Method for finding out peptide linkers among different peptides

Abstract

The invention provides a method for finding out peptide linkers among different peptides. The method comprises the following steps. A long sequence is formed, and the long sequence is composed of a plurality of different peptides and peptide connectors among the peptides. And according to the tangent point probabilities predicted by the model A, peptide connector combinations are sorted according to the expressions of the peptides and the peptide connectors, and TopN peptide connector combinations are selected. TopN peptide connector combinations are applied to other models for prediction, and peptide connector combination ranking is generated by considering the tangency probability predicted by other models. And combining the jointly selected peptide connectors, and selecting the peptide connector combination with the highest weighted average ranking.

Description

Technical Field

[0001] The present invention relates to a method for finding peptide linkers, and more particularly to a method for finding peptide linkers between different peptides. Background Art

[0002] With the progress of next-generation gene sequencing and artificial intelligence, personalized cancer vaccines have become one of the means of cancer treatment. The vector template architecture of personalized cancer vaccines includes: prioritizing and selecting tumor neoantigens, and then connecting the peptides with high priority through a peptide linker. Among them, the peptide linker not only serves to connect different possible tumor neoantigens, but also plays an auxiliary role in assisting the peptides to generate an immune response mechanism. However, currently, the selection of peptide linkers is mainly determined by the property of not causing an immune response to determine the sequence of the peptide linker. Such a method ignores the interaction between the peptide and the linker, which may prevent the peptide from being effectively cleaved to cause a downstream immune response. Therefore, it is necessary to consider the relationship between the peptide and the peptide linker and develop a method for finding suitable linkers between different peptides.

[0003] The prior art only uses a fixed linker to connect peptides, without considering that the same linker may have different reactions after being cleaved by the proteasome between different peptides, resulting in incomplete cleavage of the peptide and thus affecting the efficacy of the peptide. In addition, not all amino acid combinations can be synthesized, which may cause the peptides required for the vaccine and the fixed linker sequence to be unable to be synthesized. Therefore, it is necessary to design a dedicated linker for each different peptide to enhance the efficacy of the peptide. Summary of the Invention

[0004] The present invention provides a method for finding peptide linkers between different peptides and peptide linkers to enhance the efficacy of peptides.

[0005] The method for finding peptide linkers between different peptides of the present invention includes the following steps. Compose a long sequence, which is composed of multiple different peptides and peptide linkers between the peptides. Rank the peptide linker combinations according to the performance of the peptides and peptide linkers respectively based on the cleavage site probabilities predicted by Model A, and select the top N peptide linker combinations. Generate a ranking table according to the ranking of the peptide linker combinations, and give rankings to the peptide linker combinations using the main conditions and secondary conditions. Apply the top N peptide linker combinations to other models for prediction. Considering the cleavage site probabilities predicted by other models, rank the peptide linker combinations according to the performance of the peptides and peptide linkers, and also generate the main condition ranking table of other models, and use the secondary conditions to rank to generate the ranking of the peptide linker combinations. The combination selection order must be at least equal to or better than the number of Model A, and the co-selected peptide linkers are subjected to subsequent analysis. Combine the co-selected peptide linkers and select the peptide linker combination with the highest weighted average ranking.

[0006] In an embodiment of the present invention, Model A includes Pepsickle or NetCleave.

[0007] In an embodiment of the present invention, the main conditions include the number of peptides cleaved and the ranking of the cleavage positions of the peptide linkers.

[0008] In an embodiment of the present invention, the secondary condition includes the average probability of cleavage of the peptide linker minus the average probability of cleavage of the peptide.

[0009] In an embodiment of the present invention, selecting the top N peptide linker combinations includes selecting the top N peptide linker combinations one by one according to the combination selection order of the main condition ranking table of the Model A, and ranking the top N peptide linker combinations according to the secondary conditions to generate the ranking of the peptide linker combinations. Then use other models for prediction. If a suitable peptide linker is selected, stop; if not, continue to select in order until a peptide linker is selected.

