Method for predicting protein structure on basis of interactive attraction relationships between low-entropy hydration layers of residue side chains

Abstract

Provided is a method for predicting a protein structure on the basis of the interactive attraction relationships between low-entropy hydration layers of residue side chains. The method solves the problem of poor prediction performance in existing methods for predicting a protein folding structure. The method realizes the prediction of a protein structure using a water entropic force, wherein 'prediction' refers to the prediction of the secondary structure and tertiary structure of a protein on the basis of the primary structure of the protein; 'water entropic force' refers to the interactive attraction force between low-entropy hydration layers of amino acid residue side chains; 'interactive attraction force' refers to a force in which the entropy increase of low-entropy water molecules in the low-entropy hydration layers drives lateral association between the amino acid residue side chains; and 'lateral association' refers to a state where two residue side chains are nearly parallel to each other. The method is used for predicting the structure of a protein.

Description

Protein structure prediction method based on the attraction between low-entropy hydration layers of residue side chains Technical Field

[0001] This invention belongs to the field of structural biology, specifically relating to a method for predicting protein structure. Background Technology

[0002] Protein structure prediction is crucial for biological research. The function of a protein largely depends on its structure; therefore, obtaining the three-dimensional structure of proteins is essential for understanding their function, interactions, and pathological mechanisms. Currently, protein structure prediction mainly employs methods such as deep learning, comparative models, and molecular dynamics simulations. Deep learning, also known as artificial intelligence and machine learning, has played a key role in protein structure prediction, but its advancements can only be attributed to progress in the field of artificial intelligence. Artificial intelligence methods currently lack a direct method for deriving protein folding rules and predicting protein structures based on explicit physical and chemical mechanisms. Furthermore, using molecular dynamics simulations to simulate the protein folding process on computers is considered a physical approach to solving the protein folding problem. This approach requires significant high-performance computing resources to simulate the motion and interactions of polypeptide molecules over a certain time period to infer the most stable protein structure; however, current protein folding dynamics simulations still cannot accurately predict the molecular structure of proteins.

[0003] The prevailing view in academia today is that the physical driving forces of protein folding include hydrophobic interactions, hydrogen bonds, electrostatic forces, van der Waals forces, and ionic bonds. However, predictions based on this theory are not very effective. Precise protein folding only occurs in an aqueous environment; in any non-aqueous solvent, the vast majority of proteins cannot fold correctly. It must be emphasized that water molecules have extremely strong polarity and a very high dielectric constant, exhibiting a significant electrostatic shielding effect. Therefore, it is evident that the prediction methods based on hydrogen bonds, electrostatic forces, van der Waals forces, and ionic bonds cannot be adequately explained theoretically. This explains the poor prediction results and confirms that these methods have inherent theoretical limitations in predicting protein structure. Currently, there is no, and urgently needs to be developed, a method for predicting protein structure solely based on the hydrophobic interactions between amino acid residue side chains. Summary of the Invention

[0004] This invention aims to address the problem of poor prediction performance in current methods for predicting protein folding structures.

[0005] A protein structure prediction method based on the mutual attraction between low-entropy hydration layers of amino residue side chains is proposed. This method utilizes hydrotropic forces to predict protein structure. The prediction is based on the protein's primary structure to predict its secondary and tertiary structures. The hydrotropic forces refer to the attractive interactions between the low-entropy hydration layers of amino residue side chains. These attractive interactions refer to the lateral adhesion between amino acid residue side chains driven by the entropy increase of low-entropy water molecules in the low-entropy hydration layers. Lateral adhesion refers to two residue side chains being in a nearly parallel state.

[0006] Furthermore, the process of using water entropy force to predict protein structure includes:

[0007] I. Pre-forecast preparations:

[0008] First, entropy values ​​are assigned to the side chains of amino acid residues in the primary structure of the protein. The entropy value is a relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the side chains of amino acid residues.

[0009] Then, the amino acid residue side chains in the primary structure of the protein after entropy assignment are divided into 5 residue forms:

[0010] Residues with potential for micro-entropy increase: serine (S), threonine (T), aspartic acid (D), and asparagine (N);

[0011] Hydrophilic residues: histidine H, arginine R, lysine K, glutamic acid E, glutamine Q, aspartic acid D, asparagine N, tyrosine Y, tryptophan W, serine S;

[0012] Turning residues: glycine (G), proline (P);

[0013] Hydrophobic residues: Isoleucine I, Valine V, Leucine L, Phenylalanine F, Tyrosine Y, Tryptophan W, Alanine A, Methionine M, Cysteine ​​C, Histidine H;

[0014] Residues with high entropy increase potential: Isoleucine I, Valine V, Leucine L, Lysine K, Arginine R, Phenylalanine F, Tyrosine Y, Tryptophan W, Alanine A, Methionine M, Cysteine ​​C, Glutamic Acid E, Glutamine Q, Histidine H;

[0015] Finally, the preferred pairing between amino acid residue side chains was determined.

[0016] The preferred pairing between amino acid residue side chains is the lateral adhesion between multiple pairs of amino acid residue side chains, which leads to a sufficient entropy increase in the low-entropy hydration layer of multiple pairs of amino acid residue side chains. The pairing relationship between multiple pairs of amino acid residues is the preferred pairing between amino acid residue side chains.

[0017] The preferred pairing between amino acid residue side chains includes preferred pairing between amino acid residue side chains with a 1-4 / 5 sequence position relationship, preferred pairing between amino acid residue side chains with a 1-3 sequence position relationship, and preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship.

[0018] The 1-4 / 5 sequence position relationship refers to the relationship between the sequence positions of one side chain of an amino acid residue and another side chain of an amino acid residue in the primary structure of a protein, which are separated by 2 or 3 amino acid residues.

[0019] The 1-3 sequence position relationship refers to the relationship between the sequence positions of one amino acid residue side chain and another amino acid residue side chain in the primary structure of a protein, which are separated by one amino acid residue.

[0020] The 1-2 sequence position relationship refers to the sequence position relationship between one residue side chain and another adjacent residue side chain in the primary structure of a protein.

[0021] II. Protein secondary structure prediction:

[0022] Based on the sequence positions of the residue side chains, amino acid residue classifications, and entropy values ​​obtained in step one, as well as the preferred pairings between residue side chains, the following process is used to predict the secondary structure of proteins:

[0023] S100. On a new line below the amino acid sequence labeled with entropy values, repeat the amino acid residues in the amino acid sequence that belong to high entropy increase potential residues as high entropy increase potential residue connected segments; then, on a new line, repeat the amino acid residues in the amino acid sequence that belong to low entropy residues and turn residues.

[0024] S200. List the unblocked high-entropy-increase-potential residue connectivity segments, and assume that the listed unblocked high-entropy-increase-potential residue connectivity segments are either α-helices or β-sheets in the protein secondary structure:

[0025] S201. Assuming that a high-entropy-increase-potential residue-connected fragment forms an α-helix, if a side chain of a residue in the high-entropy-increase-potential residue-connected fragment achieves preferred pairing between amino acid residue side chains, mark it as * below the amino acid residue of the high-entropy-increase-potential residue-connected fragment;

[0026] S202. Assuming that the high entropy increase potential residue connecting fragment forms a β-sheet, if the high entropy increase potential residue connecting fragment is a preferred pairing between the side chains of amino acid residues in a 1-3 sequence position relationship, then the amino acid residues of the high entropy increase potential residue connecting fragment are marked with & below them;

[0027] S203. Analyze the * and & symbols:

[0028] S2031. If none of the high-entropy-increase-potential residues in the primary structure of a protein are marked as &-&, and more than 80% of the amino acid residues in such high-entropy-increase-potential residues are marked as *, then it is predicted that such high-entropy-increase-potential residues form an α-helix.

[0029] S2032. If the high entropy increase potential residues in the primary structure of a protein are marked with &-&, it is predicted that the high entropy increase potential residues will form a β-sheet; if the number of * marks in the high entropy increase potential residues in the primary structure of a protein is less than 50% of the number of amino acid residues in the high entropy increase potential residues, it is predicted that the high entropy increase potential residues will form a β-sheet.

[0030] For S300, the amino acid side chains on both sides of the high entropy increase potential residue connected fragment predicted by S200 to form an α-helix are searched. If the amino acid residues in the amino acid side chains on both sides have a 1-4 / 5 sequence position relationship with the amino acid residues in the high entropy increase potential residue connected fragment, the amino acid residue side chains with the predicted 1-4 / 5 sequence position relationship are preferably paired to form an α-helix.

[0031] S400, based on the predictions of the S200 hypothesis, determines whether the protein secondary structure consists of α-helices or β-sheets:

[0032] The entropy increments of amino acid residues in the primary structure of proteins predicted as α-helices and β-sheets are calculated separately. The entropy increments of the predicted α-helices and β-sheets are compared. If the larger entropy increment is greater than 20% of the other entropy increment, the prediction result of the larger entropy increment is determined as the final prediction result. If the larger entropy increment is less than 20% of the other entropy increment, the prediction of the S200 hypothesis forming α-helices or β-sheets in the secondary structure of the protein is determined as a potential result.

[0033] Among them, the entropy increment statistical method of β-sheet is: the preferred pairing between amino acid residues with a 1-3 sequence position relationship in the primary structure of the protein, and the entropy value of the amino acid residues in the preferred pairing is used as the entropy increment.

[0034] Statistical method for entropy increment of α-helix: Preferred pairing of amino acid residues in the primary structure of a protein with a 1-4 / 5 sequence position relationship is used as the entropy increment.

[0035] S500. If there are two consecutive amino acid residues with low entropy increase potential in the connected segment of high entropy increase potential residues that are not predicted to be α-helices in the primary structure of a protein, then the two amino acid residues are predicted to form a turn and are marked as T.

[0036] A sequence of four or more consecutive hydrophilic amino acid residues in the primary structure of a protein is called a "continuous hydrophilic residue sequence". If, in a high-entropy potential sequence of residues not predicted as an α-helix in the primary structure of a protein, there are sequences of more than seven amino acid residues, and the number of residues in the low-entropy potential sequence and / or turn residues and / or amino acid residue A and / or continuous hydrophilic residue sequence accounts for more than 50% of the number of sequences of more than seven amino acid residues, then the sequence of more than seven amino acid residues is predicted to form a random coil, and this is considered a potential prediction result.

[0037] S600. If a protein's primary structure consists of a fragment composed of five or more consecutive hydrophilic amino acid residues, and this fragment has been predicted as a high-entropy-increase-potential residue-connected fragment, then the fragment composed of five or more consecutive hydrophilic amino acid residues is predicted to form a random coil, and this is taken as a potential prediction result. This completes the protein structure prediction method based on the mutual attraction relationship between the low-entropy hydration layers of residue side chains.

[0038] Furthermore, preferred pairings between amino acid residue side chains include:

[0039] FF, QQ, KK, EE, TT, RR, AA, SS, MM;

[0040] MA, QE, TS, RE, KR, EK, FY, WY, WE, RW, RY, KY, HV, DE, NE, DQ, NQ;

[0041] QR / K, I / V / L / FI / V / L / F, I / V / L / F / YI / V / L / F / Y, I / V / LY, I / V / LK, I / V / LR, I / V / LA, I / V / LW, I / V / L / F / YW, I / V / LM, CM / Y, MI / V / L, A / TD / N.

