a ligand preparation example relating to chemical structure fabrication with mosquito repellent efficiency (hereinafter referred to as "mosquito insect repellent efficiency").

The invention can be divided into the following first and second parts:

Part 1-Generating simulation receptor populations with selective affinity for compounds with common functional features according to evolutionary methods;

Part 2-Create new chemical structures with common functional features.

The first part of the present invention includes several steps as follows:

(1) generating receptor genotypes and phenotypes;

(2) providing a known chemical structure (s) to a receptor (hereinafter referred to as "presentation");

(4) assessing the affinity of the receptor for chemical structure (s);

(5) assessing the selectivity of the receptor for the chemical structure (s);

(6) probabilistically develop a group of related receptors with selective affinity optimized for the chemical structure (s); Screening chemical substrates for toxicity and pharmacological activity and designing new chemical structure (s) with selective binding affinity for these receptors using optimized receptors.

In the following description of the best embodiments of the present invention, various tables of molecular and atomic radius, polarizabilities, effective polarity values and translation states and additional factors are mentioned, which are detailed in the last part of the description. It can be found in Tables I to V.

Flow diagrams showing non-limiting examples of process calculations are appended as modules 1-14 after the detailed description of the invention.

Part 1: Generation of simulated receptor populations with selective affinity for target molecules that share common functional features

(1) Genotyping and Receptor Phenotype Generation

Both simulated receptor genotypes and phenotypes are computer designed. Simulation receptor phenotypes consist of folded non-branched polymers with spherical subunits whose diameter is equal to the van der Waals radius (110 pm) of atomic hydrogen in length. The subunits can be connected to each other at any two of the six points corresponding to the intercepts of the spheres with their respective primary axes.

In this example, the connections between subunits cannot be extended or rotated and the center of the two connected subunits is always equal to their side lengths (ie 1 hydrogen radius).

Redirection occurs when two subunits are not attached to opposite sides of these common neighbors. Four types are available: right, left, right, up and down. The turn shall be made parallel to any of the primary axes. When redirection occurs at intersections with other subunits in the polymer, the subunits are allowed to occupy the same space as the other subunits for convenience of computer implementation.

The completed simulation receptor consists of one or more distinct polymers. In the case of receptors consisting of several polymers, the individual polymers may be at different space points.

For the convenience of computer implementation, all polymers making up a single receptor in this example are selected from polymers of the same length (= number of subunits). This limitation is not a functional requirement, and different polymers of different lengths can also be used for specific system modeling.

The structure of each polymer is encoded by a sequential set of redirection commands. This command confirms the individual redirection to the internal reference frame relative to the initial orientation of the first subunit in each polymer.

Instead of directly addressing the hydration of receptors and substrates in this example, it is assumed that the water molecules present at the binding site are permanently attached to the receptor surface and form an integral part of this structure. This is any estimate and it will be appreciated by those skilled in the art that this may be treated as a more accurate concept (see, eg, Van Os, 1995, Molecular Immunology 32: 199-211).

Reference is made to FIG. 1. This figure shows that the cryptographic creation module generates a random code string.

Each symbol determines a charge characteristic or responsiveness at one point in a three-dimensional form of a virtual receptor or indicates a direction change instruction. At least five different codes are needed to create a string describing the three-dimensional shape of the receptor based on a Cartesian (rectangular) coordinate frame structure. Other frame structures, such as tetrahedral structures, can also be made using different sets of redirection commands. The symbols indicate a redirection instruction defined for the current path of the virtual receptor structure in the three-dimensional space (ie, the instruction refers to the unique reference frame of the virtual receptor, not any external reference frame).

The redirection command is assigned to the current direction and orientation of the polymer, and only the reversal of left, right, up and down is allowed. If no redirection occurs, the polymer continues to progress or terminate in its current direction.

For rectangular systems, the minimum sign set is:

C ₁ = no redirection; C ₂ = right turn; C ₃ = turn left; C ₄ = phase rotation; C ₅ = turn down.

Such commands can be combined for diagonal turning. For example A _1,2 = C ₁ C ₂ ; A _2,1 = C ₂ C ₁ and so on.

The number of different symbols that determine different charge or reactive states is not limited and can be adjusted according to experimental results. Ciphers will vary in terms of length (number of codes) and how often certain codes appear in the series.

Example of Genetic Creation

It will be appreciated by those skilled in the art that the following examples of genotyping and phenotyping are used for illustrative purposes only. The following conventions are used in this embodiment:

(1) The code set used to make the code consists of five symbols representing the redirection command and two symbols identifying the charge location:

"0" = no redirection; "1" = right turn; "2" = phase rotation; "3" = left turn; "4" = turn down; "5" = positively charged position (no redirection); "6" = negatively charged position (no turn).

(2) Subunits are classified into two types, charged or uncharged. All charged subunits are assumed to have a single positive or negative charge. Such uniform charge size is arbitrary.

(3) The receptor consists of 15 distinct polymers. The length of the completed password is always a multiple of 15. The length of each polymer is equal to the total crypto length divided by 15. It will be appreciated that the receptor may be composed of several polymers of several or a certain length.

(4) The following variables are determined by the user: (a) total crypto length (and polymer length); (B) the frequency with which each code appears in the encryption string; And (c) occurrences of code combinations.

Module 1 is a flowchart showing a sample of genotyping.

(Example of Receptor Expression Creation)

Each genotype code is translated to produce a three-dimensional form of the corresponding phenotype or virtual receptor. Using a translation algorithm from a predefined starting point, the redirection instruction is converted into a series of coordinate triplets representing the locations in space of the successive subunits that make up the receptor polymer. Starting coordinates for each polymer should be given prior to translation. In translation, it is assumed that the centers of consecutive subunits are separated by the same distance as the covalent diameter of the hydrogen atom.

The translation operation decodes the cipher string continuously to generate continuous redirection and straight path parts. The interpretation of the continuous redirection to the external coordinate system depends on the previous redirection sequence. Each polymer constituting the acceptor is assumed to have the same initial orientation.

The translation operation of this embodiment is described in Table 1 showing the input and output states. In the absence of redirection, a new coordinate triplet is computed using the most recent values and new state for Δx, Δy, Δz. The charge position is considered to be a straight line (without redirection). The initial value of the old state is 20.

The following variables can be determined by the user:

A) starting coordinates for each polymer constituting the receptor;

The output can be stored with the following values:

A) 3 vectors (one for each axis: {x1, x2, x3 ... xn}, {y1 ... yn}, {z1 ... zn}).

B) three-dimensional binary matrix

C) a separate vector of charge position coordinates;

A sample process for cryptographic translation is shown in Module 2.

(2) target generation

The target is represented by a molecule composed of spherical atoms. Atoms are considered to be rigid spheres with the characteristic fixed radius of each atomic species. The radius of a rigid sphere whose repulsive force between the target atom and the virtual receptor is regarded as infinity is estimated by the exposed van der Waals radius given in Table 2. Other van der Waals radius estimates may be used instead of the values in Table 2.

The distance between the atom centers of two atoms connected by covalent bonds is expressed as the sum of their covalent radius. The covalent radius varies depending on the bonding order and atom type. Appropriate values of the covalent radius are shown in Table 3.

In the first estimate, the coupling length is assumed to be fixed (ie, the coupling vibration is to be ignored). Bond rotation is allowed and multiple atom arrays of the same structure are required for sampling with typical rotational states. The stability of the atomic arrangement is not taken into account because the bond with the virtual receptor can stabilize the atomic arrangement which would have been unstable in energy mechanics. Various energy minimization operations can be applied to target ligand generation.

The electrical charge resulting from the bond dipole moment is considered to be limited to the atomic nucleus. The negative charge is given by the atom with high electronegativity. The bipolar values used in this example are shown in Table 4. Instead of the values in Table 4, other estimates of other dipole values may be used.

(3) target presentation

The affinity of each target for the simulation receptor (s) is tested for several orientations of the target to the simulation receptor top surface. The upper surface is defined by the translation operation. Prior to binding affinity evaluation, the target should be contacted with the receptor.

Contact occurs when the distance between the center of at least one subunit of the receptor and the center of at least one atom of the target coincides with their combined van der Waals radius. To determine the relative position of the target and the receptor at the point of contact, the target is gradually moved toward the receptor surface along a path perpendicular to the receptor surface and through the receptor and the geometric center of the target.

When contact occurs, the target reaches the point of impact on the receptor. The distance between the target atoms and the receptor subunits is calculated using the translation position of the target atoms when the collision position is reached. These distances are used to calculate the strength and proximity of the electrostatic forces.

In this example it is assumed that the target moves in a straight line relative to the receptor and maintains its starting orientation upon contact. An alternative method was to gradually change its orientation as the target approaches the receptor so that the maximum affinity position at the point of contact is achieved.

Although this method is functionally similar to the present example, it is much more complicated in computer implementation. In this example, various orientations are tested with less effort in computer implementation. In this example, the pathway is mediated along the receptor's x and / or y axis to accommodate larger molecules. This feature is required for improved selectivity when different size molecules are tested for the same receptor.

Prior to the collision location calculation, the orientation of the target is randomized by randomly rotating the target by 6 ° around each x, y and z axis. Larger or smaller rotational frequencies may be used. This random orientation of the targets is unique to a given series of tests. The reliability of the optimization process depends on the number of target orientations tested and the number of target compounds evaluated. Sample processing of the target presentation is given in Module 3.

(4) affinity calculation

Estimation Strategy

This example is based on a simplified estimate that evaluates the basic components of affinity using relatively few computer operations. This estimate is illustrated in the description below.

However, it will be appreciated by those skilled in the art that more accurate affinity calculation procedures can be used to obtain more accurate affinity values. Known computer-operated packages can also be used directly in the process to calculate more accurate affinity values.

Research into crown ethers has shown that electron density distributions of smaller molecules can be used to describe the electron density of larger compounds. (Bruning, H. and Fail, D. (1991) J. Comput. Chem. 12: 1).

The Hirschfield shareholder method can be used to define strictly localized charge distributions and characterize them as charges and dipole moments (Hirsfield, F. L. (1977) Theor. Chim. Acta 44: 129). As a result, the total electron density distribution of a molecule can be divided into overlapping atomic portions, the size of which is related to the free atomic radius.

It can be demonstrated in the crown ether experiment that the main components of the electrostatic attraction are determined by local charge transfer rather than overall charge transfer between atoms. The charge distribution is mainly determined by the short range effects of different chemical bonds. In particular, non-adjacent atoms contribute little to the atomic dipole moment.

In addition, the interatomic charge transfer is affected by the electrostatic charge of the entire molecule, but only a very small influence on the charge distribution, as can be seen in the calculation results for the crown ether.

Calculation Shareholder atomic charges and dipole moments can be used to describe electrostatic attraction (Brunning, H. and Fail, D. (1991) J. Comput. Chem. 12: 1). Beyond the van der Waals radius there is only a minor contribution from the atomic quadrupole moment. Potential calculations that only consider atomic charges give very inferior results, but polar moments can be used to obtain better results.

Based on these considerations, the method of the present invention can estimate affinity between the target ligand and the simulated receptor (s) and between the simulated receptor (s) and the newly prepared chemical structure (s) based on the following two measures. It allows you to:

1. The magnitude of the electrostatic force generated between the charged subunits of the simulated receptor (s) and the atomic dipole of the target ligand (chemical structure). Since the charge subunits are assumed to have non-mobile unit charges, the magnitude of this electrostatic force is directly proportional to the size of the atomic dipole and inversely proportional to the distance between the atomic dipole of the ligand and the simulated receptor.

2. The ratio of nonpolar or uncharged subunits of the simulated receptor sufficiently close to the nonpolar region of the ligand to produce significant London dispersion.

Assumptions Used to Calculate Affinity in this Example

1. The chemical substrate targets evaluated in this example are assumed to be neutral (ie, not ionized) molecules. This is an arbitrary limitation, and the same methodology can be used to develop embodiments that can be applied to charged and uncharged targets.

2. It is assumed that the dipole moment is confined to the atomic nucleus. Similar affinity analysis can be performed assuming that the dipole moments are clustered in covalent bonds. According to Arlingham's group (1989), these assumptions are known to function the same.

3. The environment surrounding the virtual receptor is assumed to be a solvent system where the target acts as a solute. The target is effectively partitioned between the solvent and the virtual receptor.

4. At the moment the affinity is calculated, the target and the receptor are assumed to be stationary relative to one another, more specifically, in a fixed orientation.

5. The target is assumed to interact only with two types of positions on the surface of the receptor: fixed or negative (positive or positive) positions and non-polar positions.

