A method and system for evaluating target prodrug masking based on docking block model

Abstract

The application discloses a kind of based on butt joint blocking model's targeted prodrug masking evaluation method and system.The method includes the following steps: (1) obtain the contact residue pair set C of targeted drug and target molecule;(2) extract the three-dimensional structure of masking peptide and its cleavable linker peptide segment and targeted drug peptide segment;(3) obtain the blocking amino acid pair set B;(4) calculate the effective distance of blocking vector relative to contact vector, judge the contact amino acid pair is blocked;(5) calculate blocking rate, according to the principle that the greater the blocking rate, the better the masking effect of targeted prodrug is evaluated Masking effect.The system includes: contact residue pair acquisition module, targeted prodrug three-dimensional structure data acquisition module, blocking amino acid pair acquisition module, blocking judgment module and statistical evaluation module.The application calculates the frequency of being masked peptide blocking, can better reflect the masking efficiency or masking multiple of pro-antibody.

Description

Technical Field

[0001] This invention belongs to the field of targeted drug design, and more specifically, relates to a method and system for evaluating the masking of targeted prodrugs based on a docking blocking model. Background Technology

[0002] Targeted therapies, including antibodies, T-cell connectives (TCEs), and cytokines, have become an important class of drugs for the effective treatment of various diseases such as malignant tumors, transplant rejection, autoimmune diseases, and infections due to their precise targeting and long half-life. Despite the continuous development of targeted therapy, many challenges remain for current targeted drug candidates. One major problem is that although targeted therapeutic agents can specifically bind to target molecules, most of these molecules are not only specifically expressed at the lesion site but may also exist in normal cells or tissues outside the lesion site. Therefore, when drugs target and recognize antigens on normal tissues or cells, they can damage normal tissues and produce toxic side effects.

[0003] One possible approach to improving the lesion site selectivity and reducing toxic side effects of therapeutic targeted drugs is to add a masking peptide that inhibits the binding of the targeted drug to the target molecule, linked to a cleavable linker that can be specifically cleaved by one or more proteases at the lesion site. This modified drug is called a targeted prodrug. The cleavable linker of the targeted prodrug is cleaved by certain pathologically specific proteases, releasing the masking peptide. This allows the targeted prodrug to be activated in specific diseased tissues, while in normal tissues or cells it remains a targeted prodrug with the masking peptide, unable to bind to the antigen, thus achieving precise targeting of the lesion site.

[0004] Regarding the evaluation of the masking effect of masking peptides, patent application CN 102482347 B discloses a method for screening targeted prodrugs with high masking efficacy using phage display technology. Similarly, patent CN 112771066 A discloses a polynucleotide library that can be used to target prodrug masking peptides, and uses yeast display technology to screen and / or identify highly efficient masking peptides. Although a certain number of proantibodies already exist, the current phage display / yeast display screening process is slow, inefficient, labor-intensive, and very expensive.

[0005] Screening masking peptides using display techniques is time-consuming and labor-intensive. With advancements in computer hardware and breakthroughs in protein structure prediction, computer-based screening and identification of masking peptides can significantly accelerate the development process and reduce costs. Furthermore, it allows for the screening of specific masking peptides with high masking efficiency based on the characteristics of different binding fragments. Huang et al. developed a method based on the tertiary structure of proteins to calculate the cover rate of the masking peptide over the antibody's CDR region ("Development of a structure-based computational simulation to optimize the blocking efficacy of pro-antibodies." Chemical Science 12.28(2021):9759-9769.). This method first locates the crystal structure of the antibody's Fv region, then uses molecular dynamics homology modeling to simulate the tertiary structure of the pro-antibody. Based on model snapshots, the frequency at which CDR region residues are directly above the masking peptide is calculated; this frequency is the cover rate. The cover rate reflects the frequency at which the CDR region is directly above the masking peptide and is positively correlated with the masking factor of the pro-antibody, serving as a scoring function for screening pro-antibodies. However, CDR regions vary depending on their definition, and different CDR regions contribute differently to antigen binding. Even within the same CDR region, different residues contribute differently to antigen binding. Furthermore, the binding topology between antigen and antibody or antigen-binding fragments is not always a head-to-head relationship. Therefore, coverage rates that only reflect the frequency of CDR regions being directly overhead cannot accurately reflect the masking efficiency or masking factor of the pro-antibody. Summary of the Invention

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a targeted prodrug masking evaluation method and system based on a docking blocking model. The purpose is to evaluate the impact of the masking peptide and its cleavable linker on the original targeted drug's contact with amino acids through computational methods, thereby evaluating the masking effect. This solves the technical problems of existing technologies that require experimental verification, have high time and economic costs, or only provide masking effects from directly above, resulting in incomplete evaluation.

[0007] To achieve the above objectives, according to one aspect of the present invention, a method for evaluating targeted prodrug masking based on a docking blocking model is provided, comprising the following steps:

[0008] (1) Obtain the three-dimensional structural data of the specific binding of the targeted drug with its target molecule and determine the set C of contact residue pairs between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug with its target molecule, there are amino acid residues of the targeted drug with an atomic distance less than a preset threshold and their corresponding amino acid residues of the target molecule; and collect the amino acid residues of the targeted drug in the contact residue pair c to form a set A of amino acid residues to be masked;

[0009] (2) Obtain the three-dimensional structural data of the target prodrug to be evaluated, and extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the target drug peptide.

[0010] (3) For each element in the set A of amino acid residues to be masked, i.e., amino acid residue a to be masked, search for amino acid residues on the masking peptide and its cleavable linker peptide and the target drug peptide obtained in step (2) whose atomic distance is less than the preset threshold in the three-dimensional structure of the masking peptide and its cleavable linker peptide. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as the set B of blocking amino acid pairs.

