A screening method for viral drug targets and its application

Abstract

This invention discloses a method for screening viral drug targets and its application. The screening method includes obtaining genomic sequences of different lineages of evolutionary mutations in monkeypox virus, constructing a VirusEvo pedigree model and obtaining gene feature embeddings related to evolutionary mutations; obtaining viral gene sequences, converting the gene sequences into corresponding protein sequences, and extracting the lineage embedding representation of each protein sequence in the mutation lineage; predicting the set of genes most relevant to the mutation lineage at the whole genome level, and sorting gene importance according to the activation weight of each spline function in fastKAN to obtain the final set of genes most relevant to the lineage mutation, and selecting genes with a dN / dS ratio > 1 as viral drug targets. The screening method of this invention can quickly identify potential drug target proteins of viruses in about one week, and use them in antiviral drug screening to quickly screen for effective drugs targeting these targets.

Description

Technical Field

[0001] This invention relates to the field of biotechnology, specifically to a method for screening viral drug targets and its application. Background Technology

[0002] According to the theory of molecular evolution, viral epidemics are the result of viruses adaptively spreading more widely within a population (environment), representing adaptive evolution. In this evolutionary process, two types of genes play crucial roles: genes that promote viral transmission and genes that maintain normal function. The proteins encoded by these two types of genes have important functions in enhancing viral infectivity and spread.

[0003] Viruses are highly contagious and can easily cause large-scale infections, making it extremely urgent to find antiviral drugs to stop their spread. Traditional laboratory methods, involving the design and development of drugs from scratch, are time-consuming and costly. The lengthy process and high failure rate make it extremely difficult to obtain effective new antiviral drugs in a short period. However, repurposing existing approved drugs can significantly shorten the time required for antiviral drug development.

[0004] In view of this, the present invention is hereby proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method for screening viral drug targets and its application. The screening method of this invention can quickly identify potential drug target proteins of viruses in about one week and use them in the screening of antiviral drugs, which can quickly screen out effective drugs against these targets.

[0006] In order to achieve the above-mentioned objectives of the present invention, the following technical solution is adopted:

[0007] The first aspect of this invention provides a method for screening viral drug targets, the method comprising the following steps:

[0008] (a) Obtain the genome sequences of different lineages of monkeypox virus evolution mutations, encode the three-mode characters into digital tokens using a 3-mode character encoding mechanism, and then convert the genome sequence encoding into a 768-dimensional sequence embedding through the token embedding layer. The gene sequence feature extraction network is designed with a 12-layer transformer architecture as the embedding base model. The pre-training uses MLM mask loss for feature association embedding, and the fine-tuning stage uses lineage evolution labels as the target for training. Finally, the gene feature embeddings related to evolution mutations are obtained.

[0009] (b) Obtain the viral gene sequence and convert it into the corresponding protein sequence. Use the ESM2.0 protein sequence big language model to learn the amino acid sequence embedding representation and extract the lineage embedding representation of each protein sequence in the mutation lineage. In the embedding pedigree model, introduce the solvable fastKAN deep learning framework to combine the gene feature embedding related to evolutionary mutations and the lineage embedding representation of each protein sequence. Predict the set of genes most related to the mutation lineage at the whole genome level, and sort the genes by importance according to the activation weight of each spline function in fastKAN to obtain the final set of genes most related to the lineage mutation.

[0010] (d) Calculate the dN / dS ratio of each gene in the final gene set, and select genes with a dN / dS ratio > 1 as viral drug targets.

[0011] Preferably, the virus includes monkeypox virus.

[0012] A second aspect of the present invention provides an application of the above-described method for screening viral drug targets in antiviral drug screening.

[0013] A third aspect of the present invention provides a method for screening antiviral drugs, the method comprising the following steps:

[0014] (1) Viral drug targets were screened according to the screening method for viral drug targets described above;

[0015] (2) AlphaFold3 was used to screen for drugs that can bind to viral drug targets.

[0016] Preferably, in step (2), the drug is derived from Bioactive compounds in the Enamine database.

[0017] Preferably, in step (2), AlphaFold3 is used to screen for drugs that bind to the top 5 to 10 viral drug targets.