Based on these assumptions, consideration should be given to the following contributions to the strength of the interaction:

1. The charge-dipole ^{- Q 2 μ 2/6 (} 4пε) 2 kTr 4

2. Charge-Non-polar-Q ² α / 2 (4пε) ² r ⁴

3. Bipolar-Nonpolar (Debye Energy)-μ ² α / (4пε) ² r ⁶

4. Nonpolar-Nonpolar (London Energy)-.75 [hνα ² / (4пε) ² r ⁶ ]

In this example, the constants are all ignored, so only relative strength is taken into account by estimation. The fixed charge position is also assumed to be single, positive or negative. Based on this, the four components can be relabeled in the following simplified form:

Charge-dipole-μ ² / r ⁴ or -μ / r ²

2. Charge-Nonpolar-α / r ⁴

3. Bipolar-Nonpolar (Debye Energy)-μ ² α / r ⁶ or -μα ^.5 / r ³

4. Nonpolar-Nonpolar (London Energy)-α ² / r ⁶ or -α / r ³

Generally, paragraphs 2 and 3 only make a small contribution to long range interactions. However, paragraphs 1 and 4 make a significant contribution to the interaction energy. Most of the interactions between nonpolar fragments in this example are assumed to occur between neighboring alkyl and aromatic hydrogens and nonpolar subunits of the acceptor. Under these conditions, the α value is assumed to be nearly constant.

Contribution of Hydrophobic Strength and Water Rejection

The solvation effect is an important consideration for binding affinity generation. For example, hydrophobic bond formation relies on the close spatial association of nonpolar and hydrophobic groups to minimize contact between hydrophobic regions and water molecules. Hydrophobic binding can contribute as much as half the total intensity of antibody-antigen binding. Hydration of receptors and substrate surfaces is also an important factor. Water bound to the polar position of the receptor or substrate surface can form crosslinks between the surfaces to increase affinity or interfere with the binding.

Hydrophobic interactions represent a strong attraction between hydrophobic molecules in water. In the case of receptor-target interactions, this interaction is considered to refer to the attraction between the proximal regions of the nonpolar receptor subunits and the nonpolar fragments of the target. The effect is mainly caused by the entropy effect that results in rearrangement of the surface such that water is excluded between adjacent nonpolar regions.

Accurate theoretical handling of hydrophobic interactions is not available, but hydrophobic attraction is estimated to contribute up to 50% of the total attraction between antibody and antigen. In order to estimate the hydrophobic interaction between the target and the virtual receptor, this example evaluates the proportion of receptors that are effectively protected from solvation by binding to the target. All nonpolar (uncharged) subunits within a fixed distance of nonpolar atoms on the target are considered to be protected from solvation by solvent molecules having diameters above the limit distance.

Combined affinity calculation

The combinatorial affinity calculation used in this example combines two interaction measures, the close intensity and the close intensity and the close measure of the charge-dipole interaction.

In this example, it is assumed that the affinity is isotropic. It will be appreciated by those skilled in the art that using affinity calculations of affinity can result in more discriminating, but more complex computer implementations.

The charge-dipole interaction is calculated as D = Σμ _i / r _ij ^v . Where _i is the dipole moment of the i th atom of the target, r _ij is the distance between the j th charge position on the acceptor and the i th atom, and the coefficient v can be set to 2, 3 or 4. The contribution of D to the overall affinity is more sensitive to charge separation at large v values.

Proximity scale is calculated as P = Σn _i / N. Where n _i is the number of uncharged subunits of the receptor at a maximum distance ∂ from the i th atom of the target with dipole moments ≦ 0.75 Debye. In this example ∂ can be 1 to 4 subunit diameters (which is close to the van der Waals radius of water). N represents the total number of subunits that make up the receptor.

The affinity value A is calculated from D and P using the formula A = [P (D + NP / k)] ^0.5 . Where k is the junction constant (k = 10000 in this example).

In the equation, P plays two roles. In the first example, this value is a weighting factor. This is a measure of "good conjugation," which acts to impart superior affinity values to the atomic arrangement in which the nonpolar regions of the receptor and target are in close contact. In such situations hydrophobic interactions and nonpolar interaction energies show high values and will also greatly contribute to the strength and stability of the bond. Under these conditions, the number of orbits the target can escape from the receptor is much smaller, and the residence time of the target will be longer.

As a second example, P is used to estimate the contribution of distributed energy to interaction strength. Dispersion energy is only important for uncharged and nonpolar sites and is assumed to be important only when the target and the receptor are located close to each other (ie within ∂). The k and ∂ values can be adjusted to change the relative contributions of P and D. In general, P is dominant for nonpolar targets, while D is more important for targets with large local dipoles. Hydrogen bonds are formed approximately when paired negative and positively charged receptor units interact simultaneously with the target hydroxyl, carboxyl or amine functionality.

Alternative Approaches to Affinity Calculations-Possibility of Binding Polarization

In some cases, it may be beneficial to introduce a variable corresponding to the relative polarizability of the target atom (hereinafter abbreviated as "polarization") to the affinity calculation.

In this case, the equation A = [P (D + NP 2 / k)] equation P ₂ = Σn _i / N is, P ₂ = Σα _i n _i / N is not to calculate the P ₂ eseo ^0.5.

Wherein n _i is the number of charged or uncharged receptor subunits that are separated by a maximum distance ∂ from the i th atom of the target atom, and α _i is the relative degree of polarization of the i th atom of the target. For the purpose of simplicity, α _H can be designated 1.0 for aliphatic hydrogen. If polarization is used, the k value must be adjusted. Sample polarizations based on the sum of the proximity coupling polarizations are shown in Table 5.

Since the degree of polarization is related to the potential of the electron cloud, the degree of polarization of the molecule can be calculated as the sum of the characteristic polarization degrees of the covalent bonds. This additional property applies to non-aromatic molecules that do not have non-localized electrons.

Alternative Technology-Functional Specificity

The affinity estimates used in this example can be replaced by functionally similar computer calculations that preserve the relationship between local charge, dispersion energy, and target-receptor separation. In addition, affinity measures for charged targets can be devised.

In this example, only non-covalent interactions are evaluated, but this method can be extended by including subunits in the virtual receptor that can cause specific covalent bond reactions with selected target functional groups. Module 5 shows a sample flow chart of the preferred affinity calculation method used in the present invention.

(5) Evaluation of selective affinity

The conjugation between the virtual receptor and the series of target substrates is assessed by comparing the known activity or affinity value for the target with the values obtained for the virtual receptor-target complex.

The maximum affinity of the optimal selectivity virtual receptor should be closely correlated with known affinity measures. You can improve this correlation by continuously repeating point mutations to improve this correlation between a series of substrates and virtual receptors (Figure 2), or by continuously using evolutionary processes to optimize the selectivity of the virtual receptor population. (Figure 3).

Known values may be any index known or assumed to depend on binding affinity. Non-limiting examples thereof include ED ₅₀ , ID ₅₀ , binding affinity and aggregation value. The values tested should be positive. Operational conversion of data may be required. Unweighted grading data cannot be used.

The optimal atomic arrangement of the target for maximum binding affinity is unknown before testing. To obtain a representative size of the receptor-target affinity range, each target must be tested again using any different atomic arrangement for the receptor surface. Use Module 4 for affinity assessment in each test.

The confidence of the obtained maximum affinity value depends mainly on the size of the sample, because the larger the sample, the more likely it is to have a true maximum. The same set of target atoms is used for each receptor test.

Two techniques are used in this example to avoid the need to use a large sample set for optimized receptor generation:

1) use of a combination of average (or sum) affinity and maximal affinity for the selection of highly selective receptors;

2) Repeat the optimization process sequentially to increase the number of test orientations incrementally (the optimization process starts from a small set of target orientations and tests more orientations as more acceptor receptors are created).

In this example, the sum of the affinity values obtained for all test orientations of each target is calculated. This total affinity score is a measure of the average affinity between the receptor and the target. At the same time, the maximum affinity value is also measured.

The correlation between the known value and the total affinity r _SA ² and the maximum affinity r _MA ² is calculated. Target compounds that do not show activity include an origin (0,0) in this correlation, assuming that they have little or no affinity for the virtual receptor. This assumption is not always valid and some tests may require different disturbances.

Correlation using total affinity is a measure of average fit. This correlation is large, but if the correlation between maximum affinity and known affinity is weak, the result suggests that the virtual receptor is not selective. That is, the various orientations of the target may effectively interact with the receptor.

In contrast, if the maximum affinity correlates very well with the known affinity value and the correlation with the total affinity is weak, the virtual receptor will be very selective. If both affinity and maximal affinity correlate very well with known affinity, it is likely that the sample orientation confirmed the receptor's characteristic response with limited error (errors of type I and type II reduced: false positive or false negative). Likely to result).

In some cases, it may be more appropriate to select an increased correlation between maximum affinity and known affinity while minimizing the correlation between known affinity and total affinity. Such screening may require subtracting the maximum affinity value from the total value to exclude these values from being used as a source of chaotic bias.

In this example, co-correlation values are used as criteria for receptor selection. This value is calculated as the square root of the sum of the sum affinity and the maximum affinity:

F = (r _MA ² xr _SA ² ) ^0.5

This value is optimized by the evolutionary process applied to the virtual receptor. Note: If the correlation between r _MA ² and r _SA ² is high, the value contributing to r _SA ² is closely correlated with the maximum affinity value, or it is a very insignificant contribution to the sum. Alternatively, calculate a correlation (r _SA-MA ) for known affinity (total affinity-maximum affinity) and maximize the value:

F = (r _MA ² x (1-r _SA-MA ² )) ^0.5

Using this scale will allow the selection of receptors with very high affinity for a very limited set of target orientations. Module 5 shows a flow chart of a sample joint calculation.

(6) Optimization process

The purpose of the optimization process is to produce virtual receptors with selective affinity for a series of target receptors according to evolutionary methods. The total number of possible genotypes containing 300 instructions is 7 ³⁰⁰ or about 10 ^253, so a highly efficient mechanism is needed to find the answer.

The following four steps summarize the steps of the optimization process, after which each step will be described in detail with the view calculation:

Step 1: Generate a series of random genotypes and select genotypes with minimal levels of activity. The selected genotype is used as the basis for further optimization using genetic operations (recombination) and unidirectional mutation techniques.

Step 2: Each of the selected genotypes is unique but mutated to produce a breeding population of related genotypes for recombination. The most selective mutant for recombination from the population is selected from the population.

Step 3: Form new genotype by recombination of selective mutants. The genotypes with the highest affinity junctions are selected from the resulting genotypes. This population of subpopulations is then used for recombinant or mutant production.

Step 4: After selecting the best recombinants, the selectivity is enhanced by repeated introduction of point mutations.

Phase 1: Evolutionary Process-Basic Password Generation

Genetic algorithms developed by Holland (Holland, J.H. (1975) Adaptation in Natural and Processing Systems, U. Michigan Press, Ann Avo) can be used to find the optimal solution to a variety of problems. This technique is usually applied using a random large set of solutions.

In this example, significant modifications are made to the technique to reduce the number of tests and the number of iterations needed to find high selectivity virtual acceptors. This was accomplished by using a very closely related set of genotypes as an initial population and applying a high degree of mutation at each iteration.

Unique receptors with optimal affinity characteristics can be developed for any set of target compounds. For example, receptors can be optimally bound in different orientations of the same target.

The use of early populations of closely related genotypes can increase the likelihood that the optimization process will result in a single solution. Recombination of unrelated genotypes may produce new genotypes of increased conjugation, but is much more likely to be out of course.

The purpose of the first step in the optimization process is to generate a genotype with a minimum level of affinity for the target set. This genotype is used to generate populations of related genotypes. A flowchart of a sample process for genotyping of minimum affinity is given in Module 6.

Phase 2: Evolutionary Process-Mutation of the Basic Code

Mutations of the genotype include changing one or more of the codes in the code. In this example, the mutation does not change the number of subunits that make up the receptor polymer and does not affect the length of the genotype. This rule is arbitrary and, of course, various variations may be useful in some systems.

Mutations can alter the folding pattern of the phenotype, resulting in changes in receptor-like space and binding site exposure or location. Mutations affecting the atomic arrangement of the marginal region of the phenotype may result in shifting the receptor center relative to the target center.

Neutral mutation

All mutations change the structure of the phenotype, but not all mutations change receptor functionality. Such neutral mutations can alter receptor components that do not affect affinity. In some cases, these neutral mutations can be combined with subsequent mutations to produce a synergistic effect.