[0011] (4) For the set of contact amino acid pairs C obtained in step (1) and the set of blocking amino acid pairs B obtained in step (3), the contact vector v is taken as the α carbon atom of the target drug amino acid residue as the start or end point, and the α carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point. The blocking vector W is taken as the α carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point. The effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e', is calculated. If the blocking distance e' of the target drug amino acid residue is greater than 0 and less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked. The blocking distance e' is calculated as follows:

[0012] e' = e / cos(θ)

[0013] Where e is the length of the blocking vector W, and θ is the angle between the contact vector v and the blocking vector W;

[0014] (5) Based on the blocking results of each contact amino acid pair c in the contact amino acid pair set C in step (4), the probability of blocking contact amino acid pair c or the amino acid residue a to be masked is statistically analyzed as the blocking rate; and the masking effect of the targeted prodrug is evaluated according to the principle that the larger the blocking rate, the better the masking effect.

[0015] Preferably, in the targeted prodrug masking evaluation method based on docking blocking model, the preset threshold in steps (1) and (3) is between 3 and 6 angstroms, preferably 4.5 angstroms.

[0016] Preferably, in the targeted prodrug masking evaluation method based on docking blocking model, the three-dimensional structural data of the specific binding of the targeted drug to its target molecule in step (1) are as follows:

[0017] Crystal structures of targeted drugs that specifically bind to their target molecules, as verified by experiments; or

[0018] Molecular three-dimensional structure prediction data.

[0019] Preferably, in the targeted prodrug masking evaluation method based on docking blocking model, the targeted prodrug is a fusion protein constructed by the sequence of the targeted drug, the masking peptide, and the cleavable linker. The three-dimensional structure data of the targeted prodrug in step (2) is preferably the predicted three-dimensional structure data of the protein, and more preferably the three-dimensional structure data of the protein predicted by the sequence of the targeted prodrug fusion protein using a protein folding algorithm.

[0020] Preferably, in the targeted prodrug masking evaluation method based on the docking blocking model, step (5) of which statistically analyzing the probability of contact amino acid c being blocked is used as the blocking rate, specifically involves calculating the blocking rate. Where n is the number of blocked contact amino acid pairs determined in step (4), N is the cardinality of the contact amino acid pair set C, N = card(C);

[0021] The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: Where m is the number of amino acid residues a to be blocked and M is the cardinality of the set A of amino acid residues to be blocked, M = card(A); the amino acid residues a to be blocked are: all contact amino acid residues a that include the amino acid residue a to be blocked are blocked; or at least one contact amino acid residue a that includes the amino acid residue a to be blocked is blocked.

[0022] According to another aspect of the present invention, a targeted prodrug masking evaluation system based on a docking blocking model is provided, comprising: a contact residue pair acquisition module, a three-dimensional structural data acquisition module for the targeted prodrug, a blocking amino acid pair acquisition module, a blocking judgment module, and a statistical evaluation module;

[0023] The contact residue pair acquisition module is used to acquire three-dimensional structural data of the specific binding of the targeted drug and its target molecule and determine the set C of contact residue pairs between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug and its target molecule, there are amino acid residues of the targeted drug and their corresponding amino acid residues of the target molecule with an atomic distance less than a preset threshold; and the amino acid residues of the targeted drug in the contact residue pair c are collected to form a set A of amino acid residues to be masked; the set A of amino acid residues to be masked is submitted to the blocking amino acid pair acquisition module, and the set C of contact residue pairs is submitted to the blocking judgment module;

[0024] The three-dimensional structure data acquisition module for the targeted prodrug is used to acquire the three-dimensional structure data of the targeted prodrug to be evaluated, extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the targeted drug peptide, and submit them to the blocking amino acid pair acquisition module.

[0025] The blocking amino acid pair acquisition module is used to search for amino acid residues on the masking peptide and its cleavable linker peptide with an atomic distance of less than a preset threshold in the three-dimensional structure of the masking peptide and its cleavable linker peptide and the target drug peptide for each element in the set A of amino acid residues to be masked. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as a set B of blocking amino acid pairs. The set B of blocking amino acid pairs is submitted to the blocking judgment module.

[0026] The blocking judgment module is used to determine the blocking vector v for the contact amino acid pair set C and the blocking amino acid pair set B, with the α-carbon atom of the target drug amino acid residue as the start or end point, and the α-carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point, and the α-carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point; calculate the effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e'. If the blocking distance e' of the target drug amino acid residue is greater than 0 and less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked, and the judgment result is submitted to the statistical evaluation module; the blocking distance e' is calculated according to the following method:

[0027] e′=e / cos(θ)

[0028] Where e is the length of the blocking vector W, and θ is the angle between the contact vector v and the blocking vector W.

[0029] The statistical evaluation module is used to calculate the probability of blocking contact amino acid pair c or the amino acid residue a to be masked as the blocking rate based on the blocking judgment results of each contact amino acid pair c in the contact amino acid pair set C; and to evaluate the masking effect of the targeted prodrug according to the principle that the higher the blocking rate, the better the masking effect.

[0030] Preferably, the preset threshold of the targeted prodrug masking evaluation system based on docking blocking model is between 3 and 6 angstroms, and more preferably 4.5 angstroms.

[0031] Preferably, in the targeted prodrug masking evaluation system based on the docking blocking model, the three-dimensional structural data of the specific binding of the targeted drug to its target molecule are as follows:

[0032] Crystal structures of targeted drugs that specifically bind to their target molecules, as verified by experiments; or

[0033] Molecular three-dimensional structure prediction data.