[0010] Based on the above, the present invention provides a method for finding peptide linkers between different peptides. By considering the relationship between peptides and peptide linkers, and considering that the same linker may have different reactions after being cleaved by the proteasome between different peptides, the method is used to find suitable linkers between different peptides and improve the potency of this peptide. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a flowchart of the method for finding peptide linkers between different peptides according to an embodiment of the present invention;

[0012] Figure 2 is a schematic diagram of a long sequence according to an embodiment of the present invention;

[0013] Figure 3 It is a schematic diagram of the main conditions and secondary conditions according to an embodiment of the present invention. Detailed implementation mode

[0014] Examples are listed below and described in detail in conjunction with the accompanying drawings, but the provided examples are not intended to limit the scope covered by this disclosure. In addition, terms such as "including", "comprising", "having", etc. used in this text are all open-ended terms, that is, "including but not limited to".

[0015] Figure 1 It is a flowchart of a method for finding peptide linkers between different peptides according to an embodiment of the present invention. Please refer to Figure 1 , the method for finding peptide linkers between different peptides of the present invention includes four major parts, which will be specifically introduced in sequence below in conjunction with Figure 1 These four major parts will be specifically introduced in sequence.

[0016] Figure 2 It is a schematic diagram of a long sequence according to an embodiment of the present invention. Please refer to Figure 1 and Figure 2 , first, Part 1 is to form a long sequence (peptide + peptide linker + peptide), and set the initial length of the peptide linker = L, and there are 20 (amino acid)^L peptide linker combinations, such as Figure 2 which is an example with a peptide linker length of 3.

[0017] Please continue to refer to Figure 1 . After that, in Part 2, the cleavage probabilities predicted by Model A are used to rank the peptide linker combinations according to the performance of the peptides and peptide linkers respectively, and the TopN peptide linker combinations are selected. In this embodiment, Model A is usually the most commonly used model by scholars or the model with the best prediction performance, such as: Pepsickle, NetCleave. According to the peptide linker combination ranking conditions, a ranking table is generated. Rank the peptide linker combinations according to the main conditions and secondary conditions. For example, Model A can take Pepsicle as an example, and other models can take NetCleave as an example, but the present invention is not limited thereto. Apply it to other K models for prediction. Therefore, K main and secondary condition order tables will be generated, not limited to two models. If K models are applied, the same screening method is also used, and the combination selection order must be at least equal to or better than the number of Model A. In addition, usually Model A will select the model with the best performance or a large enough amount of data.

[0018] Main condition: (assuming the probability of being cleaved is >0.5)

[0019] A. The number of cleavages of peptide 1 and peptide 2 (the smaller the better)

[0020] B. Sorting of the positions where the peptide linker is cleaved (defined by the tester himself / herself, usually the peptide linker is expected to be cleaved in the middle). For example: When the length of the peptide linker is 3, the position sorting is shown in Table 1 below (a total of 2^3 = 8 sorts).

[0021] Table 1

[0022]

[0023] Secondary condition (assuming the probability of being cleaved is > 0.5)

[0024] A. The average probability of the peptide linker being cleaved minus the average probability of the peptide being cleaved (the larger the better, usually it is desired that the peptide linker is cleaved and the peptide is not cleaved). Please refer to Figure 3 for the example, Figure 3 which is a schematic diagram of the main conditions and secondary conditions according to an embodiment of the present invention.

[0025] Determine the TopN linker combinations. According to the combination selection order of the main condition sorting table of Model A, select the TopN peptide linker combinations one by one, and sort the TopN peptide linker combinations according to the secondary conditions to generate the peptide linker combination ranking. Then use other models to predict. If a suitable peptide linker is selected, stop; if not, continue to select in sequence until a peptide linker is selected. As illustrated in Table 2 below: First, according to the number of peptide linker combinations Top6 of combination selection order 1, use models B, C,..., K to predict. If not selected, then select in sequence according to combination selection orders 2, 3, 4,... until a peptide linker is selected.

