A method for implementing motif sequence constraints in an mRNA codon-optimized DFA graph

Abstract

The present application relates to the technical field of computer-aided drug research and development, and particularly relates to a method for realizing Motif sequence constraint in a DFA graph of mRNA codon optimization, which comprises the following steps: step A: locating the position of a Motif subgraph in the DFA graph; step B: extending the local DFA graph; step C: deleting the Motif subgraph; and step D: repeating step A until the Motif subgraph is completely deleted. The present application avoids explicitly arranging and combining candidate codons, and instead implicitly includes the constraint in the path of the DFA graph. The DFA graph is processed before optimization starts, thereby avoiding the large amount of calculation required for post-processing of codon arrangement and combination. Thus, the method solves the problem of Motif sequence constraint in mRNA codon optimization, and provides an effective solution for long-sequence mRNA codon design.

Description

Technical Field

[0001] This invention relates to the field of computer-aided drug development technology, specifically to a method for implementing motif sequence constraints in a DFA (finite state automaton) diagram optimized for mRNA codons. Background Technology

[0002] According to the central dogma, messenger RNA (mRNA) is translated into corresponding proteins, thereby exerting biological functions. Each codon (excluding the stop codon) consists of three consecutive residues, from the start codon to the stop codon, corresponding to one natural amino acid. In this process, mRNA acts as an intermediate carrier for the transfer of genetic material from DNA to proteins. Because mRNA is located "upstream" from proteins, theoretically, it can replace drugs at all protein levels. From a pharmaceutical industry perspective, mRNA is a highly promising drug candidate that can meet the needs of gene therapy, cancer treatment, and vaccine development. Furthermore, its overall process is simple, development is rapid, and it exhibits high stability. Therefore, mRNA drugs, as an emerging drug form, have become a research hotspot in the field of drug development.

[0003] In mRNA sequences, the codon-amino acid correspondence is one-to-many, and different codon choices affect properties such as mRNA stability and translation efficiency. Currently, due to technological limitations, some short nucleic acid sequences (motif sequences) must be avoided in mRNA design, including sequences at enzyme cleavage sites. Therefore, the actual number of possible codon permutations is astronomical, urgently requiring computer-aided methods for codon design optimization to improve the structural compactness, transcriptional stability, and translation efficiency of mRNA sequences. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by providing a method for implementing motif sequence constraints in the DFA diagram of mRNA codon optimization. This method implicitly incorporates constraints into the codon optimization process, thereby avoiding a large amount of invalid computation and improving design efficiency.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A method for implementing motif sequence constraints in a DFA graph optimized for mRNA codons is provided, comprising the following steps:

[0007] Step A: Locate the position of the Motif subgraph in the DFA graph to obtain the sequence of edges to which the Motif subgraph belongs;

[0008] Step B: Expand the local DFA graph to separate the Motif subgraph from the rest;

[0009] Step C: Delete the isolated Motif subgraph, that is, delete the corresponding edges and free vertices in the DFA graph;

[0010] Step D: Repeat the above steps until the Motif subimage is completely deleted.

[0011] In the above technical solution, step A involves searching for and locating the Motif subgraph in the DFA graph. The search steps are as follows:

[0012] A1. Start from the initial node of the DFA diagram, which is the leftmost part of the mRNA;

[0013] A2. Match the attributes on the edges of the DFA graph with the bases of the Motif using a depth-recursive approach;

[0014] A3. Iterate through each edge in the DFA graph until a sequence of matching edges is returned, or report a search failure.

[0015] In the above technical solution, step B, the step of expanding the local DFA graph, is as follows:

[0016] B1. Starting from the obtained matching motif subgraph, traverse each edge sequentially from back to front and from front to back, checking the branching of the vertices before and after each edge and applying different graph modification methods, including the following cases:

[0017] (1) From back to front: If the in-degree of a vertex is 1 and the out-degree is greater than 1, it indicates that the current vertex of the Motif subgraph shares vertices with other subgraphs and needs to be expanded. Specifically, create a new vertex, connect the previous and next vertices of the original vertex, copy the attributes of the corresponding edges, delete the original outgoing edges, and finally update the edge sequence of the Motif subgraph, such as... Figure 2 As shown;

[0018] (2) From front to back: If the in-degree of a vertex is greater than 1 and the out-degree is greater than or equal to 1, it indicates that the current vertex of the Motif subgraph shares vertices with other subgraphs and needs to be expanded. Specifically, create a new vertex, connect the previous and next vertices of the original vertex, copy the attributes of the corresponding edges, delete the original incident edges, and finally update the edge sequence of the Motif subgraph, such as... Figure 3 As shown:

[0019] B2. Repeat step B1 until no situation requiring modification of the diagram is found.