Breed improvement population

The goal of the second stage of the evolutionary process is to create a population of distinct but related genotypes from the base genotype. Members of this population are then used for recombinant production.

This breed of individual population is produced by a number of mutations of the basic genotype. The resulting genotypes are translated and screened for their selectivity. Most selective products are separated for use in recombination. Module 7 provides a flow chart for the process of multiple mutation sample of genotypes.

Phase 3: Evolutionary Process-Recombination

The purpose of recombination is to generate new genotypes with improved conjugation. Recombination promotes the combination of new instructions while preserving genotypic fragments essential for phenotypic conjugation.

When recombination is used in conjunction with selection, selectivity is usually quickly optimized. Module 8 shows a flow chart of a sample process for genotyping recombination.

In this example, populations used for recombination are selected for testing in Step 7 of Module 8. This ensures that high selectivity genotypes are not replaced by low selectivity genotypes.

In this example, the mutation (Module 7) is applied to 50% of the recombinant genotype prior to the preliminary test (Step 7-Module 8). This step increases diversity within the recombinant population.

The test populations used in this example range in size from 10 to 40 genotypes. This corresponds to a relatively small population size. In some conditions a larger population may be required.

Phase 4: Evolutionary Process-Maturation

Progressive Micromutation Technology

The final step in the optimization process is similar to the antibody maturation process in the mammalian immune system. A series of single point mutations are introduced into the genotype and evaluated for its effect on phenotyping. Unlike recombination, this process only results in a small incremental change in the selectivity of the phenotype.

The maturation process is accomplished using the Rehenberg (1 + 1) evolutionary strategy (Lehenberg, I. (1973), the evolutionary strategy, F. Promman. Sterngat). At each stage of production, the degree of conjugation of the parent genotype is compared to that of the mutant product, and genotypes of greater selectivity are selected for the next generation.

As a result, this process is strictly monodirectional because less selective mutants do not replace their parents. Module 9 presents a flow chart of a non-limiting sample of genotyping.

During each iteration of the maturation process, only one command in the code is changed. If the parent and its mutant product have the same selectivity, the parent is replaced by that product in the next generation. This method results in the accumulation of neutral mutations that can synergize with subsequent mutations. This rule is arbitrary.

If recombination or maturation does not provide improved selectivity after the iteration process, it will be necessary to repeat step 2 to improve diversity of the cultivated population genome.

Selected Application

The process of the present invention can be used in several fields, including the following 1) and 2):

1) screening for compounds with selected pharmacological or toxicological activity; And 2) development of new chemical structures with selected functional features.

The above applications and examples thereof are described below.

1A) Screening Method

Receptor populations in which evolutionary processes have been applied to have a selective affinity for certain groups of compounds that share similar pharmacological properties can be used as probes for identifying other compounds of similar activity. In this case, the activity depends on the binding affinity. For example, receptor populations can be constructed according to evolutionary methods to exhibit specific affinity for sillylate.

If the affinity of these receptors for salicylate correlates closely with the affinity of cyclooxygenase to silicylate, these receptors will at least partially mimic the functionally relevant features of the cyclooxygenase molecule binding site. Thus, these receptors can be used to select other compounds with possible binding affinity with cyclooxygenases.

Such techniques can also be applied to screen for compounds for potential toxin or oncogenic activity. For example, receptors that mimic the specific binding affinity of steroid hormone receptors can be made according to evolutionary processes. These receptors can be used to assess their possible binding affinity prior to in vitro or in vivo testing of pesticides, solvents, food additives and other synthetics.

Simulation receptors may be prepared for affinity testing for alternative target sites, transport proteins or non-target binding.

1B) Screening for Min-Max Activity

In some cases, high affinity compounds may have toxic side effects or may be unsuitable for chronic administration. In this case, compounds with low binding affinity will be needed. Techniques such as permutation synthesis do not readily produce or identify these compounds. In contrast, simulation receptors can be used as an effective screening tool for structures with any defined level of binding affinity.

1C) Measurement of molecular similarity

Selectivity of a simulated receptor can be used as a quantitative measure of molecular similarity.

Simulation Receptor Example

Virtual test values of target affinity are selected to demonstrate the ability of the receptor generation program to produce simulated receptors similar to the activity of any of the randomly selected aspects in this example.

All receptors in this example consist of 15 polymers. Area, length and depth values specify the origin coordinates of these 15 polymers with respect to the receptor center.

Example 1

Simulation receptors of the following specifications were generated:

Number of subunits: 240; Area: 6; Length: 6; Depth: 25

password:

"4100033103212204103333424052312013341024124022232334010032242

51014405133243400324620412100131310043112101132412022421302413

23112433013310032305230004334140102022302140414443502652034131

0331022051414141021402134014310010231110331235210016240 "

Each target was tested 20 times for the receptor.

The affinity score for the optimized receptor was 0.9358, which is a relatively low value.

Target substrates used for the optimization of the receptors were benzene, phenol, benzoic acid and o-salicylic acid. Aspirin precursor o-salicylic acid is an inhibitor of prostaglandin synthesis by cyclooxygenase. Benzoic acid and phenol have much lower affinity for the same site. Target affinity values and scores for receptors are shown in Table A below. The simulated receptors in this table are shown to have maximum affinity for o-salicylic acid.

Target compound Target affinity Total Affinity Score Affinity score benzene 0.6 20.88 3.38 phenol 1.2 8.03 4.99 Benzoic acid 1.6 42.23 12.98 o-salicylic acid 4.4 80.33 34.71

Three test substrates were evaluated using a simulated receptor. Two of these compounds, m-salicylic acid and p-salicylic acid, are known to have lower activity than o-salicylic acid. The third compound, diflusinal, is a fluorinated salicylic acid derivative having an effect greater than or equal to salicylic acid. The evaluation results are shown in Table B below.

Target compound Total Affinity Score Affinity score m-salicylic acid 45.9 12.3 p-salicylic acid 63.5 27.5 Diflusinal 117 71.2 o-salicylic acid 80.33 34.71

The results obtained using the simulated receptors are in close agreement with the pharmacological data for these compounds: m-salicylic acid and p-salicylic acid have lower affinity scores than o-salicylic acid and diflusinal shows greater activity than o-salicylic acid. .

To further increase the certainty of such activity predictions, it will be necessary to further refine the simulation receptors and to use additional independently optimized receptors.

Part 2: Creating New Compounds with Selected Functional Features

Creating a New Ligand

Populations of simulated receptors generated with selective affinity for a series of target compounds with similar functional characteristics can be used to design novel compounds of similar characteristics. Provided that these features should be closely correlated with the structure or binding affinity of the model compound.

Using the interaction with the receptor as a selection criterion, new chemical structures can be developed to optimally conjugate with the receptor. Since these compounds must meet the needs / sufficient requirements for receptor selectivity, these new compounds are likely to have activities similar to those of the original molecular targets.

Process outline

1. A population of simulation receptors with optimized selectivity for a series of characterized target compounds is generated. In some cases it may be desirable to create several populations with different affinity features.

For example, a first simulated receptor population that mimics the properties of a selected target site, a second simulated receptor population that mimics the site needed to transport ligand to its original target, and a third simulated receptor population that mimics a target site that mediates undesirable side effects. Three populations can be created, including.

In this example, development of new ligand structures may require simultaneous optimization of affinity for the first two receptor populations and minimization of affinity for the third population.

2. Determine a new framework for the simulation receptor population (s).

3. Remodel the basic structure and evaluate the affinity using the simulated receptor population (s). If these modifications improve the affinity trait, the modified structure is screened for the next modification. If no improvement is found, try another modification. Modifications previously rejected may be reintroduced in combination with other modalities.

4. Repeat step 3 until a compound with a suitable affinity characteristic is obtained.

Note: Appropriate differentiating simulation receptors can be used to develop chemical structures with quasi-maximal affinity for selected target sites.

1) Generating Molecular Genotyping Codes

Encoding a ligand phenotype (molecular structure) in the form of a linear genotype represented by a code string may facilitate the process of mutation, recombination and inheritance of the structural features of the ligand during evolutionary process.

The ligand developed by this example consists of a substituted carbon skeleton. Each cipher consists of three code vectors. The first code vector includes a redirection instruction for carbon skeleton generation and determines the position of each carbon atom in the skeleton. The second code vector identifies the functional groups attached to each carbon atom. The third code vector defines the position of the functional group relative to the main carbon. Molecular backbones containing atoms other than carbon (eg ethers, amides and heterocycles) can be prepared according to homogeneous methods using additional symbols in the code to specify the atomic species replacing the carbon atoms in the backbone.

The carbon skeleton is made from a series of points that form the nodules of a three-dimensional tetrahedral coordinate system. The distance between the nearest points in the initial skeletal assembly is equal to the average bond length between alkyl carbon atoms.

First Code Vector: Ligand Skeletal Determinants

The first code vector is made up of symbols that identify the direction of conversion to the current atomic position. Each direction of conversion specifies the coordinates of the next atom in the tetrahedral matrix. Four directions (1, 2, 3, 4) corresponding to the unfilled valence of sp ³ carbon are taken from each atom. Each of these carbon atoms belongs to one of four possible states (A, B, C, D). These states correspond to the number of distinct nodules in the tetrahedral coordinate system.

The relationship between the new coordinates and the direction of transition for the next atom in the skeleton is given in the following table. The two tables B1 and B2 below specify two redirection rules required for ligand assembly.

The "Sailboat" rule results in the formation of a tetrahedral matrix in which a closed ring of 6 members (cyclohexane) is shaped like a sailboat. The "chair" rule results in the formation of a matrix in which the cyclohexyl ring is chaired. It is also possible to combine these two rules during cryptography. Only sailing boat rules are used in this embodiment.

If sailboat rule is used, new position after redirection when current position = (x, y, z) Current Status Conversion = 1 Conversion = 2 Conversion = 3 Conversion = 4 A (x-.75, y + .433, z-.5) (x + .75, y + .433, z-.5) (x, y-.864, z-.5) (x, y, z + 1) B (x + .75, y-.433, z + .5) (x-.75, y-.433, z + .5) (x, y + .864, z + .5) (x, y, z-1) C (x-.75, y + .433, z + .5) (x + .75, y + .433, z + .5) (x, y-.864, z + .5) (x, y, z-1) D (x + .75, y-.433, z-.5) (x-.75, y-.433, z-.5) (x, y + .864, z-.5) (x, y, z + 1)

Each redirection characterizes the state of the new atom.

New state after redirection

Current Status Conversion = 1 Conversion = 2 Conversion = 3 Conversion = 4 A B B B C B A A A D C D D D A D C C C B

If chair rule is used, new position after redirection when current position = (x, y, z) Current Status Conversion = 1 Conversion = 2 Conversion = 3 Conversion = 4 A (x-.75, y + .433, z-.5) (x + .75, y + .433, z-.5) (x, y-.864, z-.5) (x, y, z + 1) B (x + .75, y-.433, z + .5) (x-.75, y-.433, z + .5) (x, y + .864, z + .5) (x, y, z-1) C (x-.75, y-.433, z + .5) (x + .75, y-.433, z + .5) (x, y + .864, z + .5) (x, y, z-1) D (x + .75, y + .433, z-.5) (x-.75, y + .433, z-.5) (x, y-.864, z-.5) (x, y, z + 1)

Each redirection characterizes the state of the new atom.

New state after redirection

Current Status Conversion = 1 Conversion = 2 Conversion = 3 Conversion = 4 A B B B C B A A A D C D D D A D C C C B

Using these relationships, one-dimensional arrays of carbon atoms can be made by deciphering the first codevector consisting of the strings 1, 2, 3 and 4. Short side chains or ring structures can be made by folding the resulting carbon atom strings back or by closing them with closed loops. Specific ring structures (eg cyclohexane) can be incorporated directly into specific coding sequences as described below.

Second Password Vector: Substituent

A second code vector having the same length as the first code vector is used to determine the type of substituent attached to the carbon atom defined by the first code vector. Each substituent is identified by a single sign. Substituents are added one by one to the carbon skeleton. One carbon may have more than one substituent, but this is only possible if defined more than once by the first code.

In the present embodiment, all the valences not filled with the substituents defined by the second password vector are automatically filled by the hydrogen atoms during the ligand production process. Other rules may be applied to fill the empty valence with atoms other than hydrogen.

Third Vector: Substituted Bond Vectors

A third code vector having the same length as the first code vector is used to determine the valence used for attachment of the substituents defined by the second code vector. The third cipher is composed of symbols 1, 2, 3, and 4, respectively, which refer to the redirection instructions prescribed for the first cipher.