[0018] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0019] This invention's screening method can rapidly identify potential drug target proteins for monkeypox virus within approximately one week, and then combine this with antiviral drug screening methods to identify effective drugs targeting these proteins. Specifically, target protein screening can be rapidly completed using the self-developed large-scale artificial intelligence model VirusEvo, while drug screening is achieved using the open-source AI tool AlphaFold3. The key advantage of this invention's screening method is its ability to rapidly screen suitable drugs from already marketed medications. It is not only highly efficient and reliable with no additional development costs (lower price), but also ensures that the drug safety has been clinically validated (no new side effect risks). It is expected to find targeted and effective drugs for the vast majority of viruses within approximately one week. Detailed Implementation

[0020] The embodiments of the technical solution of the present invention will be described in detail below with reference to the examples. The following embodiments are only used to illustrate the technical solution of the present invention more clearly, and are therefore only examples, and should not be used to limit the scope of protection of the present invention.

[0021] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application should have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0022] This invention provides a method for screening viral drug targets, the method comprising the following steps:

[0023] (a) Obtain the genome sequences of different lineages of monkeypox virus evolution mutations, encode the three-mode characters into digital tokens using a 3-mode character encoding mechanism, and then convert the genome sequence encoding into a 768-dimensional sequence embedding through the token embedding layer. The gene sequence feature extraction network is designed with a 12-layer transformer architecture as the embedding base model. The pre-training uses MLM mask loss for feature association embedding, and the fine-tuning stage uses lineage evolution labels as the target for training. Finally, the gene feature embeddings related to evolution mutations are obtained.

[0024] (b) Obtain the viral gene sequence and convert it into the corresponding protein sequence. Use the ESM2.0 protein sequence big language model to learn the amino acid sequence embedding representation and extract the lineage embedding representation of each protein sequence in the mutation lineage. In the embedding pedigree model, introduce the solvable fastKAN deep learning framework to combine the gene feature embedding related to evolutionary mutations and the lineage embedding representation of each protein sequence. Predict the set of genes most related to the mutation lineage at the whole genome level, and sort the genes by importance according to the activation weight of each spline function in fastKAN to obtain the final set of genes most related to the lineage mutation.

[0025] (d) Calculate the dN / dS ratio of each gene in the final gene set, and select genes with a dN / dS ratio > 1 as viral drug targets.

[0026] In one embodiment, the virus includes monkeypox virus.

[0027] Another embodiment of the present invention provides an application of the above-described method for screening viral drug targets in antiviral drug screening.

[0028] Another embodiment of the present invention provides a method for screening antiviral drugs, the method comprising the following steps:

[0029] (1) Viral drug targets were screened according to the screening method for viral drug targets described above;

[0030] (2) AlphaFold3 was used to screen for drugs that can bind to viral drug targets.

[0031] In one embodiment, in step (2), the drug is derived from bioactive compounds in the Enamine database.

[0032] In one embodiment, in step (2), AlphaFold3 is used to screen for drugs that bind to the top 5 to 10 viral drug targets.

[0033] The technical solution of the present invention will be further described in detail below through specific embodiments.

[0034] Example 1

[0035] This embodiment provides a method for screening viral drug targets, which includes the following steps:

[0036] (a) Obtain the genome sequences of different lineages of monkeypox virus evolution mutations, encode the three-mode characters into digital tokens using a 3-mode character encoding mechanism, and then convert the genome sequence encoding into a 768-dimensional sequence embedding through the token embedding layer. The gene sequence feature extraction network is designed with a 12-layer transformer architecture as the embedding base model. The pre-training uses MLM mask loss for feature association embedding, and the fine-tuning stage uses lineage evolution labels as the target for training. Finally, the gene feature embeddings related to evolution mutations are obtained.

[0037] (b) Obtain the viral gene sequence and convert it into the corresponding protein sequence. Use the ESM2.0 protein sequence big language model to learn the amino acid sequence embedding representation and extract the lineage embedding representation of each protein sequence in the mutation lineage. In the embedding pedigree model, introduce the solvable fastKAN deep learning framework to combine the gene feature embedding related to evolutionary mutations and the lineage embedding representation of each protein sequence. Predict the set of genes most related to the mutation lineage at the whole genome level, and sort the genes by importance according to the activation weight of each spline function in fastKAN to obtain the final set of genes most related to the lineage mutation.