[0020] The beneficial effects of this invention are:

[0021] This invention provides a method for implementing motif sequence constraints in a DFA graph for mRNA codon optimization, comprising the following steps: A) locating the position of the motif subgraph in the DFA graph; B) expanding the local DFA graph; C) deleting the isolated motif subgraph; and D) repeating step A until the motif subgraph is completely deleted. Because this process avoids explicitly permuting and combining candidate codons, but instead implicitly embeds the constraints within the paths of the DFA graph, it utilizes a dynamic programming algorithm for codon optimization (with a complexity of O(n log n)). 3 Before computation, the DFA graph is processed. Compared to methods that filter out non-constrained motif sequences after codon optimization, this avoids the large amount of computational overhead required for codon permutation and combination, thus significantly improving design efficiency. Therefore, the method of this invention directly avoids entering the invalid solution space during computational optimization. Without significantly increasing computational load, it implicitly incorporates constraints into the codon optimization process, supporting any number and length of motif sequences. This provides a computationally feasible and effective solution for realizing process-feasible mRNA drug design schemes, especially for long-sequence mRNA codon design. Attached Figure Description

[0022] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0023] Figure 1 This is a flowchart of a method for implementing motif sequence constraints in a DFA diagram optimized for mRNA codons, according to the present invention.

[0024] Figure 2 This is a schematic diagram illustrating the expansion of a local DFA diagram from back to front in step B1 of embodiment 1 of the present invention.

[0025] Figure 3 This is a schematic diagram illustrating the expansion of a local DFA diagram in step B1 of embodiment 1 of the present invention, in a front-to-back manner.

[0026] Figure 4 This is a DFA diagram required for mRNA optimization of the polypeptide MLP in Example 2 of the present invention.

[0027] Figure 5 This is a DFA diagram of the optimized mRNA of the polypeptide MLP in Example 2 of the present invention. Detailed Implementation

[0028] Example 1.

[0029] A method for implementing motif sequence constraints in a DFA graph optimized for mRNA codons includes the following steps:

[0030] Step A: Locate the position of the Motif subgraph in the DFA graph to obtain the sequence of edges to which the Motif subgraph belongs;

[0031] Step B: Expand the local DFA graph to separate the Motif subgraph from the rest;

[0032] Step C: Delete the isolated Motif subgraph, that is, delete the corresponding edges and free vertices in the DFA graph;

[0033] Step D: Repeat the above steps until the Motif subimage is completely deleted.

[0034] In the above technical solution, step A involves searching and locating the Motif subgraph in the DFA graph. The search steps are as follows:

[0035] A1. Start from the initial node of the DFA diagram, which is the leftmost part of the mRNA;

[0036] A2. Match the attributes on the edges of the DFA graph with the bases of the Motif using a depth-recursive approach;

[0037] A3. Iterate through each edge in the DFA graph until a sequence of matching edges is returned, or report a search failure.

[0038] In the above technical solution, step B, the step of expanding the local DFA graph, is as follows:

[0039] B1. Starting from the obtained matching motif subgraph, traverse each edge sequentially from back to front and from front to back, checking the branching of the vertices before and after each edge and applying different graph modification methods, including the following cases:

[0040] (1) From back to front: If the in-degree of a vertex is 1 and the out-degree is greater than 1, it indicates that the current vertex of the Motif subgraph shares vertices with other subgraphs and needs to be expanded. Specifically, create a new vertex, connect the previous and next vertices of the original vertex, copy the attributes of the corresponding edges, delete the original outgoing edges, and finally update the edge sequence of the Motif subgraph, such as... Figure 2 As shown:

[0041] (2) From front to back: If the in-degree of a vertex is greater than 1 and the out-degree is greater than or equal to 1, it indicates that the current vertex of the Motif subgraph shares vertices with other subgraphs and needs to be expanded. Specifically, create a new vertex, connect the previous and next vertices of the original vertex, copy the attributes of the corresponding edges, delete the original incident edges, and finally update the edge sequence of the Motif subgraph, such as... Figure 3As shown:

[0042] B2. Repeat step B1 until no situation requiring modification of the diagram is found.