Substituents are assigned only if the valence is not already filled by the carbon atom defined by the first code vector or by a pre-designated substituent. Alternatively, consecutive substituents may replace previously assigned substituents.

2) Creation of password

The first code is made by generating a random sequence of symbols belonging to the set {"1", "2", "3", "4"} to make the carbon skeleton. Heterocyclic structures, ethers, amides, imides and carboxyl compounds can be prepared by substituting carbon atoms in the backbone with other atoms as defined in the second code.

The second code is generated from a random sequence of symbols identifying the type of substituent. The frequency of the codes may be fixed or any number prior to cryptographic generation.

The third cipher consists of the symbols belonging to the set {"1", "2", "3", "4"}. The ring structure can be assembled deliberately (not randomly) by adding specific code sequences to the first code. For example, "431413" encodes a cyclohexyl ring. A total of 24 strings encode all possible orientations of the cyclohexyl ring in the tetrahedral matrix. The second and third code vectors for the ring first code are produced as described above. Module 10 shows a flow diagram of a cryptographic example that produces a looped carbon skeleton.

The relative position of the input and output points from the ring that forms part of the carbon atom skeleton is determined by the length of the code sequence used to generate the ring. In particular, if the sequence comprises six symbols, such as 431413, the input and output points will be the same member of the ring. If the sequence partially repeats the initial six symbols and additionally contains additional symbols, the input and output points will not be the same member of the ring. For example, SEQ ID NOs: 4314134 and 43141343141 will produce rings in which the output point is located in a member of the ring adjacent to the input point.

In this example, the rings are added to the backbone by adding six or more signed sequences to the code. For the ring defined as 431413 the possible sequences are:

431413

4314134

43141343

431413431

4314134314

43141343141

431413431413

431413431413

Other chemical structures can be encoded in linear form using the provisions proposed for the preparation of novel ligand genotypes (for introduction of evolutionary processes into the ligand or for storage). For example, known drugs can be encoded in a linear form and used as a starting point for developing novel ligands with similar or improved functionality. Similarly, a series of drugs that interact with the common target site can be encoded in a linear form and used for recombination.

3) Code translation and ligand production

The coding vector is converted into the three-dimensional representation of the ligand according to the three steps of translation process.

In the first step, the carbon atom skeleton is manufactured using the first password. In the second step, substituents are added to the carbon skeleton using the instructions of the second and third password vectors. The instructions of the second and third code vectors also serve to replace other carbon atoms in the skeleton. Instructions in the second and third codes may change the orientation and number of available valences present on carbon or other atoms that form part of the base skeleton. For example, the addition of carbonyl oxygen fills two empty valences. In the third step, all valences not filled by substituents in the second step are filled with hydrogen atoms (unless otherwise specified).

First Detoxification: Ligand Skeleton Construction

In the first decoding process, the position of each carbon atom is specified by using the redirection command from the first code vector. The first atom is assumed to be located at the origin of the coordinate system. It is also assumed that the first atom occupies state A in the matrix.

Detoxification proceeds sequentially. The result of the first decoding process is a 3 x n matrix containing the x, y and z coordinates of each of the n carbon atoms in the skeleton. Because loops and inversions are allowed, the same location in space may be occupied by one or more carbon atoms. In this case, it is assumed that only one carbon atom occupies that position. As a result, the number of carbon atoms constituting the completed skeleton may be less than the number of codes in the first code vector.

As the first code is decoded a list is created from the second code that identifies the substituents attached to each carbon position. At the same time, a parallel list specifying the valences occupied by each substituent is created from the third code.

Second reading: adding a substituent

Substituents are added sequentially to each carbon atom based on the list prepared from the second code during the first readout process. Corresponding values from the third code are used to specify the valence position of the substituent relative to the main carbon.

If the position is already occupied by a nearby carbon atom or a predetermined substituent, no substitution is made. Alternatively, the translation process may be performed by performing substitution at the next unoccupied position or by newly replacing a predetermined substituent.

The distance between the substituents and the carbon atoms is calculated from the table of bond lengths. The position data and the bond length are used to calculate the coordinates of the substituents. In the case of multicomponent substituents such as hydroxyl, nitro and amino groups, the coordinates of each atom in the substituent are calculated on the basis of the main carbon.

After adding all substituents specified by the second password vector to the backbone, all remaining unfilled positions in the backbone are filled with hydrogen atoms. Calculate the coordinates of each hydrogen atom using the hydrogen sp ³ -carbon bond length.

One carbon atom may take one or more non-hydrogen substituents. This situation occurs when the same position is designated more than once by the first password vector. Although multiple substitution using the second password directly is readily feasible, this embodiment does not employ it.

Substitutions are only allowed at positions not occupied by the carbon atoms forming the ligand backbone. Create a cumulative list of all occupied positions in the tetrahedral matrix.

Make a list of the types, radii, and positions of all atoms that make up the ligand during the second readout. This list serves as the basis for future target generation.

At this stage of the decryption process, the availability of structures generated from cryptographic sequences is not evaluated. In some cases, atomic coordinates can be entered into energy minimization programs to produce more realistic structures. However, this example makes no assumptions about the atomic arrangement of the ligands during binding.

In addition, in this embodiment, the unique characteristic of the structure of a specific atomic arrangement of the same molecule is preserved. For example, in this example it is possible to distinguish between the three rotational isomers of butane and each isomer is treated as a unique molecule.

The coding vector constitutes the genotype of the corresponding ligand and can undergo mutation and recombination resulting in a change in ligand structure. Ligand constructs themselves are phenotypes used in the evaluation of binding affinity of virtual receptors to selected populations.

4) Target Presentation

The chemical structure or target ligand is assembled from the original randomly generated code. After the detoxification process, the coordinates, radius, dipole moment and polarization of each atom in the target ligand are obtained from the lookup table and used for the evaluation of binding affinity between selected populations of ligands and virtual receptors.

The affinity of the target for each of the virtual receptors is tested for various orientations of the target to the receptor surface. No assumptions are made about the relative orientation of the ligand and the simulated receptor. Prior to binding affinity evaluation, the target should be contacted with the receptor. The method of calculating the affinity between the target presentation and chemical structure and the simulated receptor is essentially the same as described above for the procedure of Module 4 for known target molecules and the simulated receptor.

5) Binding affinity and bond evaluation

The binding affinity of the target ligand for each simulated receptor used for conjugation evaluation is calculated using the same effective affinity calculation method as described for the generation of simulated receptors using target molecules. As mentioned above, affinity calculations using other criteria may be incorporated into the conjugation test process, but the computer efficiencies and effects of the present invention are part of the use of the same effective affinity calculations for virtual receptor generation and chemical structure generation using simulated receptor populations. Depends.

6) Evolutionary process introduction into ligands

Level test of adhesion

The degree of conjugation between the selected population of the simulated receptor and the new ligand or chemical structure is assessed by comparing the target activity or affinity value for the ligand with the values obtained for the simulated receptor-ligand complex. The maximum affinity of the virtual receptor with optimal selectivity will be closely correlated with the target affinity measure. Evolutionary processes are repeated in order to improve this correlation.

The target value can be adjusted to any level of binding affinity. The ligand does not have to have the same binding affinity for all the virtual receptors used in the selection process. In this example, the target affinity is calculated using the maximum binding affinity of the optimized virtual receptor for the known substrate. For example, the target binding affinity can be tailored to 90% of the binding affinity of each member of the virtual receptor population for a particular substrate. Alternatively, if the interaction between the ligand and the virtual receptor should be minimized, the target binding affinity can be set to zero.

New ligands with specific binding affinity profiles can be selected by combining simulation receptors optimized for different sets of substrates and linking selected target affinity values with each receptor. Ligand conjugation determines how well the calculated ligand binding affinity matches the target affinity value. The optimization process maximizes ligand conjugation.

The optimal orientation of the ligand for maximal binding affinity is not known before testing. To obtain representative values for the receptor-ligand affinity range, each new ligand should be repeated tested with different random atomic arrangements on the receptor surface. Each test uses Module 4 as described in Part 1 for affinity assessment. In general, the maximum affinity value obtained depends on the sample size, since the larger the sample, the greater the likelihood of containing a true maximum.

Two techniques are used in this example to eliminate the need to use large sample sets for optimized new ligand or chemical structure generation:

1. Using a measure that combines maximum affinity and average (or total) affinity for ligand selection of an optimized affinity profile.

2. Successive iterations of the optimization process increase the number of test orientations incrementally. (Optimization starts from a small set of target alignments, testing more orientation as more ligands of greater conjugation are produced).

In this example, the sum is calculated for the affinity values obtained for all test orientations of each ligand. This total affinity score is a measure of the average affinity between the receptor and the ligand. At the same time the maximum affinity value is also measured.

The total and maximum affinity are used to test the degree of adhesion between the virtual receptor and the new ligand. The conjugation of each new ligand is evaluated according to the difference between the calculated values of maximum affinity and total affinity and the target values for these variables. In this example this value is calculated as the conjugation score for each novel ligand-simulated receptor pair:

+

Bondability is maximum when the bond score is zero. The target maximum affinity and target total affinity are obtained from the regression function obtained during the development of the optimized virtual receptor as described above. Target values are obtained according to the following equation:

Target affinity = f x maximum affinity of the most potent substrate used to generate the virtual receptor

Target affinity = f x total affinity of the most potent substrate used to generate the virtual receptor

(Where f = scaling factor)

If more than one simulated receptor is used for ligand conjugation evaluation, the conjugation scores of each ligand-simulated receptor pair are summed.

In this case, the bonding is maximized when the sum of the bonding scores is zero. In some cases, it may be desirable to use only the maximum affinity score when testing new ligands for different simulated receptors. In this case the junction is calculated according to the following formula:

The junction in this case is also maximized when the sum of the junction scores is zero. Other methods, such as using geometric means, can also be used to determine the total conjugation of ligands tested for a series of simulated receptors.

By using both the maximum affinity value and the total affinity value obtained for each simulated receptor, the selectivity of the virtual receptor can be implicated in the ligand conjugation evaluation. This not only reflects the affinity of the ligand, but also reflects the conformational requirements of the virtual receptor, which is the standard of selectivity.

6a) Optimization process

purpose

The purpose of this process is to evolutionarily generate new ligands with selected target affinity for a series of simulated receptors. The total number of possible genotypes, including 25 instructions, is 25625, which requires a highly effective solution-finding mechanism.

fair

(1) Step 1: Generate a series of random genotypes encoding ligands and select for a series of simulated receptors to select ligands that exceed the conjugation limit.

(2) Step 2: The selected genotype is used as the basis for subsequent optimization using genetic computation (recombination) and unidirectional mutation techniques. By mutating the selected genotypes, a group of breeders of genotypes that are unique for recombination but are related to each other are produced.

(3) The mutants with the highest selectivity are selected from the population and used for recombination.

(4) Step 3: Recombine the selective mutants to generate a new genotype. From the resulting genotypes, the genotypes with the highest affinity junctions are selected. This subpopulation is used for the next recombination (step 3 repeat) or for mutagenesis (step 4 repeat).

(5) Step 4: Improve selectivity by taking the best recombinant product and applying repeated point mutations.

(6) The optimization process is completed when the ligand of the desired conjugate is produced.

Phase 1: Evolutionary Process-Generation of the First Cryptography

The purpose of the first step of the optimization process is to generate genotypes with the lowest levels of conjugation and corresponding ligand phenotypes. This genotype is used to generate related genotype populations.

Genetic operations developed by Holland can be used to find optimal solutions to various problems. This technique is usually applied using a large random set of solutions.

In this example, the technique is significantly modified to reduce the number of tests and the number of iterations needed to find ligands of high selectivity. Such modifications are achieved by using a series of closely related genotypes as initial populations and applying high mutation rates at each iteration.

For any set of target compounds, separate ligands can be developed that have optimal affinity properties. For example, receptors can optimally bind to the same targets in different orientations. Using an initial population of closely related genotypes greatly increases the likelihood that the optimization process will be concentrated on a single solution. Recombination of unrelated genotypes can produce new genotypes with increased conjugation, but with greater potential for branching.

Step 2: Ligand Mutation

The goal of the second step in evolutionary processes is to create a population of unique and related genotypes derived from the base genotype. Members of this population are used to generate recombinants. This breeding population is created by multiple mutations of the basic genotype. The resulting genotypes are translated and their selectivity is selected. Products with the highest selectivity are used for recombination.

Ligands are mutated by changing the code in the genotype (coding vector) encoding the ligand structure. Mutations change the shape of the ligand as well as the type of functional group present on the ligand and the position of the functional group. In this example, mutations can change the number of carbons that make up the ligand backbone. Module 11 is a flowchart of a sample process for multiple point mutations.