[0038] (d) Calculate the dN / dS ratio of each gene in the final gene set, and select genes with a dN / dS ratio > 1 as viral drug targets. The viral drug targets are OPG198, OPG027, OPG199, OPG120, OPG154, OPG174, OPG092, OPG173, OPG200 and OPG112.

[0039] Example 2

[0040] This embodiment is a method for screening antiviral drugs, which includes the following steps:

[0041] (1) Obtain the SMILE format files of the drug compounds from the Bioactive compounds database (https: / / enamine.net / compound-libraries / bioactive-libraries);

[0042] (2) Input the smiles format file of the above drug compound into the json input file of the AlphaFold3 model. After running, the drugs that combine different viral drug targets in Example 1 are obtained and the top 10 drugs are selected.

[0043] Example 3

[0044] This example is an experimental verification of the antiviral effect of the drug:

[0045] To verify the physical binding of target proteins and antiviral drugs, four proteins—OPG027, OPG174, OPG198, and OPG112—were selected for expression and purification. Surface plasmon resonance (SPR) experiments were then performed to bind these proteins to seven screened drugs. The results are shown in Table 1.

[0046] Table 1

[0047] Drug General Kinetics model Channel Immobilized ligand Quality Kinetics Chi² (RU²) 1:1 binding ka (1 / Ms) kd (1 / s) KD (M) EBC-08787 1:1 binding 3 50μg / mL OPG112 pH4.0 1.25e0 1.76e+3 3.54e-2 2.01e-5 Vorapaxar Steady state affinity 3 50μg / mL OPG112 pH4.0 N / A N / A N / A 4.59e-5 EBC-07547 Steady state affinity 4 50μg / mL OPG112 pH4.0 N / A N / A N / A 1.86e-5 EBC-303150 Steady state affinity 4 50μg / mL OPG112 pH4.0 N / A N / A N / A 2.00e-5 EBC-07547 Steady state affinity 5 50μg / mL OPG027 pH4.0 5.74e0 6.94e+2 1.67e-1 1.81e-5 Vorapaxar Steady state affinity 5 50μg / mL OPG027 pH4.0 6.81e0 9.74e+1 2.21e-1 1.44e-5 EBC-08733 1:1 binding 6 50μg / mL OPG027 pH4.0 8.75e-1 9.17e+1 7.59e-2 8.28e-4 EBC-303150 Steady state affinity 6 50μg / mL OPG027 pH4.0 1.92e+1 1.71e+2 8.68e-2 1.96e-5 EBC-08787 1:1 binding 7 50μg / mL OPG027 pH4.0 1.51e0 2.66e+3 5.85e-2 2.20e-5 GNE-049 Steady state affinity 7 50μg / mL OPG027 pH4.0 9.58e0 1.52e+3 1.35e-1 1.86e-5 EBC-303150 Steady state affinity 8 50μg / mL OPG198 pH4.0 N / A N / A N / A 2.01e-5 GNE-049 Steady state affinity 8 50μg / mL OPG198 pH4.0 N / A N / A N / A 2.16e-5 EBC-08787 1:1 binding 1 100μg / mL OPG174 pH4.0 4.72e-1 4.29e+2 2.88e-2 6.72e-5 Vorapaxar 1:1 binding 1 100μg / mL OPG174 pH4.0 6.52e0 1.88e+3 1.16e-2 6.18e-6 EBC-07547 1:1 binding 2 100μg / mL OPG174 pH4.0 4.30e-1 2.07e+2 4.70e-2 2.27e-4 EBC-303150 1:1 binding 2 100μg / mL OPG174 pH4.0 2.43e+1 1.21e+3 1.54e-2 1.27e-5

[0048] As shown in Table 1, various drugs have a strong binding strength with proteins.