[0034] Preferably, in the targeted prodrug masking evaluation system based on docking blocking model, the targeted prodrug is a fusion protein constructed from the sequence of the targeted drug, the masking peptide, and the cleavable linker. The three-dimensional structure data of the targeted prodrug is preferably predicted protein three-dimensional structure data, more preferably protein three-dimensional structure data predicted based on the sequence of the targeted prodrug fusion protein using a protein folding algorithm.

[0035] Preferably, in the targeted prodrug masking evaluation system based on the docking blocking model, the probability that the contact amino acid c is blocked is used as the blocking rate, specifically by calculating the blocking rate. Where n is the number of blocked contact amino acid pairs determined in step (4), N is the cardinality of the contact amino acid pair set C, N = card(C);

[0036] The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: Where m is the number of amino acid residues a to be blocked and M is the cardinality of the set A of amino acid residues to be blocked, M = card(A); the amino acid residues a to be blocked are: all contact amino acid residues a that include the amino acid residue a to be blocked are blocked; or at least one contact amino acid residue a that includes the amino acid residue a to be blocked is blocked.

[0037] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0038] The blocking rate calculation method disclosed in this invention first uses current tertiary structure prediction tools to predict the tertiary structure of the proantibody, and then uses the crystal structure of the target drug and the target molecule complex to calculate the frequency at which the target drug is blocked by the masking peptide, which can better reflect the masking efficiency or masking fold of the target prodrug. Analysis revealed that the blocking rate and masking fold designed in this invention exhibit a strong positive correlation between different proantibodies, with a Spearman correlation coefficient exceeding 0.8. This allows for direct computer-based screening of proantibodies, significantly improving screening efficiency compared to labor-intensive phage / yeast display technology, saving manpower and experimental resources, and bringing significant advancements to the field of antibody therapy.

[0039] Furthermore, this invention has strong generalization ability and can be applied to the interaction of all amino acid sequences. It not only reflects the masking of the masking peptide directly above the antibody, but also takes into account the diverse and multidirectional ways in which proteins interact. This multidirectional masking situation can also be reflected in the blocking rate used by this method. Therefore, it has a wider range of applications and can more accurately reflect the masking efficiency. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0041] The targeted prodrug masking evaluation method based on docking blocking model provided by this invention includes the following steps:

[0042] (1) Obtain the three-dimensional structural data of the specific binding of the targeted drug with its target molecule and determine the set C of contact residue pairs between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug with its target molecule, there are amino acid residues of the targeted drug with an atomic distance less than a preset threshold and their corresponding amino acid residues of the target molecule; and collect the amino acid residues of the targeted drug in the contact residue pair c to form a set A of amino acid residues to be masked; the preset threshold is between 3 and 6 angstroms, preferably 4.5 angstroms.

[0043] The three-dimensional structural data of the targeted drug specifically binding to its target molecule is preferably the experimentally verified crystal structure data of the targeted drug specifically binding to its target molecule. When experimental data is insufficient, molecular three-dimensional structure prediction data, i.e. molecular interaction prediction data, can also be used as the three-dimensional structural data of the targeted drug specifically binding to its target molecule.

[0044] (2) Obtain the three-dimensional structural data of the target prodrug to be evaluated, and extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the target drug peptide.

[0045] The targeted prodrug is a fusion protein constructed from the sequence of the targeted drug, the masking peptide, and the cleavable linker. The three-dimensional structure data of the targeted prodrug is preferably the predicted three-dimensional structure data of the protein, more preferably the three-dimensional structure data of the protein predicted based on the sequence of the targeted prodrug fusion protein using a protein folding algorithm. The protein folding algorithm includes, but is not limited to, RoseTTAFold, trTrosetta, and Alphafold2, with Alphafold2 being preferred.

[0046] (3) For each element in the set A of amino acid residues to be masked, i.e., amino acid residue a to be masked, search for amino acid residues on the masking peptide and its cleavable linker peptide and the target drug peptide obtained in step (2) whose atomic distance is less than the preset threshold in the three-dimensional structure. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as the set B of blocking amino acid pairs. The preset threshold is the preset threshold for judging the contact residue pair c in step (1), which ensures that all possible blocking amino acid pairs are collected.

[0047] (4) For the set of contact amino acid pairs C obtained in step (1) and the set of blocking amino acid pairs B obtained in step (3), the contact vector v is taken as the α carbon atom of the target drug amino acid residue as the start or end point, and the α carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point. The blocking vector W is taken as the α carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point. The effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e', is calculated. If the blocking distance e' of the target drug amino acid residue is greater than 0 and less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked. The blocking distance e' is calculated as follows:

[0048] e' = e / cos(θ)

[0049] Where e is the length of the blocking vector W, θ is the angle between the contact vector v and the blocking vector W, and e' is less than 0, i.e., cos(θ). That is, when θ is greater than 90°, regardless of the effective distance, it is judged as not being blocked due to the direction.

[0050] (5) Based on the blocking results of each contact amino acid pair c in the contact amino acid pair set C in step (4), the probability of blocking contact amino acid pair c or the amino acid residue a to be masked is statistically analyzed as the blocking rate; and the masking effect of the targeted prodrug is evaluated according to the principle that the larger the blocking rate, the better the masking effect.

[0051] The probability that the statistically contacted amino acid c is blocked is taken as the blocking rate, specifically: calculating the blocking rate. Where n is the number of blocked contact amino acid pairs determined in step (4), N is the cardinality of the contact amino acid pair set C, and N = card(C).