[0042] Further, in step S201, the preferred pairing between amino acid residue side chains refers to the preferred pairing between amino acid residue side chains with a 1-4 / 5 sequence position relationship; if there is a preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship in the high entropy increase potential residue connected fragment, or if there is a sequence position relationship of I / V / L 1-4 / 5G / A 1-4 / 5I / V / L / F in the high entropy increase potential residue connected fragment, then an asterisk (*) will be marked below that amino acid.

[0043] Furthermore, the preferred pairing of amino acid residue side chains with the 1-3 sequence position relationship described in S202 is I / V / L / FI / V / L / F; if, starting from the amino acid in I / V / L / FI / V / L / F, there is a preferred pairing of amino acid residue side chains with the 1-4 / 5 sequence position relationship in adjacent high entropy increase potential residue connected segments, then the amino acid in the preferred pairing of amino acid residue side chains with the 1-3 sequence position relationship is not marked as &.

[0044] Further, as described in S2032, if a high-entropy-increase-potential residue connected fragment in the primary structure of a protein is labeled with &-&, and more than 85% of the amino acid residues in the amino acid side chain region of the primary structure of the protein far from the labeled &-& high-entropy-increase-potential residue connected fragment are labeled with *, and there is a preferred pairing between amino acid residue side chains that have a 1-4 / 5 sequence position relationship with the unrelated residues labeled &-&, then it is predicted that the high-entropy-increase-potential residue connected fragment labeled &-& will form an α-helix; wherein, "far from" means separated by two consecutive micro-entropy residues or corner residues, or separated by at least three hydrophilic residues, or separated by one amino acid residue P; "unrelated residues" refers to other amino acid residues in the primary structure of the protein besides those labeled &-&; and "amino acid side chain region" refers to a region composed of amino acid side chains with more than 5 amino acid residues.

[0045] Furthermore, if the high entropy increase potential residue connected fragment described in S2032 is predicted to form a β-sheet, and the high entropy increase potential residue connected fragment has two consecutive low entropy residues aspartic acid D or asparagine N, then the two consecutive low entropy residues aspartic acid D or asparagine N are predicted to form a turn, as a prediction result.

[0046] Furthermore, in the entropy increment statistics of β-sheet described in S400, if the same amino acid residue in the primary structure of a protein achieves a preferred pairing between the side chains of two amino acid residues with a 1-3 sequence position relationship, then the entropy increment of that amino acid residue is accumulated twice.

[0047] In the entropy increment statistics of α-helices, if the same amino acid residue in the primary structure of a protein achieves preferred pairing between the side chains of multiple amino acid residues with 1-4 / 5 sequence positional relationships, then the entropy increment of that amino acid residue is accumulated twice.

[0048] Furthermore, in the entropy increment statistics of the α-helix described in S400, if there is a preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship in the primary structure of the protein, and the preferred pairing is I / V / L / MI / V / L / M, then two pairs of I / V / L / M are recorded as one entropy increment.

[0049] Furthermore, the amino acid residues in the two consecutive amino acid residues described in S600 are one or both of S, T, N, D, and G.

[0050] Furthermore, the relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the amino acid residue side chains in the protein primary structure is the length of the hydrophobic portion on the amino acid residue side chains in the protein primary structure.

[0051] Furthermore, the relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the amino acid residues in the primary structure of the protein is specifically defined as the number of carbon-carbon bonds on the amino acid residue chain in the primary structure of the protein as one entropy value.

[0052] Furthermore, the entropy values ​​of the 20 amino acids are as follows:

[0053] Serine S=0.5, Threonine T=0.5, Aspartic Acid D=0.5, Asparagine N=0.5, Glycine G=0, Histidine H=3, Arginine R=3, Glutamic Acid E=3, Lysine K=4, Tryptophan W=5, Isoleucine I=6, Valine V=6, Leucine L=6, Phenylalanine F=6, Proline P=0, Cysteine ​​C=4, Glutamine Q=3, Alanine A=1, Methionine M=4, Tyrosine Y=5.

[0054] Furthermore, the sufficient entropy increase mentioned in step one refers to the lateral adhesion of two amino acid residue side chains when they are similar in the primary structure of a protein, resulting in a sufficient entropy increase in the low-entropy hydration layer of these two side chains. The similarity of the amino acid residue side chains means that the smaller the difference in entropy values ​​between two hydrophilic residues, the more similar the two hydrophilic residues are considered; the smaller the difference in entropy values ​​between two hydrophobic residues, the more similar the two hydrophobic residues are considered.

[0055] Furthermore, in determining unblocked high-entropy-increase-potential residue connectivity segments, the process first involves identifying segments in the sequence composed of consecutive high-entropy-increase-potential residues, naming the original "high-entropy-increase-potential residue connectivity segments." Then, non-high-entropy-increase-potential residues that do not block these connectivity segments are identified, thereby determining the unblocked high-entropy-increase-potential residue connectivity segments. Specifically, this includes:

[0056] (1) When there is only one amino acid residue S or T between two high entropy increase potential residues, the S or T residues participate in the formation of a high entropy increase potential residue connection segment that is not blocked.

[0057] (2) When there is only one isolated N, D or G residue between two high entropy potential residues in an amino acid sequence, and there is a high entropy pairing between I, V, L, F, Y across the D, N or G with a 1-3 or 1-4 / 5 relationship, then the D, N or G is considered to participate in the formation of a high entropy potential residue connected segment that is not blocked.

[0058] (3) When any one of the amino acid residues S, T, N, D, G appears in the amino acid sequence of the polypeptide chain and is adjacent to another amino acid residue of S, T, N, D, G, examine the 1-4 / 5 sequence position relationship across these two adjacent amino acid residues. When both amino acid residues at the 1-4 / 5 position are one of the I, V, L, F, Y residues, and at least one of the amino acid residues is I, V, or L, it is predicted that these two adjacent amino acid residues will participate in the formation of a high entropy increase potential residue connectivity segment that is not blocked.

[0059] (4) When a P residue appears in the amino acid sequence of the polypeptide chain, and the P residue is not in a 1-2 or 1-3 sequence position relationship with the D or N residue, and there is a 1-4 / 5 sequence position relationship across the P residue, and the amino acid residues on the 1-4 / 5 are I, V, L, F, then the predicted P residue participates in the formation of a high entropy increase potential residue connectivity segment that is not blocked.

[0060] (5) When a G or A residue appears in the amino acid sequence of a polypeptide chain, if an amino acid residue of type I, V, or L can be found on one side of the sequence containing the G or A residue to form a 1-4 / 5 sequence position relationship with the residue, and an amino acid residue of type I, V, L, or A can also be found on the other side of the sequence containing the G or A residue to form a 1-4 / 5 sequence position relationship with the residue, if there is a single D, N, G, or P residue that is not adjacent to other micro-entropy residues or corner residues among the I, V, and L residues, or if there are two adjacent S, T, N, or D residues, then it is predicted that these single D, N, G, or P residues or two adjacent S, T, N, D, and G residues will participate in forming a high-entropy-increasing potential residue connectivity segment that is not blocked.

[0061] Furthermore, the length of the unblocked high entropy increase potential residue connectivity segment is greater than or equal to 3 amino acid residues.

[0062] Furthermore, if a protein's primary structure is not predicted to be an α-helix or β-turn sequence containing a fragment of 5 or more amino acid residues, and this fragment is in the general initial thermodynamic metastable state of an unfolded protein, then if the height of the hydrophilic atom of any hydrophilic amino acid residue side chain minus the height of the adjacent hydrophobic residue side chain is less than or equal to 1, then this fragment is predicted to be β-sheet. Here, the height of the residue side chain refers to the number of covalent bonds in the shortest path connecting the top atom of the amino acid residue side chain to its main chain carbon atom.

[0063] A protein structure prediction method based on the attraction between low-entropy hydration layers of residue side chains is proposed. This method utilizes hydrotropic forces to predict protein structure. The prediction is based on the secondary structure of the protein predicted by this method, followed by tertiary structure prediction, and includes the following steps:

[0064] The similarity of amino acid residue side chains is defined as follows: the smaller the difference in the "entropy value" between two hydrophilic residue side chains, the more similar the two hydrophilic residues are considered; the smaller the difference in the "entropy value" between two hydrophobic residue side chains, the more similar the two hydrophobic residues are considered.

[0065] The amino acid residues that have achieved preferred pairing between I / V / L / F in the predicted secondary structure are marked. The hydrophobic surfaces of the hydrophobic side chain clusters that have achieved lateral adhesion between the side chains of I / V / L / F amino acid residues are marked and named "cluster hydrophobic surfaces". The protein structure is predicted to have surface adhesion between one I / V / L / F cluster hydrophobic surface and another adjacent I / V / L / F cluster hydrophobic surface.

[0066] If there is no preferred pairing relationship between I / V / L / F in a β-sheet or α-helix secondary structure, then search for a preferred pairing relationship between I / V / LY in the secondary structure. If it exists, then the hydrophobic surface of the hydrophobic side chain clusters that are laterally attached between the side chains of I / V / LY amino acid residues is also marked as a "cluster hydrophobic surface". Predict that an I / V / L / Y cluster hydrophobic surface and another adjacent cluster hydrophobic surface will undergo surface attachment to form a protein structure.

[0067] When an α-helical structure has two cluster hydrophobic surfaces, and the two cluster hydrophobic surfaces are not connected in the axial direction of the α-helix, it is predicted that one of the cluster hydrophobic surfaces with a smaller number of I / V / L / F / Y clusters will not adhere to another I / V / LF / cluster hydrophobic surface in a nearby β-sheet or α-helical secondary structure, as a possible prediction result.

[0068] Furthermore, the protein structure prediction method based on the mutual attraction between low-entropy hydration layers of residue side chains further includes the following steps:

[0069] For cases where the current prediction is β-sheet, it is necessary to mark the strongly hydrophobic residues in the β-sheet that are preferably paired and connected by the 1-3 positional relationship, i.e., I / V / L / F / W / Y. At the same time, these hydrophobic fragments connected by the 1-3 positional relationship are marked. The lengths of adjacent β-sheets in the sequence are compared. The α-helices with similar predicted lengths will fold into a sheet structure. Then, based on the distribution of the side chains of the strongly hydrophobic residues in the two β-sheets, the two β-sheets are laterally attached in a way that allows the most strongly hydrophobic residues on one β-sheet to achieve the preferred pairing with the amino acid residue side chains on the other α-helix. The state of attachment of the two β-sheets is the predicted sheet fold structure.

[0070] The present invention has the following beneficial effects:

[0071] This invention, based on research, posits that only hydrophobic interactions between amino acid residue side chains are the core driving force of protein folding. Hydrogen bonding, electrostatic forces, van der Waals forces, and ionic bonds are byproducts of hydrophobic interactions, i.e., products of entropy-enthalpy compensation. Therefore, this invention proposes a method for predicting protein structure solely based on the rules governing hydrophobic interactions between amino acid residue side chains. This invention can accurately analyze the folding dynamics and rules of protein structures, thereby enabling accurate prediction of protein folding structures. Based on these rules, it also provides methods for predicting secondary and tertiary protein structures. This helps scientists design new proteins and purposefully mutate existing protein structures for important biological applications. Attached Figure Description

[0072] Figure 1 is a schematic diagram of amino acid residues in the low-entropy hydration layer of 20 amino acids. The atoms in the dashed boxes are hydrophobic atoms.