[0043] The above process avoids explicitly permuting and combining candidate codons, instead embedding the constraints within the path of the DFA graph, i.e., in the dynamic programming algorithm for codon optimization (complexity O(n log n)). 3 The DFA graph is processed before the calculation, thus avoiding the large amount of post-processing computation required for codon permutation and combination, and greatly improving design efficiency. Therefore, the method of this invention directly avoids entering the invalid solution space during the computational optimization process. Without significantly increasing the computational load, it implicitly incorporates constraints into the codon optimization process, and can support any number and length of motif sequences. This provides a computationally feasible and effective solution for realizing process-feasible mRNA drug design schemes, especially for the design of long mRNA codon sequences.

[0044] Example 2.

[0045] The specific implementation of this embodiment takes the polypeptide MLP as an example, where M is an optional codon AUG and is also the start codon. The optional codons for L-leucine are UUA, UUG, CUU, CUC, CUA, and CUG; the optional codons for P-proline are CCU, CCC, CCA, and CCG; and the stop codon has three options: UAA, UGA, and UAG.

[0046] The DFA diagram required for mRNA optimization of this peptide is shown below. Figure 4 As shown:

[0047] The (0,0) node is the start node, and the (12,0) node is the end node. When it is necessary to restrict the Motif sequence to CCCG, the DFA diagram is modified as follows. Figure 5 As shown:

[0048] The constraint takes effect if no sub-path, CCCG, exists on any path from the start node to the end node.

[0049] If the constraint motif sequence is included in a necessary path in the DFA graph, it will cause the DFA graph to break, making the constraint unsolvable. This needs to be checked at the end. For example, if the constraint motif sequence is CC, the DFA graph will break because CC is a necessary path in the entire DFA graph; therefore, the addition of this constraint will make the entire optimization process unsolvable. This can be detected by checking the integrity of the DFA graph after processing.

[0050] Example 3.

[0051] In this embodiment, a typical sequence of the COVID-2019 spike protein was selected, and codon optimization was performed using the method in Example 1. Then, performance testing was conducted, including testing the running time on a computer.

[0052] The test sequence is the COVID-2019 spike protein, and its full length is:

[0053] MFVFLVLLPLVSSQCVNLTTRTQLPPAYTNSFTRGVYYPDKVFRSSVLHSTQDLFLP FFSNVTWFHAIHVSGTNGTKRFDNPVLPFNDGVYFASTEKSNIIRGWIFGTTLDSKTQSLLIVNNATNVVIKVCEFQFCNDPFLGVYYHKNNKSWMESEFRVYSSANNCTFE YVSQPFLMDLEGKQGNFKNLREFVFKNIDGYFKIYSKHTPINLVRDLPQGFSALEPLVDLPIGINITRFQTLLALHRSYLTPGDSSSGWTAGAAAYYVGYLQPRTFLLKYNENG TITDAVDCALDPLSETKCTLKSFTVEKGIYQTSNFRVQPTESIVRFPNITNLCPFGEVFNATRFASVYAWNRKRISNCVADYSVLYNSASFSTFKCYGVSPTKLNDLCFTNVYAD SFVIRGDEVRQIAPGQTGKIADYNYKLPDDFTGCVIAWNSNNLDSKVGGNYNYLY RLFRKSNLKPFERDISTEIYQAGSTPCNGVEGFNCYFPLQSYGFQPTNGVGYQPYRVVVLSFELLHAPATVCGPKKSTNLVKNKCVNFNFNGLTGTGVLTESNKKFLPFQQFG RDIADTTDAVRDPQTLEILDITPCSFGGVSVITPGTNTSNQVAVLYQDVNCTEVPVAIHADQLTPTWRVYSTGSNVFQTRAGCLIGAEHVNNSYECDIPIGAGICASYQTQTNSP RRARSVASQSIIAYTMSLGAENSVAYSNNSIAIPTNFTISVTTEILPVSMTKTSVDCTMYICGDSTECSNLLLQYGSFCTQLNRALTGIAVEQDKNTQEVFAQVKQIYKTPPIKDF GGFNFSQILPDPSKPSKRSFIEDLLFNKVTLADAGFIKQYGDCLGDIAARDLICAQKFNGLTVLPPLLTDEMIAQYTSALLAGTITSGWTFGAGAALQIPFAMQMAYRFNGIGVTQNVLYENQKLIANQFNSAIGKIQDSLSSTASALGKLQDVVNQNAQALNTLVKQLSSNFGAISSVLNDILSRLDKVEAEVQIDRLITGRLQSLQTYVTQQLIRAAEIRASANLA ATKMSECVLGQSKRVDFCGKGYHLMSFPQSAPHGVVFLHVTYVPAQEKNFTTAPAICHDGKAHFPREGVFVSNGTHWFVTQRNFYEPQIITTDNTFVSGNCDVVIGIVNNTV YDPLQPELDSFKEELDKYFKNHTSPDVDLGDISGINASVVNIQKEIDRLNEVAKNLN The sequence ESLIDLQELGKYEQYIKWPWYIWLGFIAGLIAIVMVTIMLCCMTSCCSCLKGCCSCGSCCKFDEDDSEPVLKGVKLHYT* includes the final stop codon (denoted by *). The required set of restricted motifs is GGAUCC, AAGCUU, ACUAGU, GAUAUC, CCCGGG, GGUCUC, GAGACC, CCAUGG, GAAUUC, UAAUACGACUCACUAUAG, AUCUGU CUAUAGUGAGUCGUAUUA, CCCCCC, UUUUUU. The test platform is an Intel(R) Core(TM) i7-6700K CPU@4.00GHz.