[0052] The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: Where m is the number of amino acid residues a to be blocked and M is the cardinality of the set A of amino acid residues to be blocked, M = card(A). The amino acid residues a to be blocked are: amino acid residues a to be blocked in which all contact amino acid residue pairs including the amino acid residue a to be blocked are blocked; or at least one contact amino acid residue a to be blocked in which at least one contact amino acid residue pair including the amino acid residue a to be blocked is blocked.

[0053] The targeted prodrug masking evaluation system based on docking blocking model provided by the present invention includes: a contact residue pair acquisition module, a three-dimensional structural data acquisition module for targeted prodrugs, a blocking amino acid pair acquisition module, a blocking judgment module, and a statistical evaluation module;

[0054] The contact residue pair acquisition module is used to acquire three-dimensional structural data of the specific binding of the targeted drug and its target molecule and determine the set C of contact residue pairs between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug and its target molecule, there are amino acid residues of the targeted drug and their corresponding amino acid residues of the target molecule with an atomic distance less than a preset threshold; and the amino acid residues of the targeted drug in the contact residue pair c are collected to form a set A of amino acid residues to be masked; the set A of amino acid residues to be masked is submitted to the blocking amino acid pair acquisition module, and the set C of contact residue pairs is submitted to the blocking judgment module;

[0055] The three-dimensional structure data acquisition module for the targeted prodrug is used to acquire the three-dimensional structure data of the targeted prodrug to be evaluated, extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the targeted drug peptide, and submit them to the blocking amino acid pair acquisition module.

[0056] The blocking amino acid pair acquisition module is used to search for amino acid residues on the masking peptide and its cleavable linker peptide with an atomic distance of less than a preset threshold in the three-dimensional structure of the masking peptide and its cleavable linker peptide and the target drug peptide for each element in the set A of amino acid residues to be masked. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as a set B of blocking amino acid pairs. The set B of blocking amino acid pairs is submitted to the blocking judgment module.

[0057] The blocking judgment module is used to determine the blocking vector v for the contact amino acid pair set C and the blocking amino acid pair set B, with the α-carbon atom of the target drug amino acid residue as the start or end point, and the α-carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point, and the α-carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point; calculate the effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e'. If the blocking distance e' of the target drug amino acid residue is greater than 0 and less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked, and the judgment result is submitted to the statistical evaluation module; the blocking distance e' is calculated according to the following method:

[0058] e' = e / cos(θ)

[0059] Where e is the length of the blocking vector W, and θ is the angle between the contact vector v and the blocking vector W.

[0060] The statistical evaluation module is used to calculate the probability of blocking contact amino acid pair c or the amino acid residue a to be masked as the blocking rate based on the blocking judgment results of each contact amino acid pair c in the contact amino acid pair set C; and to evaluate the masking effect of the targeted prodrug according to the principle that the higher the blocking rate, the better the masking effect.

[0061] The following is an example:

[0062] Example 1

[0063] We collected publicly available pre-antibody data. The article [Huang, Bo-Cheng, et al. "Development of a structure-based computational simulation to optimize the blocking efficacy of pro-antibodies." Chemical Science 12.28(2021):9759-9769.] disclosed nine different pre-antibodies and their masking folds. We also collected data from five pre-antibody articles [Trang, Vivian H., et al. "A coiled-coil masking domain for selective activation of therapeutic antibodies." Nature biotechnology 37.7(2019):761-765.]. After removing those without publicly disclosed antigen-antibody crystal structures, we had pre-antibody data for rituximab and trastuzumab. The results are shown in Table 1.

[0064] Table 1

[0065]

[0066]

[0067] The targeted prodrug masking evaluation method based on the docking blocking model in this embodiment includes the following steps:

[0068] The three-dimensional structural data of the specific binding of the targeted drug with its target molecule were obtained, and the set C of contact residue pairs between the targeted drug and the target molecule was determined. The three-dimensional structural data of the specific binding of the targeted drug with its target molecule was obtained from the SabDab database as the crystal structure of the antigen-antibody Fv region complex.

[0069] The contact residue pair c refers to the three-dimensional structural data of the specific binding of the targeted drug and its target molecule, in which there is an amino acid residue of the targeted drug with an atomic distance of less than 4.5 angstroms and its corresponding amino acid residue of the target molecule.

[0070] Table 2: Set of Ipilimumab-target molecule contact residue pairs C

[0071] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule L32SER C108ILE L33TYR C108ILE L33TYR C107GLY L33TYR C106LEU L33TYR C3MET L92TYR C106LEU L93GLY C106LEU L93GLY C105TYR L93GLY C104TYR L94SER C105TYR L94SER C104TYR L95SER C103PRO L95SER C104TYR L97TRP C104TYR H33THR C97GLU H50PHE C104TYR H52SER C99MET H53TYR C97GLU H53TYR C95LYS H53TYR C35ARG H53TYR C99MET H57ASN C100TYR H57ASN C99MET H57ASN C33GLU H59TYR C103PRO H59TYR C101PRO H59TYR C104TYR H59TYR C102PRO H59TYR C100TYR H59TYR C99MET H101TRP C39LEU H101TRP C48GLU H101TRP C95LYS H101TRP C46VAL H101TRP C47THR H102LEU C39LEU H102LEU C46VAL H102LEU C108ILE H102LEU C41GLN H102LEU C106LEU H103GLY C106LEU