[0073] Figure 2 shows the lateral adhesion state of two amino acid residues E and one amino acid residue K.

[0074] Figure 3 shows the lateral attachment of side chains of adjacent amino acid residues in a typical strand structure.

[0075] Figure 4 shows a schematic diagram of the sequence position relationship of an amino acid side chain with a sequence position of 2 or 3 amino acid residues, i.e., 1-4 / 5.

[0076] Figure 5 is a schematic diagram of the positional relationship of the 1-2 sequence;

[0077] Figure 6 shows the side chain of amino acid residue E with 3 carbon-carbon bonds;

[0078] Figure 7 shows the preferred entropy-increasing pairing diagram;

[0079] Figure 8 is a schematic diagram of the adhesion between the hydrophobic end of the side chain of one amino acid residue and the hydrophobic portion of the side chain of another amino acid residue.

[0080] Figure 9 is a schematic diagram of the hydrophobic surface of the cluster;

[0081] Figure 10 is a schematic diagram of the sheet's folding structure;

[0082] Figure 11 is a schematic diagram of the configuration of the "cluster hydrophobic surfaces" in a fully bonded state;

[0083] Figure 12 is a schematic diagram of a helix structure;

[0084] Figure 13 is a schematic diagram of the beta strand structure;

[0085] Figure 14 is a schematic diagram of the turn structure;

[0086] Figure 15 is a schematic diagram of an amorphous structure (random coil);

[0087] Figure 16 is a schematic diagram of the protein structure in a surface-adhesive state. Detailed Implementation

[0088] The prevailing theories regarding entropy and enthalpy changes during protein folding state generally state that as the protein transitions from an unfolded to a folded state, its conformational freedom decreases significantly, leading to a decrease in entropy. In other words, protein folding is considered to be accompanied by a decrease in entropy. Furthermore, these theories suggest that as the protein begins to fold, hydrogen bonds, hydrophobic interactions, and van der Waals forces gradually form within the molecule, releasing energy and thus causing a decrease in enthalpy. Therefore, the prevailing theory posits that entropy and enthalpy gradually decrease during protein folding. However, this invention argues that this is problematic because it fails to consider the entropy increase of the aqueous solvent in the system. Specifically, protein folding only occurs in aqueous environments; proteins cannot fold correctly in any non-aqueous solvent environment. Therefore, this invention proposes a novel theory: the protein folding process is a process of increasing system entropy and enthalpy. This new theory (the basis of this invention) posits that protein folding is driven by the entropy increase of the low-entropy hydration layer surrounding the hydrophobic atoms on the side chain surface of polypeptide residues. In other words, protein folding is driven by the entropy increase of the low-entropy hydration layer due to hydrophobic interactions between hydrophobic atoms of different residue side chains, leading to the free solvent water. Furthermore, during protein folding, the polypeptide chain molecule must first break free from the strong hydrogen bonds, electrostatic forces, and van der Waals forces formed between it and the surrounding strongly polar water molecules before forming the weaker hydrogen bonds, electrostatic forces, and van der Waals forces within the protein molecule. Therefore, protein folding is an enthalpy-increasing process, not an enthalpy-decreasing process. That is, the new concept posits that protein folding is a process where entropy increase compensates for enthalpy increase, typically achieved through a balance between entropy and enthalpy increases. Based on this theory, this invention has conducted research and simulation verification, demonstrating highly accurate prediction of protein structures. Therefore, this invention represents an original technological invention, proposing a novel theory for protein structure formation and a prediction method based on that theory. The core technology of this invention lies in its discovery that protein folding is dominated by hydrophobic interactions between the side chains of amino acid residues. This means that intramolecular hydrogen bonding, electrostatic forces, and van der Waals forces do not dominate the protein folding process. The method described in this invention ignores physical forces such as hydrogen bonds, electrostatic forces, and van der Waals forces, prioritizing the analysis of hydrophobic interactions between the side chains of amino acid residues to predict protein structure. This differs from existing methods for predicting protein structure. This invention posits that protein folding depends solely on "hydrophobic forces," meaning that the hydrophobic interactions between the low-entropy hydrated layers of amino acid residue side chains determine the precise folding pattern of the protein. The attractive force between these low-entropy hydrated layers is defined in this invention as "hydrophobic forces."

[0089] This invention posits that protein folding is primarily driven by the attraction between low-entropy hydration layers of amino acid residue side chains. The carbon and sulfur atoms in these side chains are hydrophobic, resulting in water molecules surrounding them being in a low-entropy state—that is, a low-entropy hydration layer exists around the hydrophobic atoms. The increasing entropy of this low-entropy hydration layer leads to mutual attraction and collapse between these layers, while simultaneously preventing the low-entropy hydration layer from approaching hydrophilic amino acid residues. This results in lateral attraction and adhesion between specific amino acid residue side chains, becoming one of the core mechanical mechanisms governing protein folding. During protein folding, hydrophobic interactions occur between hydrophobic amino acid residue side chains; that is, the spontaneous aggregation of these hydrophobic amino acid residue side chains drives protein folding. This invention points out that a low-entropy hydration layer also exists around the hydrophobic atoms in the side chains of hydrophilic amino acid residues. The water molecules in this layer also exhibit entropy increase, which causes some hydrophilic amino acid residue side chains to laterally adhere to neighboring hydrophilic amino acid residue side chains. This entropy increase is also an important driving force for protein folding, and it should be emphasized that this is the starting point of this invention. (See Figure 2.)

[0090] Driven by the entropy increase of water molecules in the environment, certain hydrophilic amino acid side chains spontaneously laterally adhere to each other with neighboring hydrophilic amino acid side chains (i.e., the amino acid side chains are in a parallel state). This state forms the aggregation of hydrophobic atoms in multiple hydrophilic amino acid side chains, thus achieving an entropy increase in the low-entropy hydration layer. This invention proposes that the main physical mechanism of protein folding is that the entropy increase of low-entropy water molecules drives the lateral adhesion between amino acid side chains (lateral adhesion refers to two amino acid side chains being in a nearly parallel state). The method of this invention mainly predicts which two amino acid side chains in the amino acid sequence will undergo lateral adhesion. This state is prevalent in the secondary structures of proteins, specifically the strand and helix, and is represented by the relative positional relationship of the amino acid side chain sequences. In other words, the entropy increase of the entropy hydration layer surrounding the hydrophilic amino acid side chains causes the hydrophobic portions of some hydrophilic amino acid side chains to approach each other, and also drives the hydrophilic top atoms of the amino acid side chains to approach each other. This is the first time this invention has revealed a protein folding dynamics mechanism.

[0091] For example, in the secondary structure of a strand, the side chain of an amino acid residue is usually in a nearly parallel and attached state with the side chain of an adjacent amino acid residue. In this invention, the sequence position relationship of two amino acid residue side chains separated by one amino acid residue in the amino acid residue sequence of the primary structure is named the "1-3" sequence position relationship, as shown in Figure 3.

[0092] Similarly, in the secondary structure of helix, the side chains of two amino acid residues separated by 2 or 3 amino acid residues in the side chain sequence of one amino acid residue are in an approximately parallel and attached state, as shown in Figure 4. In this invention, the relationship between the side chain of one amino acid residue and the side chain of another amino acid residue separated by 2 or 3 amino acid residues in the amino acid residue sequence of the primary structure is named the "1-4 / 5" sequence position relationship.

[0093] Furthermore, in the helix structure, the side chain of one amino acid residue will be in an approximately parallel transition state with the side chain of an adjacent amino acid residue in the sequence. In this invention, the sequence position relationship of two adjacent amino acid residues in the amino acid residue sequence of the primary structure is named the "1-2" sequence position relationship, as shown in Figure 5.

[0094] As shown in Figure 1, many hydrophilic amino acid residues also have hydrophobic carbon and sulfur atoms in their side chains. Therefore, it is believed that these hydrophobic carbon and sulfur atoms result in a low-entropy hydration layer surrounding them, meaning that a low-entropy hydration layer also exists outside the side chains of many hydrophilic amino acid residues. Furthermore, since the roots of almost all amino acid residue side chains are composed of hydrophobic carbon atoms (as shown in Figure 1), it is believed that these hydrophobic carbon atoms mask the hydrophilicity of the polypeptide backbone. Therefore, this invention treats the polypeptide backbone as hydrophobic, neglecting the hydrophilicity of the carbonyl oxygen and amide hydrogen groups in the backbone during protein structure prediction. Specifically, water molecules around the polypeptide chain must pass through the root of the side chain to approach the backbone. Since the root of the side chain is hydrophobic, the water molecules around the hydrophobic root are low-entropy. Therefore, most of the water molecules that interact hydrophilically with the polypeptide backbone are these low-entropy water molecules. This leads to the inadequate expression of the backbone's hydrophilicity; that is, there are not a large number of free water molecules that can frequently form and break hydrogen bonds with the backbone. Therefore, the hydrophilicity of the main chain is ignored when predicting protein folding. It should be noted that the side chains of amino acid residues such as E, Q, R, K, and H have relatively long hydrophobic roots, as shown in Figure 1. Therefore, it is believed that a low-entropy hydration layer also exists outside the side chains of residues such as E, Q, R, K, and H.

[0095] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the spirit of the contents disclosed in the present invention will be described in detail below. After understanding the embodiments of the present invention, any person skilled in the art can make changes and modifications based on the technology taught in the present invention without departing from the spirit and scope of the present invention.

[0096] The illustrative embodiments and descriptions of the present invention are used to explain the invention, but are not intended to limit the invention. The invention will be further described below with reference to specific embodiments.

[0097] Specific implementation method one:

[0098] This embodiment is a protein structure prediction method based on the mutual attraction between the low-entropy hydration layers of amino acid residue side chains. This embodiment mainly predicts the secondary structure of a protein by targeting the spatial structure of the amino acid sequence. First, it predicts which segments in the amino acid residue sequence of the protein's primary structure (polypeptide chain) fold into secondary structures such as α-helix (referred to as helix in this invention), β-sheet (referred to as strand in this invention), β-turn (referred to as turn in this invention), and random coil (referred to as random coil in this invention). Then, it predicts the part between the helix and strand structures: turn or random coil. Finally, it predicts the tertiary structure of the protein.

[0099] In this embodiment, the prediction method uses the common abbreviations of amino acid residues to represent each amino acid: alanine (A); valine (V); leucine (L); isoleucine (I); proline (P); phenylalanine (F); tryptophan (W); methionine (M); glycine (G); serine (S); threonine (T); cysteine ​​(C); tyrosine (Y); asparagine (N); glutamine (Q); aspartic acid (D); glutamic acid (E); histidine (H); lysine (K); arginine (R). See Figure 1. Figure 1 shows the molecular structure of 20 amino acid residues, with the hydrophobic carbon and sulfur atoms of the side chains of each amino acid residue marked with lines.

[0100] The following is a detailed description of a protein folding dynamics mechanism disclosed in this invention:

[0101] I. The preliminary work for forecasting is as follows:

[0102] First, entropy values ​​are assigned to the side chains of amino acid residues in the primary structure of a protein. These entropy values ​​are relative measures of the entropy increase potential of the low-entropy hydration layer surrounding the side chain of an amino acid residue. Specifically, the number of carbon-carbon bonds in the hydrophobic portion of the side chain is used as a metric, named "entropy value." This invention defines one carbon-carbon bond on the side chain of an amino acid residue as 1 entropy value. This "entropy value" refers to the relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the side chain of an amino acid residue. For example, the hydrophobic portion of the side chain of amino acid E has a length of 3 and contains 3 carbon-carbon bonds; therefore, the entropy value of amino acid E residue is defined as 3, as shown in Figure 6.