[0049] Example 4

[0050] This example demonstrates the in vivo viral validation of an antiviral drug:

[0051] The inhibitory effect of the drug on the Wuhan strain of monkeypox virus was verified using the Vero E6 cell line derived from monkeys. Monkeypox virus was cultured on Vero E6 cells. Cells were grown on modified Dulbecco medium (Dulbecco's modified minimal essential medium containing 2 mM L-glutamine, 100 units / ml penicillin, 100 μg / ml streptomycin, and 2% or 10% fetal bovine serum (FBS; Gibco)) at 37°C in a CO2 incubator (containing 5% CO2). Vero E6 cell monolayers were grown in 48-well plates, infected with monkeypox virus (MOI 0.05) for 72 hours, with drug concentrations of 100, 32, 10, 3.2, 1, 0.32, 0.1, and 0.032 μM. Each test concentration had three replicates. Tecoviride was used as a positive control. After infection, microplates were frozen and properly handled before being taken out of the P3 laboratory. DNA was then extracted and the amount of monkeypox virus was determined using RT-qPCR. The gene tested was G2R. The primers used were F: CACACCGTCTCTTCCACAGA; R: GATACAGGTTAATTTCCACATCG SEQ ID NO: 1; the TaqMan probe was 5′FAMAAGCCGTAATCTATGTTGTCTATCGTGTCC-3′BHQ1 SEQ ID NO: 2. The efficacy of the drug is shown in Table 2.

[0052] Table 2

[0053] Drug_name Drug protection effect (inhibition rate%) Average value AdipoRon 55% 52% 47% 53% GNE-049 27% 33% 28% 43% Vorapaxar 53% 65% 63% 79% EBC-07547 38% 40% 33% 48% EBC-08733 62% 64% 65% 66% EBC-08787 78% 78% 76% 81% EBC-303150 88% 90% 90% 92%

[0054] Based on the data in Table 3, the drugs obtained by the screening method of this invention all have a certain inhibitory effect on monkeypox virus, and the screening effectiveness rate is 100%. Therefore, the screening method of this invention can be used to screen for monkeypox virus drugs.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for screening viral drug targets, characterized in that, The method for screening viral drug targets includes the following steps: (a) Obtain the genome sequences of different lineages of monkeypox virus evolution mutations, encode the three-mode characters into digital tokens using a 3-mode character encoding mechanism, and then convert the genome sequence encoding into a 768-dimensional sequence embedding through the token embedding layer. The gene sequence feature extraction network is designed with a 12-layer transformer architecture as the embedding base model. The pre-training uses MLM mask loss for feature association embedding, and the fine-tuning stage uses lineage evolution labels as the target for training. Finally, the gene feature embeddings related to evolution mutations are obtained. (b) Obtain the gene sequence of the virus and convert the gene sequence into the corresponding protein sequence. Use the ESM2.0 protein sequence big language model to learn the amino acid sequence embedding representation and extract the lineage embedding representation of each protein sequence in the mutation lineage. In the embedding base model, the solvable fastKAN deep learning framework is introduced, which combines gene feature embeddings related to evolutionary mutations with lineage embeddings of each protein sequence. At the whole genome level, the set of genes most related to the mutation lineage is predicted, and the gene importance is sorted according to the activation weight of each spline function in fastKAN to obtain the final set of genes most related to the lineage mutation. (d) Calculate the dN / dS ratio of each gene in the final gene set, and select genes with a dN / dS ratio > 1 as viral drug targets.

2. The method for screening viral drug targets according to claim 1, characterized in that, The virus mentioned includes monkeypox virus.

3. The application of the screening method for viral drug targets as described in claim 1 or 2 in antiviral drug screening.

4. A method for screening antiviral drugs, characterized in that, The screening method for the antiviral drugs includes the following steps: (1) Viral drug targets are screened using the method for screening viral drug targets according to claim 1 or 2; (2) AlphaFold3 was used to screen for drugs that can bind to viral drug targets.

5. The method for screening antiviral drugs according to claim 4, characterized in that, In step (2), the drug is derived from the Bioactive compounds database in the Enamine database.

6. The method for screening antiviral drugs according to claim 4, characterized in that, In step (2), AlphaFold3 is used to screen for drugs that bind to the top 5 to 10 viral drug targets.