[0072] Table 3: Set of Nivolumab-target molecule contact residue pairs C

[0073] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule L32TYR G26ASP L34ALA G131LYS L46LEU G130PRO L46LEU G131LYS L49TYR G130PRO L49TYR G131LYS L49TYR G132ALA L55ALA G130PRO L56THR G130PRO L56THR G128LEU L56THR G129ALA L56THR G132ALA H28THR G59THR H28THR G60SER H30SER G30ARG H31ASN G59THR H31ASN G30ARG H31ASN G60SER H31ASN G29ASP H32SER G28PRO H32SER G29ASP H33GLY G28PRO H33GLY G29ASP H50VAL G28PRO H52TRP G26ASP H52TRP G27SER H52TRP G25LEU H52TRP G28PRO H53TYR G30ARG H53TYR G31PRO H53TYR G28PRO H53TYR G29ASP H57LYS G25LEU H99ASN G131LYS H99ASN G28PRO H99ASN G29ASP H100ASP G130PRO H100ASP G131LYS H100ASP G129ALA H100ASP G29ASP H101ASP G130PRO H101ASP G131LYS H102TYR G130PRO H102TYR G128LEU H102TYR G129ALA

[0074] Table 4: Set of residue pairs that contact the target molecule with Canakinumab C

[0075] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule L28SER I87PRO L28SER I86ASP L28SER I88LYS L30GLY I87PRO L32SER I65LYS L32SER I64GLU L50TYR I64GLU L91SER I65LYS L91SER I64GLU L92SER I87PRO L92SER I66ASN L92SER I65LYS L92SER I64GLU L93SER I66ASN L94LEU I23PRO L96PHE I65LYS H27THR I35ASP H27THR I34GLN H27THR I32GLN H30VAL I31LEU H30VAL I38GLN H30VAL I29LEU H30VAL I35ASP H31TYR I38GLN H31TYR I37GLU H31TYR I35ASP H31TYR I34GLN H51TRP I23PRO H51TRP I22GLY H51TRP I21SER H51TRP I20MET H52TYR I29LEU H52TYR I27LYS H52TYR I129ASN H52TYR I21SER H52TYR I20MET H52TYR I19VAL H53ASP I21SER H53ASP I27LYS H53ASP I129ASN H55ASP I27LYS H56ASN I23PRO H56ASN I22GLY H56ASN I21SER H58TYR I23PRO H97ARG I37GLU H98ASP I38GLN H99LEU I38GLN H99LEU I37GLU H100ARG I39GLN H100ARG I38GLN H100ARG I37GLU H100ARG I65LYS H100ARG I41VAL H100ARG I64GLU H101THR I20MET H101THR I65LYS

[0076] Table 5: Set of residue pairs that contact Adalimumab with the target molecule (C)

[0077]

[0078]

[0079] Table 6: Set of residue pairs that Rituximab contacts with the target molecule C

[0080] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule L90TRP P170ALA L90TRP P171ASN L93ASN P168GLU L93ASN P169PRO L93ASN P170ALA L95PRO P170ALA H33ASN P172PRO H33ASN P173SER H33ASN P171ASN H35HIS P170ALA H35HIS P171ASN H47TRP P170ALA H50ALA P172PRO H50ALA P170ALA H51ILE P172PRO H52TYR P172PRO H52TYR P173SER H55ASN P176ASN H57ASP P172PRO H57ASP P175LYS H57ASP P176ASN H58THR P172PRO H58THR P175LYS H59SER P172PRO H59SER P168GLU H59SER P169PRO H59SER P170ALA H99SER P171ASN H102TYR P174GLU H106TRP P173SER H106TRP P174GLU H106TRP P171ASN

[0081] Table 7: Set of Trastuzumab-target molecule contact residue pairs C

[0082] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule A28ASP C598GLU A30ASN C598GLU A30ASN C596ASP A30ASN C602GLN A30ASN C600ALA A31THR C602GLN A32ALA C571PRO A50SER C603PRO A53PHE C605PRO A53PHE C604CYS A53PHE C603PRO A91HIS C572PRO A91HIS C571PRO A92TYR C572PRO A92TYR C571PRO A92TYR C569LYS A93THR C572PRO A94THR C572PRO A94THR C560ASP B33TYR C558GLU B33TYR C573PHE B50ARG C558GLU B50ARG C573PHE B50ARG C560ASP B52TYR C558GLU B57TYR C558GLU B57TYR C557PRO B58THR C558GLU B59ARG C558GLU B59ARG C561GLN B59ARG C560ASP B99TRP C573PHE B99TRP C572PRO B103GLY C591ILE B103GLY C579PRO B103GLY C570ASP B103GLY C593LYS B104PHE C593LYS B105TYR C573PHE B105TYR C572PRO B105TYR C571PRO B105TYR C570ASP

[0083] Collect the target drug amino acid residues in contact residue pair c to form a set A of amino acid residues to be masked.

[0084] (2) The tertiary structure of the proantibody predicted by Alphafold 2.1.0 was used as the three-dimensional structural data of the target prodrug fusion protein to be evaluated, and its masking peptide and the three-dimensional structure of its cleavable linker peptide and the target drug peptide were extracted.

[0085] (3) For each element in the set A of amino acid residues to be masked, search for amino acid residues on the masking peptide and its cleavable linker peptide and the target drug peptide obtained in step (2) whose atomic distance is less than 4.5 Å. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as the set B of blocking amino acid pairs.