[0103] The entropy value of each amino acid residue on the polypeptide chain is marked below it. Considering that amino acid residues I, V, L, and F are strongly hydrophobic residues, the entropy value of I, V, L, and F is defined as 6. At the same time, considering the masking effect of hydrophilic atoms on hydrophobic atoms, the entropy value is modified accordingly. Specifically, the entropy values ​​of the side chains of amino acid residues are defined as follows: S = 0.5, T = 0.5, D = 0.5, N = 0.5, G = 0, H = 3, R = 3, E = 3, K = 4, W = 5, I = 6, V = 6, L = 6, F = 6, P = 0, C = 4, Q = 3, A = 1, M = 4, Y = 5.

[0104] Then, the amino acid residues are classified as follows:

[0105] Residues with micro-entropy increase potential include: S, T, D, and N;

[0106] Hydrophilic residues include: H, R, K, E, Q, D, N, H, Y, W, S;

[0107] Common corner residues (referred to as corner residues) include: G and P;

[0108] Hydrophobic residues include: I, V, L, F, Y, W, A, M, C, H;

[0109] Residues with high entropy increase potential include: I, V, L, K, R, F, Y, W, A, M, C, E, Q, and H.

[0110] Among them, micro-entropy residues refer to residues whose side chains contain less than or equal to 2 carbon atoms and contain at least one oxygen or hydrogen atom.

[0111] Next, determine the preferred pairings between amino acid residue side chains:

[0112] Protein folding enables the lateral attachment of multiple pairs of amino acid residue side chains within its polypeptide sequence. This invention reveals that many attachment states fully realize the entropy increase potential of the low-entropy hydration layer of each pair of amino acid residue side chains. In other words, it is considered that lateral attachment of certain pairs of amino acid residue side chains leads to the disappearance of the low-entropy hydration layer, thus fully realizing the entropy increase potential of the low-entropy hydration layer of each amino acid residue side chain. Therefore, this invention defines lateral attachment between multiple specific pairs of amino acid residue side chains as leading to a sufficient entropy increase in the low-entropy hydration layer of these two amino acid residue side chains. These amino acid residue pair relationships are defined as "preferred entropy-increasing pairings" of two amino acid residues, as shown in Figure 7.

[0113] This invention posits that when the side chains of two amino acid residues are similar, lateral adhesion between these two side chains leads to a sufficient entropy increase in the low-entropy hydration layer surrounding them. The similarity between amino acid residue side chains is defined as follows: "The smaller the difference in the entropy values ​​of the side chains of two hydrophilic amino acid residues, the more similar the two hydrophilic residues are considered. The smaller the difference in the entropy values ​​of the side chains of two hydrophobic amino acid residues, the more similar the two hydrophobic residues are considered." For example, when the entropy values ​​of two amino acid residues are both less than or equal to 4, and their entropy difference is less than or equal to 1, the side chains of the two amino acid residues are considered similar. When the entropy values ​​of two amino acid residues are both greater than or equal to 4, and their entropy difference is less than or equal to 2, the side chains of the two amino acid residues are considered similar. For example, amino acid residue L is similar to amino acid residue W. For example, the side chain of an amino acid residue E and the side chain of another amino acid residue Q are similar because they have the same entropy value and are both hydrophilic side chains. Therefore, the lateral attachment of the side chains of the two amino acid residues will cause sufficient mutual attraction and collapse between the low-entropy water layers of the two amino acid residue side chains, and will not cause the hydrophilic end of the side chain of one amino acid residue to attach to the hydrophobic part of the side chain of the other amino acid residue, as shown in Figure 8.

[0114] Based on the similarity between amino acid residue side chains, this invention defines a list of preferred entropy-increasing pairings between amino acid residue side chains as follows:

[0115] FF, QQ, KK, EE, TT, RR, AA, SS, MM;

[0116] MA, QE, TS, RE, KR, EK, FY, WY, WE, RW, RY, KY, HV, DE, NE, DQ, NQ;

[0117] QR / K, I / V / L / FI / V / L / F, I / V / LY, I / V / LK, I / V / LR, I / V / LA, I / V / LW, I / V / L / F / YW, I / V / LM, CM / Y, MI / V / L; A / TD / N. Where " / " indicates an OR relationship.

[0118] For example, in the helix structure, the two amino acid residue side chains that are separated by 2 and 3 amino acid residues in the sequence are in an approximately parallel attachment state. When the lateral attachment between amino acid residue side chains occurs between amino acid residue side chains that are "preferred entropy-increasing pairings", it is considered that these amino acid residue side chains have achieved sufficient entropy increase in their low-entropy hydration layer, that is, they have achieved "preferred entropy-increasing pairings".

[0119] II. Specific Prediction of Protein Structure

[0120] Based on the sequence position of amino acid residue side chains, amino acid residue classification and entropy value, as well as the preferred pairing between amino acid residue side chains, the following process is used to predict protein structure;

[0121] S100. Mark the entropy value below each amino acid in the primary structure amino acid sequence of the protein whose structure needs to be predicted, forming a table.

[0122] S200. In another line below the amino acid sequence of the protein's primary structure, repeat the amino acids of these "high entropy increase potential residues" below each high entropy amino acid residue. This forms a segment composed of adjacent connected segments of high entropy increase potential residues in the amino acid sequence of the protein's primary structure, which is the original "high entropy increase potential residue connected segment". In addition, on a new line, repeat the amino acids below the corresponding positions of low entropy residues and turn residues in the amino acid sequence of the protein's primary structure.

[0123] In the protein structure prediction method described in this invention, the first step is to predict which segments of the amino acid sequence will fold into typical strand and helix structures in the protein secondary structure: that is, to first predict which sequence segment in the amino acid sequence of the protein primary structure will fold into a typical strand or helix secondary structure, and then to further predict whether the segment will fold into a strand or helix structure.

[0124] The present invention describes the prediction of protein primary structure amino acid sequences folding into strand or helix secondary structures based on "unblocked high-entropy-increase-potential residue connectivity segments". The definition of an "unblocked high-entropy-increase-potential residue connectivity segment" in this invention is: a sequence of amino acids in the protein primary structure consisting of two adjacent high-entropy-increase-potential residues and a non-high-entropy-increase-potential residue that does not block the connectivity of these high-entropy-increase-potential residues. The length of an "unblocked high-entropy-increase-potential residue connectivity segment" is greater than or equal to 3 amino acid residues.

[0125] The method for determining the "unblockable high entropy increase potential residue connectivity segment" is as follows:

[0126] First, segments consisting of consecutive high-entropy-increase-potential residues adjacent to each other in the sequence are identified and named the original "high-entropy-increase-potential residue connected segments"; then, non-high-entropy-increase-potential residues that do not block these segments are determined.

[0127] (1) When there is only one amino acid residue S or T between two high entropy increase potential residues, it is considered that the S or T does not block the high entropy increase potential residue connection segment, that is, it is considered that the S or T amino acid residue does not block the high entropy increase potential residue connection segment, and participates in the formation of a continuous high entropy increase potential residue segment. The formation of a continuous high entropy increase potential residue segment means participating in the high entropy increase potential residue connection segment.

[0128] (2) When there is only one isolated N, D or G amino acid residue between two high entropy increase potential residues in an amino acid sequence, and there is a high entropy pairing between I, V, L, F, Y with a 1-3 or 1-4 / 5 relationship across the D, N or G (at least one of the paired amino acid residues is I, V or L), then it is considered that the D, N or G does not block the high entropy increase potential residue connection segment, that is, it is considered that the N, D or G form a continuous high entropy increase potential residue segment.

[0129] If the D or N is predicted as a helix by subsequent methods, and if the side entropy of the amino acid residues in the helix structure that have a 1-4 / 5 relationship with the D or N is greater than or equal to 4, then the D or N will be predicted as a turn structure.

[0130] (3) When any one of the amino acid residues S, T, N, D, or G appears in the amino acid sequence of the polypeptide chain, and it is adjacent to another amino acid residue of S, T, N, D, or G, examine the 1-4 / 5 sequence position relationship across these two adjacent amino acid residues. If both amino acid residues at the 1-4 / 5 position are one of I, V, L, F, or Y, and at least one of them is I, V, or L, then it is predicted that these two adjacent amino acid residues will not block the formation of the high-entropy-increasing potential residue connected fragment. If the high-entropy-increasing potential residue connected fragment containing these two adjacent amino acid residues is not predicted as a helix by the following method, then the adjacent amino acid residues in the two consecutive S, T, N, D, or G sequences are predicted as a turn structure T, as a prediction result.

[0131] (4) When a P amino acid residue appears in the amino acid sequence of the polypeptide chain, and the P amino acid residue is not in a 1-2 or 1-3 sequence position relationship with the D or N amino acid residue (e.g., PQD, PCCN), when there is a 1-4 / 5 sequence position relationship across the P amino acid residue, and the amino acid residues on the 1-4 / 5 are I, V, L, F, then the predicted P amino acid residue does not block the high entropy increase potential residue connection fragment. However, if the fragment containing P is not ultimately predicted as a helix, then P is predicted as a turn structure.

[0132] (5) When a G or A amino acid residue appears in the amino acid sequence of a polypeptide chain, if an amino acid residue of type I, V, or L can be found on one side of the sequence containing the G or A amino acid residue to form a 1-4 / 5 sequence position relationship with the amino acid residue, and an amino acid residue of type I, V, L, or A can also be found on the other side of the sequence containing the G or A amino acid residue to form a 1-4 / 5 sequence position relationship with the G or A amino acid residue, if there is a single D, N, G, or P amino acid residue that is not adjacent to other micro-entropy residues or corner residues among the aforementioned I, V, and L residues, or two adjacent S, T, N, D amino acid residues, then it is predicted that these single D, N, G, or P amino acid residues or two adjacent S, T, N, D, G amino acid residues will not block the high entropy increase potential residue connection fragment in the fragment.

[0133] Such as: IVL 1-4 / 5G / A 1-4 / 5IVLF (for example, VAAQGRARL), IVL 1-4 / 5A 1-4 / 5A 1-4 / 5IVLF, IVL 1-4 / 5A 1-4 / 5A 1-4 / 5A (for example, IMQDAGVTANTRA).

[0134] S300. In the amino acid sequence of the polypeptide chain, identify fragments with a length of 5 or fewer amino acid residues and containing 3 or more microentropy residues or turn residues. Predict that these fragments will fold into a turn structure. This prediction is not made for the ends of the amino acid sequence of the polypeptide chain. Alternatively, in the amino acid sequence of the polypeptide chain, identify fragments with a length of 5 or fewer amino acid residues and containing 3 or more microentropy residues or turn residues. If the fragment contains at least one turn residue (i.e., one G or P amino acid residue), predict that these amino acid residues will fold into a turn structure.