Mutations can alter the folding pattern of ligand phenotypes resulting in changes in the shape and exposure or location of the functional group. Mutations affecting the arrangement of the marginal regions of the ligand phenotype can result in position shifts to the receptor center.

Neutral mutation

Not all mutations change the structure of the phenotype, but not all mutations change ligand functionality. Such neutral mutations can alter ligand components that do not affect affinity. In some cases, these neutral mutations can be combined with subsequent mutations to produce a synergistic effect.

Sequence mutation

Sequence mutations do not directly change the code within the code but instead rearrange the sequence of code within the code. Sequence mutations can alter the size, structural form and presence and position of functional groups of the ligand. Four types of sequence mutations are used in this example:

a) Delete: Remove a sequence of codes from a password.

ABCDEA ⇒ ABEA

b) inversion: reverse the order of the symbols that make up the sequence within the code.

ABCDEA ⇒ ABDCEA

c) redundancy: repeating a sequence of codes that form part of a cipher.

ABCDEA ⇒ ABCDCDEA

d) Insertion: Insert a sequence of codes into a cipher.

ABCDEA ⇒ ABCDBCEA

In this example, the mutations are applied in combination. Module 12 shows a flow chart of sample sequence mutations.

Step 3: Generation of Recombinant Passwords

During the recombination process, randomly selected complementary sites are exchanged between selected genotypes. The purpose of recombination is to create novel genotypes with improved conjugation. Recombination preserves the genotyping fragments essential for phenotypic conjugation while allowing the introduction of a combination of new instructions. In general, the use of recombination in combination with selection can result in rapid optimization of selectivity. Module 13 shows a sample flow chart for recombination.

In this example, the population used for recombination is taken and tested. This prevents the replacement of high selectivity genotypes with low junctional genotypes. In this example, multiple mutations are applied to 50% of the recombinant genotype prior to testing. This process increases diversity within recombinant populations.

The test population used in the Examples ranges in size from 10 to 40 genotypes. This corresponds to a relatively small population size. Under some conditions, larger populations may be needed.

Step 4: Ligand Maturation

Progressive Micromutation Technology

The final stage of the optimization process is similar to the antibody maturation process of the mammalian immune system. A series of single point mutations are introduced into the genotype and their effects on phenotypic conjugation are evaluated. Unlike recombination, this process only results in a small incremental change in the selectivity of the phenotype.

The maturation process is achieved using Rehenberg (1 + 1) evolutionary strategy. At each stage of production, the degree of conjugation of the parental genotype is compared to that of the mutant product, and a greater selectivity genotype is selected for the next generation.

As a result, this process is strictly unidirectional because less selective mutants do not replace their parents. During each iteration of the maturation process, only one command in the code is changed.

If the parent and the mutant product have the same selectivity, the parent is replaced by that product in the next generation. This method results in the accumulation of neutral mutations that can synergize with subsequent mutations. This rule is arbitrary. Module 14 presents a flow chart of the sample maturation process.

If recombination or maturation does not result in improved selectivity after the iteration process, multiple mutations (step 2) will need to be repeated to improve diversity of the cultivated population genome.

Ligand Generation Example

survey

Mosquito Aedes aegypti can be kicked out by benzaldehyde. Benzene and toluene show much lower levels of mosquito repellent effects (Table 1). Cyclohexane or hexane does not show significant insect repellent effect on this species (Table 1).

In the novel ligand production test below, this method is used to generate the first benzaldehyde-like compound in terms of insect repellent activity.

In the first stage of ligand production, we simulate a receptor that has a high affinity for benzaldehyde and a low effect on benzene. In the second step, a simulated receptor binding affinity develops ligands similar to benzaldehyde.

Mosquito reaction

Mosquitoes were incubated in the laboratory (parental bodies 7-14 days old, not fed). The test was run between 12:00 and 17:00 EDT. The test population used for the four sets of tests consisted of 200, 175, 105 and 95 female mosquitoes. The mosquitoes were given a drink.

The test was carried out in a 35 x 35 x 35 cm transparent plexiglass box with walls facing two screens. The insect screen consists of two layers: an inner layer of coarse plastic mesh and an outer layer of fine nylon mesh. The box was placed in the gas hood to allow air to enter either side of the screen and to the other side. Air flow was <0.5 cm / sec.

The mosquitoes sat on the wall of the box with their heads upwards. Triangular pieces (4 x 4 x 1 mm) of Whatman # 1 filter paper were used to provide stimulant compounds. The tip of the filter paper was immersed in the test solution to a depth of 0.5 cm and used immediately. The response to the test solution was measured as follows:

1. A female mosquito sitting stationary on the inner screen of the backwind was selected for testing.

2. The end of the treated filter paper was placed on the outside of the screen and opposite the mosquito's midthoracic section. In all cases, the initial approach was from the bottom of the mosquito's sitting position.

3. The tip of the filter paper was maintained for up to 3 seconds and the reaction of the mosquito was observed.

This process was repeated for other mosquitoes. Mosquitoes were tested only once a day with each compound. The filter paper ends were replaced with new ones after each 5 trials (total duration of 30 seconds or less). Compounds were tested in random order and each compound was tested twice on separate days.

Two sets of controls were performed using the ends of the untreated (dry) filter paper and the ends of the filter paper moistened with distilled water, arranged in the same way as the treated filter paper. This control was regularly dispersed during the insect repellent compound test. The response to the control test did not change during the course of the experiment (p> 0.25).

Four behavioral responses were recorded:

1. No Reaction: Mosquito stays in motion.

2. Takeoff: Mosquitoes fly away from their rest.

3. Hibernating leg elevation: Mosquitoes raise the middle chest leg in the same direction as the stimulant compound source.

4. Rear leg elevation: Mosquito raises the middle chest leg in the opposite direction to the source of stimulant compound.

Following hibernation bridges, takeoffs frequently occurred and all of these actions were recorded. Polyethylene gloves were worn during all steps and tests of compound preparation.

Mosquito Reaction to Selected Volatile Compounds compound Boiling Point (℃) N % flight % Leg lift Relative Response Insect Rate * Benzaldehyde 178 130 90 10 178 benzene 80 72 72 12.5 68 toluene 110 166 67 27 94 Cyclohexane 81 80 6 0 4.9 Hexane 69 100 4 0 2.8 Contrast test (blank) - 450 5 0

Relative insect repellent rate = [(% flight reaction +% leg reaction) x boiling point] / 100

Simulation receptor and ligand generation

Two simulation receptors were prepared using the same selection criteria. Each receptor was independently used for the preparation of a series of ligands.

Molecular Assembly 1

Step 1: Create Receptor

Receptors with selective affinity for benzaldehyde were generated by evolutionary methods. Benzene and benzaldehyde were used as test targets. Affinity values were calculated using the 15 orientations of each target. The results of the evolutionary process are as follows:

Target Active level Total affinity Affinity benzene 1.0 6.87 2.21 Benzaldehyde 5.9 75.87 13.02

The affinity score for the receptor was 0.992.

The code for the optimized 25 x 6 x 7 benzaldehyde receptor is as follows:

231014406145 053400324221 412100131300 063112101132

412061421302 413231124335 133100333032 300043541401

022224-31514 143431012321 341310334122 101414141021

402131114311 010233120331 260214016231

Step 2: Create Ligand

New ligands were generated in an evolutionary manner using optimized simulation receptors as templates. Four different ligands were assembled by random mutation and selection. Ligands were selected for similarity with benzaldehyde. The affinity values of the ligands are as follows:

Benzaldehyde Ligand 1.1C ₉ H ₁₇ C _l2 OH Ligand 1.2C ₈ H ₁₅ Cl Ligand1.3C ₈ H ₁₃ Cl (= O) Ligand 1.4C ₁₃ H ₁₆ OH (= O) Total affinity 75.87 74.03 67.88 72.55 72.94 Affinity 13.02 12.82 15.14 12.58 11.2

The resulting ligands 1.1 to 1.4 are shown in Figure 4b. At least one orientation of each ligand was structurally similar to benzaldehyde.

Molecular Assembly 2

Step 1: Create Receptor

25 x 6 x 7 receptors with selective affinity for benzaldehyde were generated by evolutionary methods. Benzene and benzaldehyde were used as test targets. Affinity values were calculated using the 15 orientations of each target. The results of the evolutionary process are as follows:

Target Active level Total affinity Affinity benzene 1.0 25.88 8.53 Benzaldehyde 5.8 162.23 42.74

The affinity score for the receptor was 0.996.

The code of the receptor is as follows:

031264441313 004422243042 223140112054 302122330134

543301114446 210043042311 323431131340 130020120133

224223503403 432003432122 002221221113 411440003113

323030313214 002321144010 000243013133

Step 2: Create Ligand

New ligands were generated in an evolutionary manner using optimized simulation receptors as templates. Four different ligands were assembled by random mutation and selection. Ligands were selected for similarity with benzaldehyde. The affinity value of the ligand is as follows:

Benzaldehyde Ligand 2.1C ₈ H ₁₅ Cl (= 0) Ligand 2.2C ₉ H ₁₅ Cl (= 0) Ligand 2.3C ₆ H ₁₀ CN (= O) Ligand 2.4C ₉ H ₁₃ (= O) ₂ Total affinity 162.23 182.4 166.5 159.7 156.8 Affinity 42.74 48.97 43.0 39.0 40.5 Bond score 42.74 0.135 0.02 0.05 0.06

The ligands 2.1 to 2.4 thus obtained are shown in FIG. 4C. At least one orientation of each ligand was structurally similar to benzaldehyde.

Compounds 2.1 to 2.4 are substituted cyclohexanone derivatives. Ligand 2.2 is 5-chloro-2,7-nonadione and ligand 2.3 is 2-cyano-5-hexanone. Ligand 1.4 comprises a fragment corresponding to methylcyclohexylketone in structure. Test results of the insect repellent rate of cyclohexanone, menton, methylcyclohexylketone and 2-octanone (see FIG. 4A) suggest that these ligands also have insect repellent effects against mosquitoes (Table E2).

Mosquito Reaction to Selected Volatile Compounds compound Boiling Point (℃) N % flight % Leg lift Relative insect repellency rate * Benzaldehyde 178 130 90 10 178 2-octanone 2-acetylcyclo- 173 80 82 12.5 162 Hexanon 225 100 54 24 175 Cyclohexanone 156 134 99 One ＞ = 154 Menton 207 110 72 11 172 Control test (blank) - 450 5 0

Relative insect repellent rate = [(% flight response +% leg response) x boiling point] / 100

The methods described herein for creating new chemical structures of preselected functional properties or properties are described for illustrative purposes only. For example, this method may be readily implemented using other known or acceptable polarization degrees, dipole moments, covalent radius values, and the like.

The flow chart of the process calculation steps of the following modules is also described by way of example. For example, affinity may be calculated using a commercial computer calculation package that uses fewer estimates than described herein. New chemical structure generation methods include first generating one or more simulated receptors that exhibit a preselected affinity for known compounds with similar functional properties and using these receptors to generate new structures that exhibit such characteristics to the desired extent. do.

Receptors can also be used as tools for screening pharmacological or toxicological properties of known compounds other than for the production of new chemical structures. Thus, it will be apparent to those skilled in the art that various modifications may be made to the methods described herein within the scope of the present invention.

Transition state and additive factors Old state Additive factors New state for redirection = Δx Δy Δz Right lift Left Downward One 0 One 0 2 4 3 5 2 -One 0 0 15 6 One 24 3 One 0 0 One 7 15 22 4 0 0 -One 12 23 14 One 5 0 0 One 9 One 16 23 6 0 0 -One 11 20 10 2 7 0 0 -One 13 21 8 3 8 0 -One 0 7 9 24 14 9 -One 0 0 17 10 5 8 10 0 One 0 6 14 22 9 11 0 -One 0 22 16 6 12 12 -One 0 0 18 11 4 13 13 0 One 0 24 12 7 16 14 One 0 0 4 8 18 10 15 0 -One 0 3 17 2 18 16 One 0 0 5 13 17 11 17 0 0 -One 16 19 9 15 18 0 0 One 14 15 12 19 19 0 One 0 20 18 21 17 20 One 0 0 23 24 19 6 21 -One 0 0 19 22 23 7 22 0 0 One 10 3 11 21 23 0 -One 0 21 5 20 4 24 0 0 One 8 2 13 20

Calculation formula: input (old state, direction change) ⇒ output (Δx, Δy, Δz, new state)

Eg initial position (12, 34, -18); Input: sphere state = 10, turn = right:

Output: scene state = 6, Δx = 0, Δy = 1, Δz = 0, subsequent position (12, 35, -18)

Van der Waals radius element H F O N C Cl S Br P I Van der Waals Radius (pm) 110 140 150 150 170 180 180 190 190 200 Relative Radius (H = 0.5) 0.5 0.64 0.68 0.68 0.77 0.82 0.82 0.86 0.86 0.91

N.S. Based on Isaacs (1987), Physical Organic Chemistry, Longman Scientific and Technical, New York, 828 pp.