[0086] Table 8: L1-ipi blocking amino acid residue pair set B

[0087]

[0088] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0089] Table 9: L2-ipi blocking amino acid residue pairs set B

[0090]

[0091]

[0092] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.) Table 10: L2-niv blocking amino acid residue pair set B

[0093]

[0094] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.) Table 11: L3-niv blocking amino acid residue pair set B

[0095]

[0096] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.) Table 12: L3-can blocking amino acid residue pairs set B

[0097]

[0098]

[0099] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.) Table 13: L4-can blocking amino acid residue pairs set B

[0100]

[0101] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.) Table 14: L5-ada blocking amino acid residue pair set B

[0102]

[0103] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0104] Table 15: L6-ada blocking amino acid residue pairs set B

[0105]

[0106] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0107] Table 16: L7-ada blocking amino acid residue pairs set B

[0108]

[0109] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0110] Table 17: L8-rit blocking amino acid residue pairs set B

[0111]

[0112] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0113] Table 17: L9-tra blocking amino acid residue pairs set B

[0114]

[0115] (Note: To correspond with the numbering of targeted drugs, the targeted drug portion of the prodrug has been renumbered.)

[0116] (4) For the set of contact amino acid pairs C obtained in step (1) and the set of blocking amino acid pairs B obtained in step (3), the contact vector v is taken as the α carbon atom of the target drug amino acid residue as the start or end point, and the α carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point. The blocking vector W is taken as the α carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point. The effective distance of the blocking vector relative to the contact vector is calculated, i.e., the blocking distance e'. If the blocking distance e' of the target drug amino acid residue is less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked. The blocking distance e' is calculated as follows:

[0117] e' = e / cos(θ)

[0118] Where e is the length of the blocking vector W, and θ is the angle between the contact vector v and the blocking vector W.

[0119] (5) For the set of contact amino acid pairs C, the probability of the amino acid residue a to be masked being blocked is used as the blocking rate; and the masking effect of the targeted prodrug is evaluated according to the principle that the larger the blocking rate, the better the masking effect.

[0120] The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: Where m is the number of amino acid residues a to be blocked and masked, M is the cardinality of the set A of amino acid residues to be masked, and M = card(A). The amino acid residues a to be blocked and masked are:

[0121] At least one contact amino acid residue including the amino acid residue to be masked a is blocked from the amino acid residue to be masked a.

[0122] The results of the blocking rate calculation are shown in Table 1.

[0123] By converting the blocking rate and masking factor into rank data, the Spearman correlation coefficient between the blocking rate and the masking factor was calculated to be 0.845, indicating a strong positive correlation between the blocking rate and the masking factor, which can be used as an evaluation standard for the masking efficiency of the pre-antibody.

[0124] Example 2

[0125] The article [Mansurov A, Hosseinchi P, Chang K, et al. Masking the immunotoxicity of interleukin-12 by fusing it with a domain of its receptor via a tumor-protease-cleavable linker[J]. Nature Biomedical Engineering, 2022:1-11.] publicly disclosed a masked cytokine that uses IL-12Rβ1 as a masking peptide, allowing IL-12 to be specifically activated in tumors. We have compiled three targeted prodrugs with masking efficiency, as shown in Table 8:

[0126] Table 18

[0127]

[0128]

[0129] The targeted prodrug masking evaluation method based on the docking blocking model in this embodiment includes the following steps:

[0130] (1) The cytokine in the article is murine IL-12, but the PDB does not contain the crystal structure of the complex between murine IL-12 and its target molecule. Alphafold 2.1 was preferred to predict the three-dimensional structural data of the specific binding of the targeted drug to its target molecule and to determine the set C of contact residue pairs between the targeted drug and the target molecule; the three-dimensional structural data of the specific binding of the targeted drug to its target molecule was obtained from the SabDab database as the crystal structure of the antigen-antibody Fv region complex.

[0131] The contact residue pair c refers to the three-dimensional structural data of the specific binding of the targeted drug and its target molecule, in which there is an amino acid residue of the targeted drug with an atomic distance of less than 4.5 angstroms and its corresponding amino acid residue of the target molecule.

[0132] Table 19: Set of IL-12 contact residue pairs with target molecules C

[0133] Targeted drugs Target molecule Targeted drugs Target molecule Targeted drugs Target molecule B13VAL C71LEU B14ASP C71LEU B15TRP C9ARG B15TRP C71LEU B15TRP C97TYR B15TRP C15TYR B15TRP C64GLU B15TRP C70VAL B16THR C97TYR B16THR C71LEU B17PRO C95TYR B17PRO C71LEU B17PRO C97TYR B17PRO C98ASN B17PRO C70VAL B21GLY C40ASN B42GLN C50ARG B42GLN C48ARG B43ARG C41HIS B43ARG C40ASN B43ARG C48ARG B43ARG C39PRO B43ARG C38PRO B44HIS C48ARG B46VAL C43HIS B47ILE C41HIS B47ILE C42THR B47ILE C43HIS B48GLY C42THR B55ILE C40ASN B56THR C40ASN B57VAL C40ASN B58LYS C39PRO B58LYS C71LEU B58LYS C40ASN B59GLU C72SER B59GLU C39PRO B59GLU C71LEU B59GLU C69PRO B59GLU C70VAL B60PHE C65GLN B60PHE C64GLU B61LEU C36PHE B61LEU C68ILE B61LEU C50ARG B61LEU C48ARG B61LEU C39PRO B61LEU C69PRO B61LEU C38PRO B61LEU C67GLY B62ASP C39PRO B62ASP C40ASN B84LYS C12LYS B84LYS C9ARG B84LYS C64GLU B84LYS C13THR B86GLU C12LYS B86GLU C97TYR B86GLU C9ARG B91SER C12LYS B92THR C12LYS B93GLU C12LYS B93GLU C9ARG B93GLU C64GLU B93GLU C13THR B101LYS C164GLY B101LYS C165SER B101LYS C163ARG B123ARG C63TRP B123ARG C13THR B123ARG C14ASP B192GLN C65GLN B192GLN C63TRP B193GLN C65GLN B193GLN C63TRP B194ASN C65GLN B195LYS C65GLN B195LYS C63TRP B195LYS C64GLU B195LYS C13THR

[0134] Collect the target drug amino acid residues in contact residue pair c to form a set of amino acid residues to be masked, A.