[0135] Based on the above judgment, the predicted high entropy increase potential residue connectivity fragments are marked below the original amino acid sequence;

[0136] S400. Predict whether the high-entropy-increase-potential residues in the amino acid sequence of the polypeptide chain will fold into a helix structure or a strand structure:

[0137] S401. Assume that the high entropy increase potential residues in the amino acid sequence of the polypeptide chain are connected segments as helix or strand structures.

[0138] (A) Assuming a high-entropy-increasing-potential residue-connected fragment folds into a helix structure, mark whether each amino acid residue side chain in the fragment achieves a preferred pairing relationship in the helix state. If a side chain of an amino acid residue in the high-entropy-increasing-potential residue-connected fragment achieves a preferred pairing, mark it with * below that amino acid; otherwise, leave it unmarked. Here, preferred pairing refers to pairing achieved through a "1-4 / 5" sequence position relationship. Furthermore, if the high-entropy-increasing-potential residue-connected fragment contains preferred pairings of I / V / L and I / V / L achieved through a 1-2 sequence position relationship (e.g., EAQRLL), or if the high-entropy-increasing-potential residue-connected fragment contains a series of positional relationships such as I / V / L 1-4 / 5G / A 1-4 / 5I / V / L / F, mark these IVFGAs with * as well.

[0139] (B) Assuming that a high entropy increase potential residue connectivity segment is predicted to be a strand structure, mark whether each amino acid residue side chain in the strand state has achieved a preferred pairing of I / V / L / F and I / V / L / F of a certain 1-3 sequence position relationship (e.g. IDV). The two I / V / L / F amino acid residues that have achieved the preferred pairing of I / V / L / F and I / V / L / F of the 1-3 sequence position relationship are marked with &.

[0140] Furthermore, when a pair of I / V / L and I / V / L with a 1-3 sequence position relationship is marked with &, it is called marked with &-&. And both I / V / L can find I / V / L / F / Y or A with a 1-4 / 5 sequence position relationship in the high entropy increase potential residue connected fragment to form a preferred pair. At this time, the &-& marked by the two I / V / L is considered invalid, that is, the &-& is no longer marked.

[0141] S402. The * and & marked below the amino acid residue sequences are compared and analyzed as follows:

[0142] When an amino acid residue in a high entropy increase potential residue-connected fragment is not tagged with &-& and more than 80% of the amino acid residues are tagged with *, the fragment is predicted to be helix;

[0143] When a high-entropy-increase-potential residue-connected fragment is marked with &-& (excluding I / V / LSI / V / L marked with &-&, such as VSV), the fragment is usually predicted to be a strand. An exception is when, in a region "far from" the &-&-marked area of ​​a high-entropy-increase-potential residue-connected fragment (where "far from" is defined as separated by two consecutive micro-entropy residues or corner residues, or by at least three hydrophilic residues or one amino acid residue P), any fragment of a high-entropy-increase-potential residue-connected fragment of length greater than 5 can be found as a sub-fraction. In this sub-fraction, more than 85% of the amino acid residues are marked with *, and there are 1-4 / 5 sequence position relationships between I / V / L / F / Y and I / V / L / F / Y for other amino acid residues besides those marked with &-&. In this case, the sub-fraction is predicted to fold into a helix structure, and the remaining portion of the high-entropy-increase-potential residue-connected fragment is predicted to be a strand. When the number of "high entropy increase potential residue connectivity fragments" marked with * is less than 50% of the number of amino acid residues in the fragment, the fragment is predicted as a strand.

[0144] In a segment predicted as a strand, if two consecutive microentropy residues D or N appear in the amino acid residue sequence, then these two consecutive microentropy residues D or N are predicted as a turn structure as a prediction result.

[0145] S403. Search for amino acid residues adjacent to the amino acid residue sequence fragment predicted as a helix on both sides of the fragment. Determine whether the side chains of the searched amino acid residues have a preferred pairing with the amino acid residues in the predicted helix structure in a 1-4 / 5 sequence position relationship. If a preferred pairing exists, predict that the residues in this preferred pairing will fold into a helix structure. Repeat this operation until the first adjacent amino acid residue does not have a preferred pairing with the amino acid residues in the initially predicted helix structure in a 1-4 / 5 sequence position relationship.

[0146] S404. Calculate the entropy increase achieved by each amino acid residue in the helix and strand states, respectively. (Two initiation scenarios)

[0147] The method for calculating the entropy increment of a folded strand is as follows: when a preferred pairing occurs in the 1-3 sequence position relationship of the amino acid sequence of the polypeptide chain, the entropy value of the preferred pairing amino acid residue is marked below the amino acid residue and recorded as the entropy increment of the amino acid residue. When there are two preferred pairings in the 1-3 sequence position relationship for a certain amino acid residue, the entropy increment of the amino acid residue is accumulated twice.

[0148] The method for calculating the entropy increment in helix sequences is as follows: When a preferred pairing occurs in the 1-4 / 5 sequence position relationship, the entropy value of the amino acid residue with the preferred pairing is marked below that amino acid residue and recorded as its entropy increment. If an amino acid residue achieves multiple preferred pairings in the 1-4 / 5 sequence position relationship, its entropy increment is accumulated twice. In the helix entropy increment statistics, when two I / V / L / M residues are adjacent in the sequence, i.e., have a "1-2" sequence position relationship, these two I / V / L / M residues are considered to have achieved one entropy increment, and this is accumulated once.

[0149] Compare the entropy increment achieved by each amino acid residue side chain in the two schemes of folding into helix and strand. If the larger entropy increment calculated by the two schemes is 20% greater than the smaller one, the one with the larger overall entropy increment is the predicted result. If the difference in the overall entropy values ​​predicted as strand and helix does not exceed 20% of the smaller value, then the fragment is predicted as either strand or helix as two possible prediction results. In the entropy statistics of strand, I / V / LGI / V / L is considered to represent the optimal entropy pairing of two IVLs, where "-" indicates an arbitrary residue.

[0150] S405. When a fragment of 5 or more residues is in a straight state, and no hydrophilic atom (nitrogen or oxygen atom) of an amino acid residue side chain is at the same height (the height of the residue side chain) as a hydrophobic atom of an adjacent amino acid residue side chain (in this method, amino acid residues S and T are considered hydrophobic, meaning the oxygen atom at the top of the S and T amino acid residue side chains is considered hydrophobic), then the fragment is ultimately predicted to be a straight fragment. If a hydrophilic atom (nitrogen or oxygen atom) of an amino acid residue side chain is at the same height as a hydrophobic atom of an adjacent amino acid residue side chain, hydrophilic-hydrophobic repulsion will occur between the hydrophilic and hydrophobic atoms, disrupting the straight state. The height of the residue side chain refers to the number of covalent bonds in the shortest path connecting the top atom of the amino acid residue side chain to its main chain carbon atom.

[0151] When the predicted length of a strand is greater than or equal to 12 amino acid residues, the search is conducted to determine if there are three consecutive hydrophilic residues in the middle of the segment. If so, the prediction is that these three amino acid residues form a turn structure.

[0152] S406. When a segment of amino acid residues in the amino acid sequence of a polypeptide chain is predicted to be a helix, the IVLF group of connected segments that are laterally attached to each other through a 1-4 / 5 or 1-2 sequence position relationship in the helix structure is marked. The fragment sequence is searched for an isolated I / V / LF amino acid residue that is not laterally attached to any I / V / LF side chain of any IVLF amino acid residue in the group. If there is such a residue, and the I / V / LF amino acid residue fragment is adjacent to or separated from the S / T / N / D amino acid residue fragment by only one amino acid residue, then the fragment containing the I / V / LF and S / T / N / D is predicted to no longer be a helix.

[0153] When the predicted length of the helix structure is less than or equal to 6 amino acid residues, predicting the fragment as a random coil is also a possible prediction result.

[0154] S500, the presence of two consecutive amino acid residues in a high entropy-increasing potential connected fragment that was not predicted as a helix is ​​predicted as a turn structure and labeled as T; the two consecutive amino acid residues are any one or two of S, T, N, D, and G, for example SS or ST.

[0155] In the primary structure of a protein, a sequence of four or more consecutive hydrophilic amino acid residues is called a "continuous hydrophilic residue sequence," such as EEQQK. When a high-entropy-increase-potential connected fragment not predicted as a helix contains a segment longer than seven amino acid residues, and the residues in the low-entropy-increase-potential residues and / or turn residues and / or amino acid residue A and / or continuous hydrophilic residue sequence comprise 50% or more of these residues, the segment is predicted to fold into a random coil structure.

[0156] S600: Search for a fragment consisting of 5 or more consecutive hydrophilic residues in the amino acid sequence of the search polypeptide chain. If the fragment has been predicted as a high entropy increase potential residue-connected fragment, then the fragment consisting of 5 or more consecutive hydrophilic residues is predicted as a random coil as a prediction result.

[0157] Based on the above judgment, the predicted fragments containing the turn structure were marked below the original amino acid sequence.

[0158] In addition, hydrophilic residues with entropy values ​​greater than or equal to 3 in the current predicted turn structure and random coil structure are marked, and the preferred pairing relationship between these hydrophilic residues is marked. The lateral attachment state of the side chains of amino acid residues with preferred pairing relationship is taken as a predicted result. When it is not yet possible to determine that the predicted turn structure is a turn structure, the micro-entropy residues in the predicted turn structure are marked, and the preferred pairing relationship between these micro-entropy residues is marked. The lateral attachment state of the side chains of these amino acid residues with preferred pairing relationship is taken as a predicted result.

[0159] Simultaneously, the hydrophobic residues I / V / L / F / M in the turn structure and random coil structure are labeled, and the side chain of the hydrophobic residue will attach to the side chain of the adjacent I / V / L / F / Y / W / K amino acid residues as a prediction result; the I / V / L / F amino acid residues in the two secondary structures linked by the turn structure and random coil structure are labeled, and the attachment of I / V / L / F in the two secondary structures is taken as a prediction result.

[0160] After the above judgment, if the protein's primary structure does not contain a segment with 5 or more amino acid residues in the sequence predicted as helix or turn, and this segment is in the general initial thermodynamic metastable state of unfolded proteins, then if the height of the hydrophilic atom of any hydrophilic amino acid residue side chain minus the height of the adjacent hydrophobic residue side chain is less than or equal to 1, then this segment is predicted as a strand. The determination of the general initial thermodynamic metastable state of unfolded proteins can be based on the "General Initial Thermodynamic Metastable State of Unfolded Proteins" by Yang Lin et al.

[0161] Specific Implementation Method Two:

[0162] This embodiment is a protein structure prediction method based on the mutual attraction between the low-entropy hydration layers of amino acid residue side chains. This embodiment is based on the prediction of the spatial structure of the amino acid sequence of the protein in Specific Embodiment 1, and then predicts the tertiary structure of the protein. That is, the secondary structure is first predicted using Specific Embodiment 1, and then the tertiary structure is further predicted based on the secondary structure.

[0163] First, the similarity of amino acid residue side chains is defined: the smaller the difference in the "entropy value" between two hydrophilic residue side chains, the more similar the two hydrophilic residues are considered; the smaller the difference in the "entropy value" between two hydrophobic residue side chains, the more similar the two hydrophobic residues are considered.

[0164] The amino acid residues that have achieved preferred pairing between I / V / L / F in the predicted secondary structure are marked, and the hydrophobic surfaces of the hydrophobic side chain clusters that have achieved lateral adhesion between the side chains of I / V / L / F amino acid residues are marked and named "cluster hydrophobic surfaces", as shown in Figure 9. The protein structure is predicted to have surface adhesion between one I / V / L / F cluster hydrophobic surface and another adjacent I / V / L / F cluster hydrophobic surface.