Covalent Radius (pm) Join order First order H28 1st 2nd 2nd 3rd aromatic B C N O F88 77 70 66 6466.5 60 5560.2 5570 First order Si P S Cl117 110 104 99 First order Br114

N.S. Based on values found in Isaacs (1987).

Sample effective dipole used for charge location assignment Combination atom Debye C-H HC +0.35 or + 0.84 * charge not assigned ArC-H HC + 0.6-0.366 or no charge assigned = C-H HC + 0.336-0.6 or charge not assigned * C = O OC -2.7 * charge not assigned or +1.35 C-O-C O -0.8 C-OH HO +1.5 or + 1.7-1.1 C-NH ₂ HN + 1.3-1.3 C-NO ₂ ON -2.0 + 4.0 C≡N NC -3.7 charge not assigned C-S-C S + 1.5 * in thiophene or dimethyl sulfide (can be negatively charged throughout the context) C-N = C N +1.5 or +1.3 in pyridine or CH -N = CH Ar-F or C = C-f F -1.3 C-f F -1.8 Ar-Cl or C = C-Cl C -1.7 C-Cl Cl -2.1 Ar-Br or C = C-Br Br -1.7 C-Br Br -2.0 C-I I -2.0

* Preferred under most conditions.

Each target atom is fully described as a set of eight values {x _i , y _i , z _i , r _i , br _i , cr _i , d _i , α _i }. (Where x _i , y _i and z _i are the coordinates of the geometric center of the molecule, r _i = van der Waals radius, br _i = bonding or covalent radius, cr _i = collision radius (= r _i + 0.5), α _i = polarization degree, d _i = effective dipole moment value.

Selected relative effective polarity for selected target atoms atom context Relative polarization degree (α _i ) H C-H 1.0 H N-H 1.1 H O-H 1.1 H S-H 3.0 * F C-F 1.5 * Cl C-Cl 4.0 Br C-Br 5.8 I C-I 8.9 * C C-CH ₃ 3.7 C C-CH ₂ -C 3.5 C C-CC ₂ -H 3.2 C C = CH ₂ 4.5 C C = CH-C 4.3 C C = CC ₂ 4.0 C C≡C-H 4.9 * C C≡C-C 4.6 * C Arengori 4.3 * or 2.6 (based on benzene (non-localized electron cloud)) C C-C≡N 4.0 C C ₃ -CO- 3.6 C C ₂ HCO- 3.8 C CH ₂ -CO- 4.1 C H ₃ -CO- 4.4

Module 1: Cryptography for Simulation Receptors

Step 1: Enter password generation variables: i) password length; And ii) command frequency.

Step 2: Set the initial value of the empty code string to store the password.

Step 3: Generate a random number

Step 4: Select codes {'0', '1', ...., '6'} based on the random number and the command frequency and connect them to the cipher string.

The fourth step is repeated until the string length is equal to the predetermined encryption length.

Step 5 Output the password.

Module 2: Cryptographic Translation for Simulated Receptors

Step 1: Enter the origin coordinates for the polymers that make up the receptor.

Step 2: Enter the password for the polymer.

Step 3: Decrypt the first code of the cipher.

Step 4: If the sign is a direction change command, determine the subordinate coordinates by using the translation operation. In all other cases, perform step 7.

Step 5: Store the subunit coordinates. The charge value 0 is assigned to the subunit.

Step 6: If the code is not the last code in the cipher, repeat step 3; otherwise, perform the first step.

Step 7: If the sign is a full load command, subordinate coordinates are determined using a translation operation on the assumption that there is no redirection.

Step 8: Store the subunit coordinates. Based on the sign, assign a charge of +1 or -1 to the subunit.

Step 9: If the code is not the last code in the cipher, repeat step 3; otherwise, perform the first step.

Step 10: Repeat steps 2 to 9 for each polymer constituting the receptor.

Step 11: Output the charge values and the coordinates of the subunits.

Module 3: Target Presentation

Step 1: Enter the coordinates and radius of the target atom (xt _i , yt _i , zt _i , radius _i )

(i = number of atoms in the target)

Enter the coordinates of the receptor (xr _j , yr _j , zr _j , charge _j )

(j = number of subunits in the receptor)

Step 2: Generate random angles (ㅿ θ, ㅿ Φ) and translation values (k _x , k _y ).

Step 3: Rotate and translate atomic coordinates by random amounts

Step 3a: The target coordinates are converted into the polar form.

(xt _i , yt _i , zt _i , radius _i ) ⇒ (θ _i , Φ _i , ρ _i , radius _i )

Step 3b: Add a random change to the angle.

(θ _i , Φ _i , ρ _i , radius _i ) ⇒ (θ _i + ㅿ θ, Φ _i + ㅿ Φ, ρ _i , radius _i )

Step 3c: Convert to rectangular coordinates.

(θ _i + ㅿ θ, Φ _i + ㅿ Φ, ρ _i , radius _i ) ⇒ (x _i , y _i , z _i , radius _i )

Step 3d: Add Random Translation.

(xn _i , yn _i , zn _i , radius _i ) ⇒ (x _i + k _x , y _i + k _y , z _i , radius _i )

Step 4: Concentrate the target coordinate on the origin (0, 0, 0).

Step 4a: find the maximum and minimum values of xn _i , yn _i and zn _i .

Step 4b: find the geometric center of the receptor

xn _center = (xn _max -xn _min ) / 2,

yn _center = (yn _max -yn _min ) / 2, zn _center = (zn _max -zn _min ) / 2

Step 4c: Compute the center coordinates:

(xnc _j , ync _j , znc _j ) = (xn _i -xn _center , yn _i -yn _center , zn _j -zn _max )

Step 5: Assemble the collision surface of the target using atomic radius and conversion coordinates (xnc _i , ync _i , znc _i , radius _i ): g (x _g , y _g ) = z _g

Step 5a: Create a grid with an interval equal to the diameter of the receptor subunit (= 1).

Grid coordinates:

x _g ∈ (Int (xn _min -xn _center ), Int (xn _min -xn _center )

+ 1 ... 0, .. Int (xn _max -xn _center )-1, Int (xn _max -xn _center )｝

y _q ∈ {Int (yn _min -yn _center ), Int (yn _min -yn _center )

+ 1 .... 0, .. Int (yn _max -yn _center )-1, Int (yn _max -yn _center )｝

Set the initial value of g (x _g , y _g ) to 0 at all points on the grid.

Step 5b: For each atom (i), adjust the g (x _g , y _g ) (height) value of each lattice point (x _g , y _g ) according to the following rules:

i = 1 to the number of atoms in the target

If (xnc _i -x _p ) ² + (ync _i + y _p ) ² <radius _i ² ,

g (x _g , y _g ) = minimum (g (x _g , y _g ), znc _i -radius _i ); or

If (xnc _i -x _p ) ² + (ync _i + y _p ) ² <(radius _i + .5) ² ,

g (x _g , y _g ) = minimum (g (x _g , y _g ), znc _i -radius _i / 2)); or

g (x _g , y _g ) = minimum (g (x _g , y _g ), 0) then i

Step 6: Concentrate the receptor coordinates at the origin (0, 0).

Step 6a: find the maximum and minimum values of xr _j , yr _j and zr _j .

Step 6b find the geometric center of the receptor:

xr _center = (xr _max -xr _min ) / 2,

yr _center = (yr _max -yr _min ) / 2,

zr _center = (zr _max -zr _min ) / 2

Step 6c: Compute the center receptor coordinates:

(xc _j , yc _j , zc _j ) = (xr _j -x _center , yr _j -y _center , zr _j -z _minimum )

Step 7: Assemble the collision surface s (x _s , y _s ) = z _s of the receptor using the central receptor coordinates according to the following rules:

Set all initial values of s (xc _j , yc _j ) to zero.

j = 1 to the number of subunits in the receptor

If zc _j > s (xc _j , yc _j ) then s (xc _j , yc _j ) = zc _j

Then j

Step 8: Find the minimum separation between the collision surface of the receptor and the collision surface of the target. Compute the difference matrix d (x _g , y _g ) according to the following rules:

In all cases

x _g ∈ (Int (xn _min -xn _center ), Int (xn _min -xn _center )

+ 1 .... 0, .. Int (xn _max -xn _center )-1, Int (xn _max -xn _center )｝ and

y _g ∈ {Int (yn _min -yn _center ), Int (yn _min -yn _center )

+ 1 .... 0, .. Int (yn max-yn _center )-1, Int (yn max-yn _center )｝

d (x _g , y _g ) = (h (x _g , y _g )-(zn _min + zn _max ) + (s (x _g , y _g )

Calculate + (zr _min + zr _max ).

Find the minimum of d (x _g , y _g ) = d _minimum for all x _g and y _g, where d _min is the minimum separation distance.

Step 9: Invert target and receptor coordinates for collisional atom arrays.

For receptors:

(x receptor _j, y _j receptors, receptor z _j) = _(j xc, yc _{j, j} + zc zr _least -zr _max)

For targets:

(x target _i , y target _i , z target _i )

= (xnc _i , ync _i , znc _i -zn _min + zn _{max-d min} )

Step 10: (x target _i , y target _i , z target _i ) and (x receptor _j , y receptor _j , z receptor _j ) are used for affinity calculations.

Repeat steps 2 through 9 for each target atom array to be tested.

Module 4: Affinity Calculation

Step 1: Enter the collision coordinates (x target _i , y target _i , z target _i ) and (x receptor _j , y receptor _j , z receptor _j ) of the target and receptor.

(i = number of atoms in the target, j = number of subunits in the receptor)

Step 2: Enter the dipole moment for the target: dipole (i)

Enter the charge value for the receptor: charge (j)

Step 3: Enter the limits for proximity calculations: Limits

Step 4: Calculate the bipolar affinity value.

Step 4a: For each charged subunit (charge (j) ≠ 0), e (i, j) = dipole (i) / ((x target _i -x receptor _j ) ² + (y target _i -y _Calculate receptor _j ) ² + (z target _i − z receptor _j ) ² ) ^1.5 .

Step 4b: Compute the sum of e (i, j) for all combinations of i and j with charge (j) ≠ 0.

Dipole = Σe (i, j)

Step 5: Compute proximity value (this step can be replaced by calculation based on polarization)

Step 5a: | Dipole (j) | For each target atom ≤ 0.75

1 (i, j) = ((x target _i -x receptor _j ) ² + (y target _i -y receptor _j ) ²

Calculate + (z target _i -z receptor _j ) ² ) ^0.5 .

1 (i, j) <Proximity (i, j) = 1 if limit

Step 5b: | Dipole (j) | Calculate the sum of proximity (i, j) for all combinations of i and j with ≤ 0.75: Proximity = Σ proximity (i, j)

Step 6: Compute the affinity value for the target substrate combination = affinity

Affinity = (Proximity / j) ((Proximity / 10000) + Bipolar)

Module 5: Joint Diagram Calculation

Step 1: Enter a known target efficiency or affinity value (y _k ) (k = number of test targets)

Step 2: Enter the collision coordinates (x target _i , y target _i , z target _i ) and (x receptor _j , y receptor _j , z receptor _j ) of the target and receptor (i _k = number of atoms in target k , j = number of subunits in the receptor)

Step 3: Enter the number of target orientations tested (= m).

Step 4: Obtain affinity values (= affinity _{k, m} ) for each target and target orientation using Module 5.

Step 5: Determine the maximum affinity (MA _k ) and the total affinity (SA _k ) for each target.

Step 6: Correlation coefficient for known target efficiency or affinity value (y _k ) versus maximum affinity (MA _k ) r _MA ² and known target efficiency or affinity value (y _k ) vs. total affinity (SA) Calculate the correlation coefficient r _SA ² for _k ).

Step 7: Calculate the joint coefficient F

F = (r _MA ² xr _SA ² ) ^0.5

Alternative steps

Step 6: correlation coefficient for known target efficiency or affinity value (y _k ) versus maximum affinity (MA _k ) r _MA ² and known target efficiency or affinity value (yk) vs. total affinity (SAk) )-Calculate the correlation coefficient r _SA-MA ² for maximum affinity.