[0135] (2) Using Alphafold 2.1.0 to predict the tertiary structure of the targeted prodrug as the three-dimensional structural data of the targeted prodrug fusion protein to be evaluated, the three-dimensional structures of its masking peptide and its cleavable linker peptide and the targeted drug peptide were extracted.

[0136] (3) For each element in the set A of amino acid residues to be masked, search for amino acid residues on the masking peptide and its cleavable linker peptide and the target drug peptide obtained in step (2) whose atomic distance is less than 4.5 Å. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as the set B of blocking amino acid pairs.

[0137] Table 20: L2-IL12 Blocked Amino Acid Residue Pair Set B

[0138]

[0139] Table 21: L3-IL12 Blocked Amino Acid Residue Pair Set B

[0140]

[0141]

[0142] Table 22: Lnc-IL12 Blocked Amino Acid Residue Pairs Set B

[0143]

[0144]

[0145] (4) For the set of contact amino acid pairs C obtained in step (1) and the set of blocking amino acid pairs B obtained in step (3), the contact vector v is taken as the α carbon atom of the target drug amino acid residue as the start or end point, and the α carbon atom of the target molecule residue of the contact residue pair as the corresponding end or start point. The blocking vector W is taken as the α carbon atom of the masking peptide and the cleavable linker peptide residue of the blocking residue pair as the corresponding end or start point. The effective distance of the blocking vector relative to the contact vector is calculated, i.e., the blocking distance e'. If the blocking distance e' of the target drug amino acid residue is less than the length d of its contact vector, then the contact amino acid pair is determined to be blocked. The blocking distance e' is calculated as follows:

[0146] e' = e / cos(θ)

[0147] Where e is the length of the blocking vector W, and θ is the angle between the contact vector v and the blocking vector W.

[0148] (5) Based on the blocking judgment results of each contact amino acid pair c in the contact amino acid pair set C in step (4), the probability of contact amino acid pair c being blocked is statistically analyzed as the blocking rate; and the masking effect of the targeted prodrug is evaluated according to the principle that the larger the blocking rate, the better the masking effect.

[0149] The probability that the statistically contacted amino acid c is blocked is taken as the blocking rate, specifically: calculating the blocking rate. Where n is the number of blocked contact amino acid pairs determined in step (4), N is the cardinality of the contact amino acid pair set C, and N = card(C).

[0150] The masking effect of the targeted prodrug was evaluated according to the principle that the higher the blocking rate, the better the masking effect. The results of the blocking rate calculation are shown in Table 8.

[0151] The Spearman correlation coefficient between the blocking rate and the masking factor was calculated to be 1, indicating a strong positive correlation between the blocking rate and the masking factor.

[0152] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating targeted prodrug masking based on a docking blocking model, characterized in that, Includes the following steps: (1) Obtain the three-dimensional structural data of the specific binding of the targeted drug with its target molecule and determine the set C of contact residue pairs c between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug with its target molecule, there are amino acid residues of the targeted drug with an atomic distance less than a preset threshold and their corresponding amino acid residues of the target molecule; and collect the amino acid residues of the targeted drug in the contact residue pair c to form a set A of amino acid residues a to be masked; (2) Obtain the three-dimensional structural data of the target prodrug to be evaluated, and extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the target drug peptide. (3) For each element in set A, namely the amino acid residue a to be masked, search for amino acid residues on the masking peptide and its cleavable linker peptide and the target drug peptide obtained in step (2) whose atomic distance is less than the preset threshold. These amino acid residues are used as blocking amino acids of the target drug amino acid residues of the contact residue pair c. The target drug amino acid residue and its blocking amino acid are used as blocking amino acid pair b. All blocking amino acid pairs b are collected as the set B of blocking amino acid pairs. (4) For set C obtained in step (1) and set B obtained in step (3), take the α carbon atom of the amino acid residue of the target drug as the starting point or ending point, and the α carbon atom of the target molecule residue of its contact residue pair c as the corresponding ending point or starting point as the contact vector v, and take the α carbon atom of the masking peptide of the amino acid pair b and its cleavable linker peptide residue as the corresponding ending point or starting point as the blocking vector W; calculate the effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e'. If the blocking distance e' of the amino acid residue of the target drug is greater than 0 and less than the length d of its contact vector, then it is determined that the contact residue pair c is blocked; the blocking distance e' is calculated as follows: ； in, Let W be the length of the blocking vector. Let v be the angle between the contact vector v and the blocking vector W. (5) Based on the blocking results of each contact residue in set C in step (4), the probability that the contact residue is blocked or the amino acid residue a to be masked is used as the blocking rate; and the masking effect of the targeted prodrug is evaluated according to the principle that the larger the blocking rate, the better the masking effect.

2. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 1, characterized in that, The preset thresholds in steps (1) and (3) are between 3 and 6 angstroms.

3. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 2, characterized in that, The preset threshold in steps (1) and (3) is 4.5 angstroms.

4. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 1, characterized in that, The three-dimensional structural data of the targeted drug and its specific binding to the target molecule in step (1) are as follows: Crystal structures of targeted drugs that specifically bind to their target molecules, as verified by experiments; or Molecular three-dimensional structure prediction data.

5. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 1, characterized in that, The targeted prodrug is a fusion protein constructed from the sequence of the targeted drug, the masking peptide, and the cleavable linker. The three-dimensional structural data of the targeted prodrug in step (2) is the predicted three-dimensional structural data of the protein.

6. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 5, characterized in that, The predicted protein three-dimensional structure data is obtained by using a protein folding algorithm to predict the protein three-dimensional structure data based on the sequence of the targeted prodrug fusion protein.

7. The targeted prodrug masking evaluation method based on docking blocking model as described in claim 1, characterized in that, Step (5) Calculate the probability that the contact residues block c as the blocking rate. Specifically, calculate the blocking rate. ,in N represents the number of blocked contact residue pairs c determined in step (4), where N is the cardinality of set C. ; The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: ,in M represents the number of amino acid residues a to be blocked and masked, where M is the cardinality of the set A of amino acid residues to be masked. The blocked amino acid residue a is: the amino acid residue a whose contact amino acid residue pair c is blocked; or at least one amino acid residue a whose contact amino acid residue pair c is blocked.

8. A targeted prodrug masking evaluation system based on a docking blocking model, characterized in that, include: The system includes modules for acquiring contact residue pairs, acquiring three-dimensional structural data of targeted prodrugs, acquiring blocking amino acid pairs, determining blocking effects, and statistical evaluation. The contact residue pair acquisition module is used to acquire three-dimensional structural data of the specific binding of the targeted drug and its target molecule and determine the set C of contact residue pairs c between the targeted drug and the target molecule; the contact residue pair c is: in the three-dimensional structural data of the specific binding of the targeted drug and its target molecule, there are amino acid residues of the targeted drug and their corresponding amino acid residues of the target molecule with an atomic distance less than a preset threshold; and the amino acid residues of the targeted drug in the contact residue pair c are collected to form a set A of amino acid residues a to be masked; set A is submitted to the blocking amino acid pair acquisition module, and set C is submitted to the blocking judgment module; The three-dimensional structure data acquisition module for the targeted prodrug is used to acquire the three-dimensional structure data of the targeted prodrug to be evaluated, extract the three-dimensional structure of its masking peptide and its cleavable linker peptide and the targeted drug peptide, and submit them to the blocking amino acid pair acquisition module. The blocking amino acid pair acquisition module is used to search for amino acid residues on the masking peptide and its cleavable linker peptide with an atomic distance less than a preset threshold in the three-dimensional structure of the masking peptide and its cleavable linker peptide and the target drug peptide for each element in set A, namely the amino acid residue to be masked a. This amino acid residue and its blocking amino acid are used as blocking amino acid pairs b. All blocking amino acid pairs b are collected as a set B of blocking amino acid pairs b. The set B of blocking amino acid pairs is submitted to the blocking judgment module. The blocking judgment module is used to, for sets C and B, take the α-carbon atom of the amino acid residue of the target drug as the starting or ending point, and the α-carbon atom of the target molecule residue of its contact residue as the corresponding ending or starting point as the contact vector v, and the α-carbon atom of the masking peptide and its cleavable linker peptide residue of its blocking residue pair as the corresponding ending or starting point as the blocking vector W; calculate the effective distance of the blocking vector relative to the contact vector, i.e., the blocking distance e'. If the blocking distance e' of the amino acid residue of the target drug is greater than 0 and less than the length d of its contact vector, then it is determined that the contact residue pair c is blocked, and the judgment result is submitted to the statistical evaluation module; the blocking distance e' is calculated according to the following method: ； in, Let W be the length of the blocking vector. Let v be the angle between the contact vector v and the blocking vector W. The statistical evaluation module is used to calculate the probability that the contact residues block c or the amino acid residue a to be masked is blocked based on the blocking judgment results of each contact residue in set C; and to evaluate the masking effect of the targeted prodrug according to the principle that the larger the blocking rate, the better the masking effect.

9. The targeted prodrug masking evaluation system based on the docking blocking model as described in claim 8, characterized in that, The preset threshold is between 3 and 6 angstroms.

10. The targeted prodrug masking evaluation system based on the docking blocking model as described in claim 9, characterized in that, The preset threshold in steps (1) and (3) is 4.5 angstroms.

11. The targeted prodrug masking evaluation system based on docking blocking model as described in claim 8, characterized in that, The three-dimensional structural data of the targeted drug specifically binding to its target molecule are as follows: Crystal structures of targeted drugs that specifically bind to their target molecules, as verified by experiments; or Molecular three-dimensional structure prediction data.

12. The targeted prodrug masking evaluation system based on the docking blocking model as described in claim 8, characterized in that, The targeted prodrug is a fusion protein constructed from the sequence of the targeted drug, the masking peptide, and the cleavable linker. The three-dimensional structural data of the targeted prodrug is the predicted three-dimensional structural data of the protein.

13. The targeted prodrug masking evaluation system based on the docking blocking model as described in claim 8, characterized in that, The predicted protein three-dimensional structure data is obtained by using a protein folding algorithm to predict the protein three-dimensional structure data based on the sequence of the targeted prodrug fusion protein.

14. The targeted prodrug masking evaluation system based on the docking blocking model as described in claim 8, characterized in that, The probability that contact residues block c is used as the blocking rate. Specifically, the blocking rate is calculated as follows: ,in N represents the number of blocked contact residue pairs c determined in step (4), where N is the cardinality of set C. ; The probability that the amino acid residue a to be masked is blocked is taken as the blocking rate. Specifically, the blocking rate is calculated as follows: ,in M represents the number of amino acid residues a to be blocked and masked, where M is the cardinality of the set A of amino acid residues to be masked. The blocked amino acid residue a is: the amino acid residue a whose contact residue pair c including the amino acid residue a is blocked; or at least one amino acid residue a whose contact amino acid residue pair including the amino acid residue a is blocked.

Images