[0165] If there is no preferred pairing relationship between I / V / L / F in a strand or helix secondary structure, then search for a preferred pairing relationship between I / V / LY in the secondary structure. If it exists, then the hydrophobic surface of the hydrophobic side chain clusters that are laterally attached between the side chains of I / V / LY amino acid residues is also marked as a "cluster hydrophobic surface". It is predicted that an I / V / L / Y cluster hydrophobic surface and another adjacent cluster hydrophobic surface will undergo surface attachment to form a protein structure.

[0166] When a helix structure has two cluster hydrophobic surfaces, and the two cluster hydrophobic surfaces are not connected in the axial direction of the helix, it is predicted that one of the cluster hydrophobic surfaces with a smaller number of I / V / L / F / Y clusters will not adhere to another I / V / LF / cluster hydrophobic surface in a nearby strand or helix secondary structure, as a possible prediction result.

[0167] For cases where the current prediction is a strand, it is necessary to mark the strongly hydrophobic residues (I / V / L / F / W / Y) of the preferred paired connected segments in the 1-3 positional relationship of the strand. At the same time, mark these hydrophobic segments connected by the 1-3 positional relationship, compare the lengths of adjacent strands in the sequence, and strands with similar predicted lengths will fold into a sheet structure. Then, based on the distribution of the strongly hydrophobic residue side chains of the two strands, the strands are laterally attached in a way that the most strongly hydrophobic residues (I / V / L / F / W / Y) on one strand can achieve preferred pairing with the amino acid residue side chains on the other strand. The attached state of the two strands is the predicted sheet folded structure (in fact, the βsheet is composed of βstrands), as shown in Figure 10.

[0168] In predicting that the surface adhesion of a cluster of hydrophobic surfaces will result in a protein structure, we can actually start by predicting a secondary structure that is predicted to be a turn structure. For the two predicted secondary structures connected to the turn structure, the hydrophobic surface of the hydrophobic side chain cluster is marked and named "cluster hydrophobic surface". Using the "cluster hydrophobic surface", the relative position of the two secondary structures connected to the turn structure is used to determine the local tertiary structure of the predicted secondary structure. At the same time, we also need to consider the adhesion state of the two secondary structures, as well as the adhesion state of other secondary structure "cluster hydrophobic surfaces", to ensure that the "cluster hydrophobic surfaces" are fully adhered to each other. The configuration of the corresponding fully adhered state is the prediction result, as shown in Figure 11.

[0169] Example 1

[0170] For ease of explanation, in the embodiments, amino acid residues folded into helix structures are labeled H; amino acid residues folded into strand structures are labeled E; turn structures are labeled T; and amorphous structures (random coils) are labeled C. It should be noted that the present invention includes, but is not limited to, these labeling methods, and other labeling forms may be used in other embodiments.

[0171] A method for progressively predicting the tertiary structure of a protein is achieved by predicting four secondary structures:

[0172] I. The four secondary structures are as follows:

[0173] ① Helix structure (see Figure 12);

[0174] ② Beta strand (see Figure 13);

[0175] ③ Turn structure (see Figure 14);

[0176] ④ Random coil (see Figure 15).

[0177] II. Preprocessing steps:

[0178] Amino acid residue fragments predicted as helix were labeled H, amino acid residue fragments predicted as strand were labeled E, and amino acid residue fragments predicted as turn structures were labeled T.

[0179] In this embodiment, protein number 1054 from the PDB database was selected to predict its structure:

[0180] 1. Label the entropy value below each amino acid in the protein amino acid sequence whose structure needs to be predicted, forming a table, where 'a' represents an entropy value of 0.5. See below:

[0181] 2. On a new line below the protein amino acid sequence, repeat these "high entropy increase potential residues" below each high entropy increase potential residue. This marks the segments formed by adjacent connected segments of high entropy increase potential residues in the sequence, which are the original "high entropy increase potential residue connected segments". Additionally, on a new line below the protein amino acid sequence, repeat these low entropy residues and corner residues below their corresponding positions, as follows:

[0182] 3. Identify segments in the sequence composed of consecutive high-entropy-increase-potential residues, and name them the original "high-entropy-increase-potential residue connected segments". Furthermore, if two high-entropy-increase-potential residues are separated only by one amino acid residue S or T, it is considered that the S or T residue does not block the high-entropy-increase-potential residue connected segment, and the high-entropy-increase-potential residue connected segment remains valid; that is, it is considered that the S or T amino acid residue participates in the formation of a consecutive high-entropy-increase-potential residue segment.

[0183] The remaining micro-entropy residues and turn residues that may block the high-entropy-increase-potential connected segments after removing these S and T residues are marked in the fourth row of the table below.

[0184] 4. When two high-entropy-increasing-potential residues in an amino acid sequence are separated only by an isolated N, D, or G amino acid residue, and there is a high-entropy pairing between I / V / L / F / Y with a 1-3 or 1-4 / 5 sequence position relationship across that D, N, or G (at least one of the paired amino acid residues is I / V / L), then it is considered that the D, N, or G does not block the high-entropy-increasing-potential residue connection segment, and the high-entropy-increasing-potential residue connection segment still exists; that is, it is considered that the N, D, or G form a continuous high-entropy-increasing-potential residue segment. The remaining micro-entropy residues and corner residues that may block the high-entropy-increasing-potential residue connection segment are marked in the last row (add a row below the table above, which is the last row in this step).

[0185] When two adjacent S / T / N / D / G residues appear in an amino acid sequence, if there is a 1-4 / 5 sequence positional relationship between I / V / L / F amino acid residues spanning these two residues, then the two adjacent S / T / N / D / G residues are predicted not to block the formation of a high-entropy-increasing-potential residue connectivity segment. If the high-entropy-increasing-potential residue connectivity segment containing these two adjacent S / T / N / D / G residues is not predicted as a helix by the following method, then the two consecutive S / T / N / D / G residues are predicted as a turn structure T, as a prediction result.

[0186] When an amino acid residue P appears in an amino acid sequence, and P is not in a 1-2 or 1-3 sequence position relationship with D / N (e.g., PQD, PCCN), if there is a 1-4 / 5 sequence position relationship between I / V / L / F amino acid residues across P, then the predicted P amino acid residue does not block the high entropy increase potential residue connectivity segment. However, if the segment containing P is not ultimately predicted as a helix, then P is predicted as a turn structure.

[0187] The remaining micro-entropy residues and turn residues that may block high-entropy-increment potential connected segments are listed in the last line:

[0188] 5. When a fragment with the amino acid residue positional relationship of IVL 1-4 / 5G / A 1-4 / 5IVLF (e.g., VAAQGRARL), IVL 1-4 / 5A 1-4 / 5A 1-4 / 5IVLF (e.g., IMQDAGVTANTRV), or IVL 1-4 / 5A 1-4 / 5A 1-4 / 5A (e.g., IMQDAGVTANTRA) appears in the amino acid sequence, if a single D, N, G, P or two adjacent S / T / N / D amino acid residues are present in the fragment, it is predicted that these single D, N, G, or P residues or two adjacent S / T / N / D / G amino acid residues will not block the connection of residues with high entropy increase potential in the fragment.

[0189] When a fragment with the amino acid residue positional relationship IVL 1-4 / 5G / A 1-4 / 5IVLF (e.g., VAAQGRARL), IVL 1-4 / 5A1-4 / 5A 1-4 / 5IVLF (e.g., ), IVL 1-4 / 5A 1-4 / 5A 1-4 / 5A appears in the amino acid sequence, if there is a single D, N, G, or P amino acid residue or two adjacent S / T / N / D amino acid residues that are not adjacent to other low-entropy residues or turn residues in the fragment, it is predicted that these single D, N, G, or P amino acid residues or two adjacent S / T / N / D / G amino acid residues will not block the connection of high-entropy-increasing potential residues in the fragment.

[0190] 6. According to S300: If a sequence contains 5 or fewer amino acid residues and at least 3 microentropy residues or turn residues, it is predicted that the fragment will fold into a turn structure. If the fragment contains two adjacent microentropy residues or turn residues that are not adjacent to another microentropy residue or turn residue, it is accurately predicted that the two adjacent microentropy residues or turn residues will fold into a turn structure. For specific prediction methods for turn structures, see Sections 30, 31, and 32. If a sequence contains 5 or fewer amino acid residues and at least 3 microentropy residues or turn residues, and if the fragment contains at least one turn residue (i.e., one G or P amino acid residue), it is predicted that these amino acid residues will fold into a turn structure (T).

[0191] Based on this, the predicted turn structure of the protein is marked in the third row:

[0192] 7. In summary, the segments with preliminarily predicted corner structures are marked in the third row of the table below:

[0193] 8. The corresponding predicted micro-entropy residue connectivity segments are shown in the table below.

[0194] 9. The method for predicting whether a high-entropy-increase-potential residue connected fragment in an amino acid sequence folds into a helix or a strand structure is as follows: First, assume that a high-entropy-increase-potential residue connected fragment folds into a helix structure. Mark whether each amino acid residue side chain in the helix state achieves a preferred pairing relationship. If a side chain of an amino acid residue achieves a preferred pairing, mark it with an asterisk (*); otherwise, leave it unmarked. Here, preferred pairing refers to pairing achieved through the "1-4 / 5" sequence position relationship. Furthermore, if there are preferred pairings of IVL and IVL achieved through the 1-2 sequence position relationship (e.g., EAQRLL), or if there are IVL 1-4 / 5G / A1-4 / 5IVLF series position relationships, mark these IVL GAs with an asterisk (*).

[0195] 10. Assuming a high-entropy-increase-potential residue connected fragment is predicted to be a strand structure, each amino acid residue side chain is labeled to indicate whether it achieves a preferred pairing (e.g., IDV) of a certain 1-3 positional IVLF and IVLF in the strand state. The amino acid residues of the two IVLFs that achieve this preferred pairing are marked with &. Furthermore, when a pair of 1-3 positional IVLs and IVLs are marked with & (referred to as marked &-&), and both IVLs can find a preferred pairing of IVLFY or A with a 1-4 / 5 positional sequence in the high-entropy-increase-potential residue connected fragment, the &-& marking of these two IVLs is considered invalid, meaning the &-& marking is no longer applied.

[0196] 11. Compare the * and & in the first two tables, and perform the following comparative analysis on the * and & marked below the amino acid residue sequence: When no amino acid residues in a high entropy increase potential residue connected fragment are marked &-&, and more than 80% of the residues are marked with *, the fragment is predicted to be helix.

[0197] 12. When a segment of a high-entropy-increase-potential residue-connected fragment is marked with &-& (excluding I / V / LSI / V / L marked with &-&, such as the sequence VSV), the fragment is usually predicted to be a strand. An exception is if, in a region “far from” the &-&-marked area of ​​a high-entropy-increase-potential residue-connected fragment (where “far from” is defined as separated by two consecutive micro-entropy residues or corner residues, or at least three hydrophilic residues), a sub-fraction of one amino acid sequence longer than 5 is found, where more than 85% of the amino acid residues are marked with *, and there is a 1-4 / 5 sequence position relationship between IVLFY and IVLFY unrelated to the &-&-marked amino acid residues, then the sub-fraction is predicted to fold into a helix structure, and the remaining portion of the high-entropy-increase-potential residue-connected fragment is predicted to be a strand. When the number of *-marked segments in a high-entropy-increase-potential residue-connected fragment is less than 50% of the total number of amino acid residues in the fragment, the fragment is predicted to be a strand.