Step 7: calculate the joint coefficient F.

F = (r _MA ² x (1-r _SA-MA ² )) ^0.5

Module 6: Genotyping with Minimal Affinity

Step 1: Determine the minimum junction limit

Step 2: form a random genotype (Module 1).

Step 3: Translate the genotypes to assemble the phenotypes (Module 2).

Step 4: Test the affinity of the phenotype for the target (modules 3, 4, 5, 6).

Step 5: If the degree of concatenation of the phenotype exceeds the limit of concatenation, the generation of encryption is stopped and the password is sent to step 2. If not, repeat steps 1-5.

Module 7: Multiple Mutations

Step 1: Enter the first password (basic password) (from the detailed description "Step 1" of the invention)

Step 2: Determine the number of mutations per code (= q) (in our example 2.5-5% of the genotypes are mutated).

Step 3: Enter the size of the population (= p).

Step 4: Randomly select a location within the genotype.

Step 5: The cipher code of the position is replaced with another randomly selected code.

Step 6: Repeat steps 4 and 5 until q is reached.

Step 7: Repeat steps 4-6 to generate a total of p new passwords.

Step 8: Apply Modules 1-6 to test the conjugation of the mutant population. The group of subjects with the highest selectivity is selected and used in the detailed description "Step 3" of the invention.

Module 8: Recombination

Step 1: Determine the population size (= P).

Step 2: Randomly select two ciphers from the population created by "Step 2".

Step 3: Randomly select a location within the genotype.

Step 4: Generate a random number for the number of codes to be exchanged.

Step 5: Start at the selected location and perform code exchange between ciphers.

Step 6: Repeat steps 2-5 until P new genotypes are generated.

Step 7: Apply Modules 2-6 to test the conjugation of the mutant population. The group of subpopulations with the highest selectivity is selected and used for subsequent recombination or maturation of "Step 4".

Module 9: Maturation

Step 1: Enter the parent password derived from "Step 3".

Step 2: Determine the number of iterations.

Step 3: Randomly select a location in the parent genotype.

Step 4: Replace the cipher code of the position with another randomly selected code.

Step 5: Use Modules 2-6 to test the selectivity of the parent code (F _p ) and the selectivity of the mutation product (F _M ).

Step 6: Replace the parent genotype with the mutant product if F _M ≥ F _p .

Step 7: Repeat steps 3-6 for the required number of repetitions.

Module 10: Generating a password to create a ringed carbon backbone (ring with 6 members, entry point = output point)

Step 1: Determine the length of the password.

Determine v1, v2, v3, ... vn (frequency of substituents).

The probe _ ring (frequency of ring coding sequence) is determined.

(0 ≤ probe _ ring ≤ 1)

Step 2: Initialize the first_password = "".

Initialize the second_password = "".

Initialize the third_password = "".

Step 3: Generate a sign string.

Repeat step 4 until the encryption length is obtained.

Step 4a: Probe _ Ring> Random (0 ≤ Random ≤ 1)

Code assignments for rings (sailboat rules).

Assign a new_sign_1 to the following randomly selected numbers: {'431413', '314134', '141343', '132132', '321321',

'213213', '123123', '231231', '312312', '421412',

'214124', '141242', '324234', '242343', '423432'}

Sign assignment for substituents.

Using frequency v1, v2, v3, ... vn, new_signal_2 is assigned to six members randomly selected from {c1, c2, c3, ..., cn} (c1..cn respectively) Code to specify other functional groups).

Specifies the sign for a substituent atom. The new_code_3 is assigned to six members randomly selected from {'1', '2', '3', '4'}.

or

Step 4b: Specify a single (non-ambient) code for the first password.

The new_sign_1 is assigned to a member randomly selected from {'1', '2', '3', '4'}.

Specifies the sign for the substituent.

Using the frequencies v1, v2, ... vn, the new _sign_2 is assigned to the member randomly selected from {c1, c2, ..., cn}.

Specifies the sign for the substituent valency.

The new_signal_3 is assigned to a member randomly selected from {'1', '2', '3', '4'}.

Step 4c: Concatenate the new codes to the cipher string.

First_password = first_password & new_code_1

Second_password = second_password & new_code_2

Third_password = third_password & new_code_3

Module 11: Multiple Point Mutations

Step 1: Enter the first password.

Step 2: Determine the number of mutations per code (= q) (in our example 2.5-5% of the genotypes are mutated).

Step 3: Enter the size of the population (= p).

Step 4: Randomly select a location within the genotype.

Step 5: Replace the cipher code of the position in each cipher vector with another randomly selected code.

Step 6: Repeat steps 4 and 5 until q is reached.

Step 7: Repeat steps 4-6 to generate a total of p new passwords.

Step 8: Test the conjugation of each member in the population of mutants. Subgroups of the highest selectivity are selected and used for recombination or additional multiple mutations.

Module 12: Sequence Mutation

Step 1: Set P _deletion , P _inversion , P _insertion and P _overlap as the threshold for mutagenesis (0 ≤ P _X ≤ 1).

Step 2: Generate one random position (= x) in the cipher (0 ≦ p ≦ password length).

Step 3: Generate a random length (= L) of the sequence (0 ≦ L ≦ coding length minus x).

Step 4: Copy the sequence from the code starting at x up to a total of L codes.

Step 5: When 0 ≦ P _inversion ≦ random number ≦ 1, reverse the order of the codes in the string.

Step 6: If 0 ≦ P _overlap ≦ random number ≦ 1, copy the sequence and link the copy to the sequence.

Step 7: When 0 ≤ P _deleting ≤ random number ≤ 1, L codes are deleted from the starting code at position x. or

The coded sequence is replaced with the sequence generated in steps 5 and 6.

Eighth Step: When 0 ≤ P _Insertion ≤ Random Number ≤ 1, randomly select a position (= y) in the cipher (0 ≤ y ≤ cipher length).

The sequence generated by steps 5 and 6 is inserted at position y.

Module 13: Recombination

Step 1: Determine the population size (= P).

Step 2: Randomly select two codes from the population created by multiple mutations.

Step 3: Randomize a location within the genotype

Step 4: Generate a random number for the number of codes to be exchanged.

Step 5: Starting from the selected position, perform code exchange between each of three encryption vectors.

Step 6: Repeat steps 2-5 until P new genotypes are generated.

Step 7: Test the conjugation of each ligand in the resulting mutant population. The group of subpopulations with the highest selectivity is selected for subsequent use in recombinant series or maturation.

Module 14: Maturation

Step 1: Enter the parent code derived from recombination.

Step 2: Determine the number of iterations.

Step 3: Randomly select a location in the maternal genotype.

Step 4: Replace the cipher code of the positions in each cipher vector with other randomly selected codes.

Step 5: Use Modules 4 and 5 to test the maternal password's conjugation (F _p ) and the mutation product's conjugation (F _M ).

Step 6: Replace the parent genotype with the mutant product if F _M ≥ F _p .

Step 7: Repeat steps 3-6 for the required number of repetitions.

Claims (53)