[0054] Meanwhile, the above sequence was optimized using methods from Comparative Examples 1 and 2, and the running time was tested, where:

[0055] Comparative Example 1: A method of combined filtering after partial optimization is adopted, that is, basic optimization is performed first (without restricting the Motif sequence), and then the corresponding codons are combined according to the position of the Motif sequence in the result;

[0056] Comparative Example 2: The method of complete combination, that is, enumerating all possible combinations of bases.

[0057] The test results are shown in Table 1.

[0058] Table 1. Performance testing of the COVID-2019 spike protein sequence under different methods (average of 3 tests)

[0059]

[0060] As shown in Table 1, since each secondary structure calculation takes approximately 14 seconds, when using the complete combination method in Comparative Example 2, the entire mRNA sequence will have a length of 2.4 × 10⁻⁶. 632 The number of codon combinations is limited, therefore the total time required for combination computation far exceeds the capabilities of an electronic computer; when using the method of Comparative Example 1, which involves first performing basic optimization and then combining the corresponding codons based on the positions of the Motif sequence in the result, the number of combinations is within 10. 9 While this method significantly reduces the time required compared to Comparative Example 2, it is still unacceptable in practical production. The method of this invention, however, can solve this problem within 10 minutes, avoiding the massive computational demands of codon permutation and combination. Therefore, the method of this invention is feasible in practice and provides a computationally feasible and effective solution for mRNA drug design, especially for long-sequence mRNA codon design.

[0061] The above-described embodiments are preferred embodiments of the present invention and are only used to facilitate the illustration of the present invention. They are not intended to limit the present invention in any way. Any person skilled in the art who makes local modifications or alterations to the technical content disclosed in the present invention without departing from the scope of the technical features of the present invention shall still fall within the scope of the technical features of the present invention.

Claims

1. A method for implementing motif sequence constraints in an mRNA codon-optimized DFA graph, characterized by: Includes the following steps: Step A: Locate the position of the Motif subgraph in the DFA graph to obtain the sequence of edges to which the Motif subgraph belongs; Step B: Expand the local DFA graph to separate the Motif subgraph from the rest; Step C: Delete the isolated Motif subgraph, that is, delete the corresponding edges and free vertices in the DFA graph; Step D: Repeat the above steps until the Motif sub-image is completely deleted; In step A, the search for and location of the Motif subgraph within the DFA graph involves the following steps: A1. Start from the initial node of the DFA diagram, which is the leftmost part of the mRNA; A2. Match the attributes on the edges of the DFA graph with the bases of the Motif using a depth-recursive approach; A3. Iterate through each edge in the DFA graph until a sequence of matching edges is returned, or report a search failure. In step B, the step of expanding the local DFA graph is as follows: B1. Starting from the obtained matching motif subgraph, traverse each edge sequentially from back to front and from front to back, checking the branching of the vertices before and after each edge and applying different graph modification methods, including the following cases: (1) From back to front: If the in-degree of a vertex is equal to 1 and the out-degree is greater than 1, it means that the vertex of the Motif subgraph shares a vertex with other subgraphs and needs to be expanded. The specific method is to create a new vertex, connect the vertices before and after the original vertex, copy the corresponding edge attributes, delete the original outgoing edge, and finally update the edge sequence of the Motif subgraph. (2) From front to back: If the in-degree of the vertex is greater than 1 and the out-degree is greater than or equal to 1, it indicates that the vertex of the Motif subgraph shares the vertex with other subgraphs and needs to be expanded. The specific method is to create a new vertex, connect the vertices before and after the original vertex, copy the corresponding edge attributes, delete the original incident edge, and finally update the edge sequence of the Motif subgraph. B2. Repeat step B1 until no situation requiring modification of the diagram is found.

Images