[0198] 13. When a segment of a high-entropy-increase-potential residue-connected fragment is marked with &-& (excluding I / V / LSI / V / L marked with &-&, such as VSV), the fragment is usually predicted to be a strand. An exception is if, in a region “far from” the &-&-marked area of ​​a high-entropy-increase-potential residue-connected fragment (where “far from” is defined as separated by two consecutive micro-entropy residues or corner residues, or at least three hydrophilic residues), a sub-fraction of one amino acid sequence longer than 5 is found, where more than 85% of the amino acid residues are marked with *, and there is a 1-4 / 5 sequence position relationship between IVLFY and IVLFY unrelated to the &-&-marked amino acid residues, then the sub-fraction is predicted to fold into a helix structure, and the remaining portion of the high-entropy-increase-potential residue-connected fragment is predicted to be a strand. When the number of *-marked segments in a high-entropy-increase-potential residue-connected fragment is less than 50% of the total number of amino acid residues in the fragment, the fragment is predicted to be a strand.

[0199] 14. If the high entropy increase potential residue connected segment of the two adjacent S / T / N / D / G residues is not predicted as a helix by the following method, then the two consecutive S / T / N / D / G residues are predicted as a turn structure T, as a prediction result.

[0200] 15. Search for adjacent amino acid residues flanking the amino acid residue sequence fragment predicted as a helix. Determine if the side chain of this amino acid residue has a preferred pairing with the amino acid residues in the predicted helix structure (1-4 / 5 relationship). If a preferred pairing exists, predict that the amino acid residues in this preferred pairing will fold into a helix structure. Repeat this operation until the first adjacent amino acid residue does not have a preferred pairing with the amino acid residues in the predicted helix structure (1-4 / 5 relationship).

[0201] 16. The prediction results are listed in the third row of the table below, and the software's analysis of the actual protein structure is predicted in the fourth row, showing that the prediction is accurate.

[0202] 17. When the turn structure cannot be predicted, mark the micro-entropy residues in the predicted turn structure, and then mark the preferred pairing relationship between these micro-entropy residues. The lateral attachment state of the side chains of these amino acid residues with preferred pairing relationship is a prediction result.

[0203] 18. Mark the amino acid residues that have achieved optimal IVLF pairing in the predicted secondary structures:

[0204] The hydrophobic surfaces of hydrophobic side chain clusters that achieve lateral adhesion between IVLF amino acid residue side chains are labeled and named "cluster hydrophobic surfaces". The state in which one IVLF cluster hydrophobic surface adheres to another adjacent IVLF cluster hydrophobic surface in the amino acid residue sequence is predicted as the protein structure. When there is no preferred pairing relationship between IVLFs in a secondary structure, the search is conducted to see if there is a preferred pairing relationship between IVLYs in the secondary structure. The hydrophobic surfaces of hydrophobic side chain clusters with lateral adhesion between IVLY amino acid residue side chains are then labeled as "cluster hydrophobic surfaces", and the state in which one IVLY cluster hydrophobic surface adheres to another adjacent cluster hydrophobic surface in the amino acid residue sequence is predicted as the protein structure. When a helix structure exhibits two cluster hydrophobic surfaces, and the two cluster hydrophobic surfaces do not have a axially connected segment in the helix, it is predicted that one cluster with a smaller number of IVLFYs will not adhere to the other adjacent IVLF cluster hydrophobic surface, as shown in Figure 16.

[0205] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A protein structure prediction method based on the mutual attraction between low-entropy hydration layers of residue side chains, characterized in that: The prediction method utilizes hydrotropic force to predict protein structure; the prediction is based on the prediction of the protein's secondary structure from its primary structure; the hydrotropic force refers to the attractive force between the low-entropy hydration layers of amino residue side chains; the attractive force refers to the lateral adhesion between amino acid residue side chains driven by the entropy increase of low-entropy water molecules in the low-entropy hydration layers; the lateral adhesion refers to the two residue side chains being in a nearly parallel state.

2. The prediction method according to claim 1, characterized in that: The process of using water entropy force to predict protein structure includes: I. Pre-forecast preparations: First, entropy values ​​are assigned to the side chains of amino acid residues in the primary structure of the protein. The entropy value is a relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the side chains of amino acid residues. Then, the amino acid residue side chains in the primary structure of the protein after entropy assignment are divided into 5 residue forms: Residues with potential for micro-entropy increase: serine (S), threonine (T), aspartic acid (D), and asparagine (N); Hydrophilic residues: histidine H, arginine R, lysine K, glutamic acid E, glutamine Q, aspartic acid D, asparagine N, tyrosine Y, tryptophan W, serine S; Turning residues: glycine (G), proline (P); Hydrophobic residues: Isoleucine I, Valine V, Leucine L, Phenylalanine F, Tyrosine Y, Tryptophan W, Alanine A, Methionine M, Cysteine ​​C, Histidine H; Residues with high entropy increase potential: Isoleucine I, Valine V, Leucine L, Lysine K, Arginine R, Phenylalanine F, Tyrosine Y, Tryptophan W, Alanine A, Methionine M, Cysteine ​​C, Glutamic Acid E, Glutamine Q, Histidine H; Finally, the preferred pairing between amino acid residue side chains was determined. The preferred pairing between amino acid residue side chains is the lateral adhesion between multiple pairs of amino acid residue side chains, which leads to a sufficient entropy increase in the low-entropy hydration layer of multiple pairs of amino acid residue side chains. The pairing relationship between multiple pairs of amino acid residues is the preferred pairing between amino acid residue side chains. The preferred pairing between amino acid residue side chains includes preferred pairing between amino acid residue side chains with a 1-4 / 5 sequence position relationship, preferred pairing between amino acid residue side chains with a 1-3 sequence position relationship, and preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship. The 1-4 / 5 sequence position relationship refers to the relationship between the sequence positions of one side chain of an amino acid residue and another side chain of an amino acid residue in the primary structure of a protein, which are separated by 2 or 3 amino acid residues. The 1-3 sequence position relationship refers to the relationship between the sequence positions of one amino acid residue side chain and another amino acid residue side chain in the primary structure of a protein, separated by one amino acid residue. The 1-2 sequence position relationship refers to the sequence position relationship between one residue side chain and another adjacent residue side chain in the primary structure of a protein. II. Protein secondary structure prediction: Based on the sequence positions of the residue side chains, amino acid residue classifications, and entropy values ​​obtained in step one, as well as the preferred pairings between residue side chains, the following process is used to predict the secondary structure of proteins: S100. On a new line below the amino acid sequence labeled with entropy values, repeat the amino acid residues in the amino acid sequence that belong to high entropy increase potential residues as high entropy increase potential residue connected segments; then, on a new line, repeat the amino acid residues in the amino acid sequence that belong to low entropy residues and turn residues. S200. List the unblocked high-entropy-increase-potential residue connectivity segments, and assume that the listed unblocked high-entropy-increase-potential residue connectivity segments are either α-helices or β-sheets in the protein secondary structure: S201. Assuming that a high-entropy-increase-potential residue-connected fragment forms an α-helix, if a side chain of a residue in the high-entropy-increase-potential residue-connected fragment achieves preferred pairing between amino acid residue side chains, mark it as * below the amino acid residue of the high-entropy-increase-potential residue-connected fragment; S202. Assuming that the high entropy increase potential residue connecting fragment forms a β-sheet, if the high entropy increase potential residue connecting fragment is a preferred pairing between the side chains of amino acid residues in a 1-3 sequence position relationship, then the amino acid residues of the high entropy increase potential residue connecting fragment are marked with & below them; S203. Analyze the * and & symbols: S2031. If none of the high-entropy-increase-potential residues in the primary structure of a protein are marked as &-&, and more than 80% of the amino acid residues in such high-entropy-increase-potential residues are marked as *, then it is predicted that such high-entropy-increase-potential residues form an α-helix. S2032. If the high entropy increase potential residues in the primary structure of a protein are marked with &-&, it is predicted that the high entropy increase potential residues will form a β-sheet; if the number of * marks in the high entropy increase potential residues in the primary structure of a protein is less than 50% of the number of amino acid residues in the high entropy increase potential residues, it is predicted that the high entropy increase potential residues will form a β-sheet. For S300, the amino acid side chains on both sides of the high entropy increase potential residue connected fragment predicted by S200 to form an α-helix are searched. If the amino acid residues in the amino acid side chains on both sides have a 1-4 / 5 sequence position relationship with the amino acid residues in the high entropy increase potential residue connected fragment, the amino acid residue side chains with the predicted 1-4 / 5 sequence position relationship are preferably paired to form an α-helix. S400, based on the predictions of the S200 hypothesis, determines whether the protein secondary structure consists of α-helices or β-sheets: The entropy increments of amino acid residues in the primary structure of proteins predicted as α-helices and β-sheets are calculated separately. The entropy increments of the predicted α-helices and β-sheets are compared. If the larger entropy increment is greater than 20% of the other entropy increment, the prediction result of the larger entropy increment is determined as the final prediction result. If the larger entropy increment is less than 20% of the other entropy increment, the prediction of the S200 hypothesis forming α-helices or β-sheets in the secondary structure of the protein is determined as a potential result. Among them, the entropy increment statistical method of β-sheet is: the preferred pairing between amino acid residues with a 1-3 sequence position relationship in the primary structure of the protein, and the entropy value of the amino acid residues in the preferred pairing is used as the entropy increment. Statistical method for entropy increment of α-helix: Preferred pairing of amino acid residues in the primary structure of a protein with a 1-4 / 5 sequence position relationship is used as the entropy increment. S500. If there are two consecutive amino acid residues with low entropy increase potential in the connected segment of high entropy increase potential residues that are not predicted to be α-helices in the primary structure of a protein, then the two amino acid residues are predicted to form a turn and are marked as T. A sequence of four or more consecutive hydrophilic amino acid residues in the primary structure of a protein is called a "continuous hydrophilic residue sequence". If, in a high-entropy-increasing-potential residue sequence in the primary structure of a protein that is not predicted to be an α-helix, there are sequences of more than seven amino acid residues, and the number of residues in the low-entropy-increasing-potential residue sequence and / or turn residues and / or amino acid residue A and / or continuous hydrophilic residue sequence accounts for more than 50% of the number of sequences of more than seven amino acid residues, then the sequence of more than seven amino acid residues is predicted to form a random coil, and this is considered a potential prediction result. S600. If a protein's primary structure consists of a fragment composed of five or more consecutive hydrophilic amino acid residues, and this fragment has been predicted as a high-entropy-increase-potential residue-connected fragment, then the fragment composed of five or more consecutive hydrophilic amino acid residues is predicted to form a random coil, and this is taken as a potential prediction result. This completes the protein structure prediction method based on the mutual attraction relationship between the low-entropy hydration layers of residue side chains.