[0026] Table 2

[0027]

[0028] Continue to refer to Figure 1 . Next, the third part is to apply the TopN peptide linker combinations to other models for prediction, consider the cleavage point probability predicted by other models, sort the peptide linker combinations according to the performance of the peptide and the peptide linker, and also generate the main condition sorting table of other models, and use the secondary conditions to sort to generate the peptide linker combination ranking. The combination selection order must be at least equal to or better than the number of Model A, and the jointly selected peptide linkers are subjected to subsequent analysis. In this example, the Top6 peptide linkers in combination selection order 1 are predicted using the NetCleave model to obtain the main condition A, main condition B, and secondary conditions of these six peptide linkers. Select the main condition A: the peptide linker with 0 peptide cleavage numbers (that is, TWG & SWG) and sort according to the secondary condition probability difference to obtain SWG as rank1 and TWG as rank2. (Please refer to Table 3 below)

[0029] Table 3

[0030]

[0031] Please continue to refer to Figure 1 。Then, the fourth part is to combine the co-selected peptide linkers and select the peptide linker combination with the highest weighted average ranking.

[0032] Weighted_Average=(W A *R A +W B *R B +…) / K

[0033] W i : The weight of the i-th model, R i : The ranking of the i-th model, i = 1, 2, … K

[0034] For example: Suppose the peptide linker length is 3, and the weights of model A and model B are both 0.5. After screening by the combination label 1 of model A (peptide cleavage number = 0, peptide linker cleavage position = 010), only the main conditions of TWG and SWG are met. After calculating the weighted average ranking, the peptide linker of TWG will be selected (0.5 * 1 + 0.5 * 2 = 1.5). (Please refer to Table 4 below)

[0035] Table 4

[0036]

[0037] In summary, the present invention provides a method for finding peptide linkers between different peptides. By considering the relationship between peptides and peptide linkers and the fact that the same linker may have different reactions after being cleaved by the proteasome between different peptides, the method is used to find suitable linkers between different peptides. Therefore, the found peptide linkers can be used to link tumor neoantigens, design exclusive linkers for each different peptide, and enhance the potency of this tumor neoantigen.

Claims

1. A method for finding peptide connectors between different peptides, characterized in that: include: Composing a long sequence, wherein the long sequence is composed of a plurality of different peptides and peptide connectors between the peptides; The cut point probability predicted by model A is used to sort the peptide connector combinations according to the performance of peptides and peptide connectors, and the top N peptide connector combinations are selected and sorted according to the peptide connector combinations to generate a sorting table, and the peptide connector combinations are ranked using the main conditions and the secondary conditions; Apply the TopN peptide connector combinations to other models for prediction, consider the cut point probability predicted by other models, sort the peptide connector combinations according to the performance of peptides and peptide connectors, and similarly generate a primary condition ranking table for other models, and use the secondary condition ranking to generate a ranking of peptide connector combinations. The combination selection order must be at least equal to or better than the number of model A, and the peptide connectors selected together will be subjected to subsequent analysis; and The peptide connector combinations selected together are combined, and the peptide connector combination with the highest weighted average ranking is selected.

2. The method for finding peptide connectors between different peptides according to claim 1, characterized in that: The model A includes Pepsickle or NetCleave.

3. The method for finding peptide connectors between different peptides according to claim 1, characterized in that: The main conditions include the number of peptides to be cut and the order of positions at which the peptide connectors are cut.

4. The method for finding peptide connectors between different peptides according to claim 1, characterized in that: The secondary condition includes the average probability of a peptide linker being cleaved minus the average probability of a peptide being cleaved.

5. The method for finding peptide connectors between different peptides according to claim 1, characterized in that: Selecting the TopN peptide connector combinations includes selecting the TopN peptide connector combinations one by one according to the combination selection order of the main condition ranking table of the model A, and sorting the TopN peptide connector combinations by secondary conditions to generate a peptide connector combination ranking, and then using other models to predict. If a suitable peptide connector is selected, stop, if not, select in order until a peptide connector is selected.

Images