(A) preparing a physical model of a simulated receptor phenotype encoded by a linear code sequence and providing a set of target molecules that share one or more quantifiable functional features; (B) for each target molecule (i) calculating the affinity between the target molecule and the receptor for each of the various orientations using the effective affinity calculation; (ii) adding up the calculated affinity to calculate the total affinity; (iii) confirming maximum affinity; (C) using the calculated total affinity and maximum affinity; (i) calculating the maximum affinity correlation coefficient between the maximum affinity and the quantifiable functional feature; (ii) calculating a total affinity correlation coefficient between the total affinity and the quantifiable functional feature; (D) calculating a joint coefficient using the maximum correlation coefficient and the total sum correlation coefficient; (E) changing the receptor structure and repeating steps (b) to (d) until a receptor population having a preselected conjugation coefficient is obtained; (F) providing a physical model of a chemical structure encoded by a linear code sequence on a molecule, calculating the affinity between the chemical structure and each receptor for each of the various orientations using the effective affinity calculation, and calculating the calculated affinity. Calculating an affinity junction score using; (G) changing the chemical structure to generate chemical structural variants and repeating step (f); (H) computer-based design of a chemical structure having a pre-selected functional feature comprising the step of selecting a chemical structural variant having an affinity score close to a preselected affinity score and adding a further change. The method of claim 1, wherein preparing a simulated receptor genotype comprises generating a receptor linear code sequence that encodes space occupancy and charge, and preparing a physical model of the chemical structure comprises: encoding the space occupancy and charge. A computer-based design of a chemical structure having a preselected functional feature comprising generating a molecular linear sequence. 3. The method according to claim 2, wherein the effective affinity calculation is based on a proximity scale at which the ratio of the uncharged sites on the simulated receptor sufficiently close to the nonpolar sites of the molecular structure to generate an effective London dispersion force, and the charged sites of the simulated receptors. And a computer-based design of a chemical structure having a preselected functional feature comprising two measures of the total intensity of charge-dipole electrostatic forces generated between the two poles present in the molecular structure. The method of claim 2, wherein the affinity scoring score calculating step comprises calculating the sum and maximum affinity between the molecular structure and each receptor, wherein the affinity score is calculated as follows: Σ {| Calculated Maximum Affinity-Target Maximum Affinity | / Target affinity}, Computer-based design of a chemical structure having a preselected functional feature, wherein the preselected jointness score is substantially zero. The method of claim 2, wherein the affinity scoring score calculating step includes calculation of the molecular structure and the sum and maximum affinity between each receptor, wherein the affinity score is calculated as follows: Σ {(| Calculated maximum affinity-target maximum affinity | / 2 x target maximum affinity) + | Calculated total affinity-target total affinity | / 2 x target affinity}, Computer-based design of a chemical structure having a preselected functional feature, wherein the preselected jointness score is substantially zero. The joint affinity correlation coefficient of claim 2, wherein the total affinity correlation coefficient is r _SA ² , the maximum affinity correlation coefficient is r _MA ² , and the bonding coefficient is F = (r _MA ² xr _SA ² ) ^0.5 . A computer-based design of a chemical structure having a preselected functional feature, wherein the coefficient is substantially one. The method of claim 2, wherein the total affinity correlation coefficient is r _SA-MA ² , the maximum affinity correlation coefficient is r _MA ² , and the bonding coefficient is F = (r _MA ² x (1-r _SA ² )) ^0.5 And wherein said preselected bond coefficient is substantially one. 3. The method of claim 2, wherein the molecular linear sequence comprises a plurality of sequence code triplets, wherein the first sign of the triplet is randomly selected from the first set of codes defining the location and identity of the occupants in the molecular backbone of the molecular structure. The second sign of the triplet is randomly selected from the second set of codes that define the identity of the substituents attached to the occupant, and the third sign of the triplet defines the position of the atomic substituent defined by the first sign of the triplet. A computer-based design of a chemical structure having a preselected functional feature characterized in that it is randomly selected from a third set of codes. The method according to claim 8, wherein the molecular linear coding sequence is decoded using an effective molecular assembly operation method of sequentially translating each triplet from the molecular linear sequence and then filling an unfilled position on the molecular skeleton with a hydrogen atom. Computer-based design of chemical structures with preselected functional features. 10. The method of claim 9, wherein the step of mutating the molecular structure comprises at least one of the following steps: Iii) mutating said molecular genotype by randomly exchanging at least one of the first, second and third symbols of at least one triplet from the linked code set; Ii) deleting the pre-plate from the molecular genotype; Iii) an overlapping step of overlapping triplets in the molecular genotype; Iii) an inversion step of inverting the sequence order of one or more triplets in the molecular genotype; Iii) an insertion step of inserting the triplet in the molecular genotype to another position in the molecular genotype. The chemistry of claim 10, wherein the step of mutating the molecular genotype comprises recombining randomly selected pairs of the selected mutant molecular genome such that corresponding symbols in the molecular linear sequence are interchanged. How to design a structure based on computer. 3. The computer-based method of claim 2, wherein each code in the receptor linear code sequence defines a charged position without spatial redirection instructions or redirection. 13. The preselected function of claim 12, wherein said receptor phenotype comprises at least one linear polymer having a plurality of subunits, wherein one of said subunits is the first subunit in said at least one linear polymer. Computer-based design of a characteristic chemical structure. The method of claim 13, wherein the linear receptor sequence is decoded using an effective receptor assembly operation, wherein the redirection instruction for each subsequent subunit following the first subunit is applied with respect to the initial position of the first subunit. Computer-based design of a chemical structure having a preselected functional feature. 15. The position and direction of claim 14, wherein the codes defining the space direction change instruction encode no turn, right turn, left turn, up turn, down turn, and the codes defining charge positions are positively charged positions and directions without a turn. A computer-based design of a chemical structure having a preselected functional feature characterized by encrypting a negatively charged position without conversion. The computer-based design of a chemical structure having a preselected functional feature in accordance with claim 14, wherein said subunits are in the form of a substantially sphere having a van der Waals radius substantially equal to the van der Waals radius of hydrogen. Way. 16. The method of claim 15, wherein said step of mutating said receptor genotype comprises a chemical structure having a preselected functional feature comprising one or more of the following steps: Iii) deleting the code from the receptor genotype; Ii) an overlapping step of overlapping the code in the receptor genotype; Iii) an inversion step of inverting the sequence sequence of one or more symbols in the receptor genotype; Iii) an insertion step of inserting the code in the receptor genotype at another location in the genotype. 18. The chemistry of claim 17, wherein said mutating said receptor genotype comprises recombining randomly selected pairs of said selected mutant receptor genotype such that corresponding symbols in said receptor linear sequence are interchanged. How to design a structure based on computer. (A) preparing a simulated receptor genotype by generating receptor linear coding sequences encoding space occupancy and charge; (B) deciphering the genotype to generate a receptor phenotype, providing one or more target molecules with the selected functional feature, and using affinity calculations to calculate the affinity between each target molecule and receptor for various orientations, Calculate the total and maximum affinity between each target molecule and the receptor, and calculate the total affinity correlation coefficient for the functional feature versus the total affinity of the target molecule and the maximum affinity correlation coefficient for the functional feature versus the maximum affinity. Calculating a joint coefficient based on the sum and the maximum affinity correlation coefficient; (C) mutating the receptor genotype and repeating step (b), selecting a receptor that exhibits an improved conjugation coefficient and mutating it until a group of receptors having a preselected conjugation coefficient is obtained; (D) Calculate the affinity between each receptor for the different orientations and the selected chemical structure for the various orientations using the effective affinity calculation, calculate the total and maximum affinity between the compound and each receptor, and then determine at least one of the total and maximum affinity. Calculating affinity junction scores by comparing the total and maximal affinities between one or more targets and the receptor population, and using the comparisons as a guide for the level of functional activity of the chemical structure for the one or more target molecules Method for screening chemical structures with functional features. The proximity affinity calculation of claim 19, wherein the effective affinity calculation is a proximity measure from which the proportion of uncharged sites on the simulated receptor sufficiently close to the nonpolar sites of the molecular structure to produce an effective London dispersion force and charged sites of the simulated receptor. And two measures of the total intensity of the charge-dipole electrostatic force generated between the dipoles present in the molecular structure. 21. The method of claim 20, wherein said joint score is calculated as follows: Σ {| Calculated Maximum Affinity-Target Maximum Affinity | Target affinity} 21. The method of claim 20, wherein said joint score is calculated as follows: Σ {(| Calculated maximum affinity | target maximum affinity | / 2 x target maximum affinity) + (| Calculated total affinity-target total affinity | / 2 x target total affinity)} 21. The method of claim 20, wherein the total affinity correlation coefficient is r _SA ² and the maximum affinity correlation coefficient is r _MA ² , the bonding coefficient is F = (r _MA ² xr _SA ² ) ^0.5 , the preselected bonding degree A method for screening chemical structures with preselected functional features, wherein the coefficient is substantially one. The method of claim 20, wherein the total affinity correlation coefficient is r _SA ² and the maximum affinity correlation coefficient is r _MA ² , and the junction coefficient is F = (r _MA ² x (1-r _SA-MA ² )) ^0.5 And wherein the preselected bonding coefficient is substantially one. 21. The method according to claim 20, wherein each code in the receptor linear code sequence defines a charged position without spatial redirection instruction or redirection. The preselected function of claim 25, wherein the receptor phenotype comprises at least one linear polymer having a plurality of subunits, wherein one of the subunits is the first subunit in the at least one linear polymer. Method for screening chemical structures with features. 27. The method of claim 26, wherein the linear receptor sequence is decoded using an effective receptor assembly algorithm, wherein redirection instructions for each subsequent subunit following the first subunit are applied with respect to the initial position of the first subunit. Method for screening a chemical structure having a pre-selected functional feature characterized in that it is. 29. The position and direction of claim 27, wherein the codes defining a space direction change instruction encode no turn, right turn, left turn, up turn, and down turn, and codes defining charge positions are positively charged positions and directions without a turn. A method for screening chemical structures with preselected functional features, characterized by encoding a negatively charged position without conversion. 29. The method of claim 28, wherein said subunits are in the form of substantially spherical spheres having van der Waals radii substantially equal to the van der Waals radii of hydrogen. 28. The method of claim 27, wherein said step of mutating said receptor genotype comprises one or more of the following steps: Iii) deleting the code from the receptor genotype; Ii) an overlapping step of overlapping the code in the receptor genotype; Iii) an inversion step of inverting the sequence sequence of one or more symbols in the receptor genotype; Iii) an insertion step of inserting the code in the receptor genotype at another location in the genotype. 31. The chemistry of claim 30, wherein said mutating said receptor genotype comprises recombining randomly selected pairs of said selected mutant receptor genotype such that corresponding symbols in said receptor linear sequence are interchanged. Method of screening structure. (A) preparing a simulated receptor genotype by generating receptor linear coding sequences encoding space occupancy and charge; (B) deciphering the genotype to generate a phenotype of the receptor, providing a series of target molecules that share similar functional features, and using affinity calculations to calculate the affinity between each target molecule and receptor for various orientations; Calculate the total affinity and maximum affinity between each target molecule and the receptor, calculate the total affinity correlation coefficient for the functional trait versus the total affinity for each target molecule, and calculate the functional trait versus the affinity for each target molecule. Calculating a maximum affinity correlation coefficient for the maximum affinity and calculating a joint coefficient for each target molecule based on the sum and the maximum affinity correlation coefficient; (C) mutagenesis of the genotype and repeating step (b), and then selecting a receptor that exhibits an improved conjugation coefficient and mutating until a population of receptors having a preselected conjugation coefficient is obtained. A method of designing a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having a. 33. The method of claim 32, wherein the design of a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having a similar functional feature that defines a position where each code in the receptor linear code sequence is charged without spatial redirection or redirection. . 34. The method of claim 33, wherein the receptor phenotype comprises a plurality of linear polymers having a plurality of subunits, each linear polymer encoded by a corresponding linear code sequence, wherein one of the subunits is one or more linear A method of designing a simulated receptor similar to a bioreceptor with selective affinity for a compound having a similar functional feature, characterized in that it is the first subunit in the polymer. 35. The method of claim 34, wherein the linear receptor sequence is decoded using an effective receptor assembly algorithm, wherein a redirection instruction for each subsequent subunit following the first subunit is applied with respect to the initial position of the first subunit. A method of designing a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having similar functional features, characterized in that 36. The position and direction of claim 35, wherein the codes defining a space direction change instruction encode no turn, right turn, left turn, up turn, down turn, and the codes defining charge positions are positively charged positions and directions without turning. A method of designing a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having similar functional features characterized by encoding a negatively charged position without conversion. 37. The bioreceptor of claim 36, wherein said subunits are in the form of substantially spherical spheres having van der Waals radius substantially equal to the van der Waals radius of hydrogen. How to design similar simulated receptors. 36. The method of claim 35, wherein said step of mutating said receptor genotypes comprises a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having a similar functional feature comprising one or more of the following steps: Iii) deleting the code from the receptor genotype; Ii) an overlapping step of overlapping the code in the receptor genotype; Iii) an inversion step of reversing the sequence of one or more codes in the receptor genotype; Iii) an insertion step of inserting the code in the receptor genotype at another location in the genotype. See translation of PCT corrections (paragraph 39 is missing at the time of international application). 40. The compound of claim 39, wherein the step of mutating the receptor genotype comprises recombining randomly selected pairs of the selected mutant receptor genotype such that corresponding symbols in the receptor linear sequence are interchanged. A method of designing a simulated receptor similar to a bioreceptor that exhibits selective affinity for humans. 34. The method of claim 33, wherein the effective affinity calculation is a proximity measure from which the proportion of uncharged sites on the simulated receptor sufficiently close to the nonpolar sites of the molecular structure to produce an effective London dispersion force and charged sites of the simulated receptor. And a bioreceptor-like simulation receptor that exhibits selective affinity for a compound having similar functional features, including two measures of the total intensity of charge-dipole electrostatic forces generated between the dipoles present in the molecular structure. 42. The method according to claim 41, wherein the total affinity correlation coefficient is r _SA ² and the maximum affinity correlation coefficient is r _MA ² , and the bonding coefficient is F = (r _MA ² xr _SA ² ) ^0.5 , A method of designing a simulated receptor similar to a bioreceptor with selective affinity for a compound having a similar functional feature, wherein the coefficient is substantially 1. 42. The method according to claim 41, wherein the total affinity correlation coefficient is r _SA - _MA ² and the maximum affinity correlation coefficient is r _MA ² , the bonding coefficient is F = (r _MA ² x (1-r _SA-MA ² ) ^0.5 ), wherein said preselected conjugation coefficient is substantially 1, and design a simulated receptor similar to a bioreceptor that exhibits selective affinity for a compound having a similar functional feature. (A) providing a series of target molecules that share a physical model of the receptor and one or more quantifiable functional features; (B) for each target molecule (i) calculating the affinity between the target molecule and the receptor for each of the various orientations using the effective affinity calculation; (ii) adding up the calculated affinity to calculate the total affinity; (iii) confirming maximum affinity; (C) using the calculated total affinity and maximum affinity; (i) calculating the maximum affinity correlation coefficient between the maximum affinity and the quantifiable functional feature; (ii) calculating a total affinity correlation coefficient between the total affinity and the quantifiable functional feature; (D) calculating a joint coefficient using the maximum correlation coefficient and the total sum correlation coefficient; (E) changing the receptor structure and repeating steps (b) to (d) until a receptor population having a preselected conjugation coefficient is obtained; (F) provide a physical model of the chemical structure, calculate the affinity between the chemical structure and each receptor for each of the various orientations using the effective affinity acid, and calculate the affinity junction score using the calculated affinity Making a step; (G) changing the chemical structure to generate chemical structural variants and repeating step (f); (H) Computer-based design of a chemical structure having a preselected functional property comprising the step of selecting a chemical structural variant close to the pre-selected affinity score and applying further changes. 45. The method of claim 44, wherein providing a physical model of the receptor comprises generating a receptor linear code sequence encoding space occupancy and charge, and preparing a physical model of the chemical structure comprises a linear encoding space occupancy and charge. Computer-based design of chemical structures with preselected functional characteristics including generation of code sequences. 46. The method of claim 45, wherein the linear coding sequence for the chemical structure comprises a plurality of sequence code triplets, wherein the first sign of the triplet is randomized from a first set of codes that defines the location and identity of skeletal occupants of the chemical structure. Wherein the second sign of the triplet is randomly selected from a second set of codes that define the identity of the substituents attached to the occupant, and the third sign of the triplet identifies the position of the atomic substituent defined by the first sign of the triplet. A computer-based design of a chemical structure having preselected functional characteristics, characterized in that it is randomly selected from a third set of codes. 47. The method of claim 46, wherein the linear structural sequence is decoded using an effective molecular assembly operation method of sequentially translating each triplet from the linear sequence and then filling an unfilled position on the molecular skeleton with hydrogen atoms. Computer-based design of chemical structures with preselected functional properties. And a linear code sequence encoding the space occupancy and charge for each atom in the chemical structure. 49. The set of claim 48, wherein said linear code sequence for said chemical structure comprises a plurality of sequence code triplets, wherein the first code of said triplet defines a location and identity of occupants in the molecular backbone of said chemical structure. Wherein the second sign of the triplet is selected from the second set of codes that define the identity of the substituents attached to the occupant, and the third sign of the triplet identifies the position of the atomic substituent defined by the first sign of the triplet. A method of encrypting a chemical structure comprising atomic elements, characterized in that it is selected from a third set of codes. 48. The atom of claim 46, wherein the linear encoding sequence is decoded by an effective molecular assembly operation of sequentially translating each triplet from the linear sequence and then filling an unfilled position on the molecular skeleton with hydrogen atoms. A method of encrypting a chemical structure comprising components. 51. The method of claim 50, comprising storing the linear code sequence in computer-accessible storage means. 20. The method of claim 19, wherein said functional feature is biologically toxic. 20. The method according to claim 19, wherein said functional feature is catalytically active.