3. The prediction method according to claim 2, characterized in that: The determination of preferred pairings between amino acid residue side chains includes: FF, QQ, KK, EE, TT, RR, AA, SS, MM; MA, QE, TS, RE, KR, EK, FY, WY, WE, RW, RY, KY, HV, DE, NE, DQ, NQ; QR / K, I / V / L / FI / V / L / F, I / V / L / F / YI / V / L / F / Y, I / V / LY, I / V / LK, I / V / LR, I / V / LA, I / V / LW, I / V / L / F / YW, I / V / LM, CM / Y, MI / V / L, A / TD / N.

4. The prediction method according to claim 2, characterized in that: In step S201, the preferred pairing between amino acid residue side chains refers to the preferred pairing between amino acid residue side chains with a 1-4 / 5 sequence position relationship; if there is a preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship in the high entropy increase potential residue connected fragment, or if there is an IVL 1-4 / 5G / A 1-4 / 5IVLF sequence position relationship in the high entropy increase potential residue connected fragment, then an asterisk (*) will be marked below that amino acid.

5. The prediction method according to claim 2, characterized in that: The preferred pairing of amino acid residue side chains in the 1-3 sequence position relationship described in S202 is I / V / L / FI / V / L / F; if, starting from the amino acid in I / V / L / FI / V / L / F, there is a preferred pairing of amino acid residue side chains in the adjacent high entropy increase potential residue connected fragments, then the amino acid in the preferred pairing of amino acid residue side chains in the 1-3 sequence position relationship is not marked as &.

6. The prediction method according to claim 2, characterized in that: As described in S2032, if a high-entropy-increase-potential residue connected fragment in the primary structure of a protein is labeled with &-&, and more than 85% of the amino acid residues in the amino acid side chain region of the primary structure of the protein far from the labeled &-& high-entropy-increase-potential residue connected fragment are labeled with *, and there is a preferred pairing between amino acid residue side chains that have a 1-4 / 5 sequence position relationship with the unrelated residues labeled &-&, then it is predicted that the labeled &-& high-entropy-increase-potential residue connected fragment will form an α-helix; wherein, "far from" means separated by two consecutive micro-entropy residues or corner residues, or separated by at least three hydrophilic residues, or separated by one amino acid residue P; "unrelated residues" refers to other amino acid residues in the primary structure of the protein besides those labeled &-&; and "amino acid side chain region" refers to a region composed of amino acid side chains with more than 5 amino acid residues.

7. The prediction method according to claim 2 or 6, characterized in that: If the high entropy increase potential residue connected fragment described in S2032 is predicted to form a β-sheet, and the high entropy increase potential residue connected fragment has two consecutive micro-entropy residues aspartic acid D or asparagine N, then the two consecutive micro-entropy residues aspartic acid D or asparagine N are predicted to form a turn, as a prediction result.

8. The prediction method according to claim 2, characterized in that: In the entropy increment statistics of β-sheet described in S400, if the same amino acid residue in the primary structure of a protein achieves a preferred pairing between the side chains of two amino acid residues with a 1-3 sequence position relationship, then the entropy increment of that amino acid residue is accumulated twice. In the entropy increment statistics of α-helices, if the same amino acid residue in the primary structure of a protein achieves preferred pairing between the side chains of multiple amino acid residues with 1-4 / 5 sequence positional relationships, then the entropy increment of that amino acid residue is accumulated twice.

9. The prediction method according to claim 2 or 8, characterized in that: In the entropy increment statistics of the α-helix described in S400, if there is a preferred pairing between amino acid residue side chains with a 1-2 sequence position relationship in the primary structure of the protein, and the preferred pairing is I / V / L / MI / V / L / M, then two pairs of I / V / L / M are recorded as one entropy increment.

10. The prediction method according to claim 2, characterized in that: The amino acid residues in two consecutive amino acid residues described in S600 are one or both of S, T, N, D, and G.

11. The prediction method according to claim 2, characterized in that: The relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the side chains of amino acid residues in the primary structure of the protein is the length of the hydrophobic portion on the side chains of amino acid residues in the primary structure of the protein.

12. The prediction method according to claim 2 or 11, characterized in that: The relative measure of the entropy increase potential of the low-entropy hydration layer surrounding the amino acid residues in the primary structure of the protein is specifically defined as the length of a carbon-carbon bond on the amino acid residue chain in the primary structure of the protein as one entropy value.

13. The prediction method according to claim 2, characterized in that: The entropy values ​​of the 20 amino acids are as follows: Serine S=0.5, Threonine T=0.5, Aspartic Acid D=0.5, Asparagine N=0.5, Glycine G=0, Histidine H=3, Arginine R=3, Glutamic Acid E=3, Lysine K=4, Tryptophan W=5, Isoleucine I=6, Valine V=6, Leucine L=6, Phenylalanine F=6, Proline P=0, Cysteine ​​C=4, Glutamine Q=3, Alanine A=1, Methionine M=4, Tyrosine Y=5.

14. The prediction method according to claim 2, characterized in that: The sufficient entropy increase mentioned in step one refers to the lateral adhesion of two amino acid residue side chains when they are similar in the primary structure of a protein, resulting in a sufficient entropy increase in the low-entropy hydration layer of these two side chains. The similarity of the amino acid residue side chains means that the smaller the difference in entropy values ​​between two hydrophilic residues, the more similar the two hydrophilic residues are considered; the smaller the difference in entropy values ​​between two hydrophobic residues, the more similar the two hydrophobic residues are considered.

15. The prediction method according to claim 2, characterized in that: In determining unblocked high-entropy-increase-potential residue connectivity segments, the process first involves identifying segments in the sequence composed of consecutive high-entropy-increase-potential residues, naming these segments as the original "high-entropy-increase-potential residue connectivity segments." Then, non-high-entropy-increase-potential residues that do not block these segments are identified, thus determining the unblocked high-entropy-increase-potential residue connectivity segments. Specifically, this includes: (1) When there is only one amino acid residue S or T between two high entropy increase potential residues, the S or T residues participate in the formation of a high entropy increase potential residue connection segment that is not blocked. (2) When there is only one isolated N, D or G residue between two high entropy potential residues in an amino acid sequence, and there is a high entropy pairing between I, V, L, F, Y across the D, N or G with a 1-3 or 1-4 / 5 relationship, then the D, N or G is considered to participate in the formation of a high entropy potential residue connected segment that is not blocked. (3) When any one of the amino acid residues S, T, N, D, G appears in the amino acid sequence of the polypeptide chain and is adjacent to another amino acid residue of S, T, N, D, G, examine the 1-4 / 5 sequence position relationship across these two adjacent amino acid residues. When both amino acid residues at the 1-4 / 5 position are one of the I, V, L, F, Y residues, and at least one of the amino acid residues is I, V, or L, it is predicted that these two adjacent amino acid residues will participate in the formation of a high entropy increase potential residue connectivity segment that is not blocked. (4) When a P residue appears in the amino acid sequence of the polypeptide chain, and the P residue is not in a 1-2 or 1-3 sequence position relationship with the D or N residue, and there is a 1-4 / 5 sequence position relationship across the P residue, and the amino acid residues on the 1-4 / 5 are I, V, L, F, then the predicted P residue participates in the formation of a high entropy increase potential residue connectivity segment that is not blocked. (5) When a G or A residue appears in the amino acid sequence of a polypeptide chain, if an amino acid residue of type I, V, or L can be found on one side of the sequence containing the G or A residue to form a 1-4 / 5 sequence position relationship with the residue, and an amino acid residue of type I, V, L, or A can also be found on the other side of the sequence containing the G or A residue to form a 1-4 / 5 sequence position relationship with the residue, if there is a single D, N, G, or P residue that is not adjacent to other micro-entropy residues or corner residues among the I, V, and L residues, or if there are two adjacent S, T, N, or D residues, then it is predicted that these single D, N, G, or P residues or two adjacent S, T, N, D, and G residues will participate in forming a high-entropy-increasing potential residue connectivity segment that is not blocked.

16. The prediction method according to claim 15, characterized in that: The length of the unblocked high entropy increase potential residue connectivity fragment is greater than or equal to 3 amino acid residues.

17. The prediction method according to claim 15, characterized in that: If a protein's primary structure is not predicted to be an α-helix or β-turn sequence containing a fragment of 5 or more amino acid residues, and this fragment is in the general initial thermodynamic metastable state of an unfolded protein, then if the height of the hydrophilic atom of any hydrophilic amino acid residue side chain minus the height of the adjacent hydrophobic residue side chain is less than or equal to 1, then this fragment is predicted to be β-sheet. Here, the height of the residue side chain refers to the number of covalent bonds in the shortest path connecting the top atom of the amino acid residue side chain to its main chain carbon atom.

18. A protein structure prediction method based on the mutual attraction between low-entropy hydration layers of residue side chains, characterized in that: The prediction method utilizes hydrotropic forces to predict protein structure; the prediction is based on the protein structure prediction method according to any one of claims 6 to 17, which predicts the secondary structure of the protein based on the mutual attraction between the low-entropy hydration layers of residue side chains, and then performs tertiary structure prediction, including the following steps: The similarity of amino acid residue side chains is defined as follows: the smaller the difference in the "entropy value" between two hydrophilic residue side chains, the more similar the two hydrophilic residues are considered; the smaller the difference in the "entropy value" between two hydrophobic residue side chains, the more similar the two hydrophobic residues are considered. The amino acid residues that have achieved preferred pairing between I / V / L / F in the predicted secondary structure are marked. The hydrophobic surfaces of the hydrophobic side chain clusters that have achieved lateral adhesion between the side chains of I / V / L / F amino acid residues are marked and named "cluster hydrophobic surfaces". The protein structure is predicted to have surface adhesion between one I / V / L / F cluster hydrophobic surface and another adjacent I / V / L / F cluster hydrophobic surface. If there is no preferred pairing relationship between I / V / L / F in a β-sheet or α-helix secondary structure, then search for a preferred pairing relationship between I / V / LY in the secondary structure. If it exists, then the hydrophobic surface of the hydrophobic side chain clusters that are laterally attached between the side chains of I / V / LY amino acid residues is also marked as a "cluster hydrophobic surface". Predict that an I / V / L / Y cluster hydrophobic surface and another adjacent cluster hydrophobic surface will undergo surface attachment to form a protein structure. When an α-helical structure has two cluster hydrophobic surfaces, and the two cluster hydrophobic surfaces are not connected in the axial direction of the α-helix, it is predicted that one of the cluster hydrophobic surfaces with a smaller number of I / V / L / F / Y clusters will not adhere to another I / V / LF / cluster hydrophobic surface in a nearby β-sheet or α-helical secondary structure, as a possible prediction result.

19. The protein structure prediction method based on the mutual attraction between low-entropy hydration layers of residue side chains according to claim 18, characterized in that: It also includes the following steps: For cases where the current prediction is β-sheet, it is necessary to mark the strongly hydrophobic residues in the β-sheet that are preferably paired and connected by the 1-3 positional relationship, i.e., I / V / L / F / W / Y. At the same time, these hydrophobic fragments connected by the 1-3 positional relationship are marked. The lengths of adjacent β-sheets in the sequence are compared. The α-helices with similar predicted lengths will fold into a sheet structure. Then, based on the distribution of the side chains of the strongly hydrophobic residues in the two β-sheets, the two β-sheets are laterally attached in a way that allows the most strongly hydrophobic residues on one β-sheet to achieve the preferred pairing with the amino acid residue side chains on the other α-helix. The state of attachment of the two β-sheets is the predicted sheet fold structure.

Images