A method and device for acquiring a multi-site mutant sequence sample

Abstract

Embodiments of the present application provide a multi-site mutant sequence sample acquisition method and device. The method comprises: performing saturated single mutation processing on elements at each position in a protein sequence to obtain a saturated single mutant sequence, and acquiring a function value matrix corresponding to the saturated single mutant sequence; preprocessing the function value matrix to obtain a probability distribution matrix for indicating an amino acid probability distribution at each position; and performing sampling processing on the probability distribution matrix based on a specified sampling method to obtain a multi-site mutant sequence sample. Embodiments of the present application can realize that the sample obtained by sampling from the mutant amino acid probability distribution not only controls the number of mutations between the sample and the natural coat, but also the proportion of available samples in the sample set is significantly better than random mutation, and the generation efficiency is significantly better than rational artificial design.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of chemistry, in particular to a multi-site mutant sequence sample acquisition method and device. BACKGROUND

[0002] AAV (Adeno-associated viruses) capsids have become a powerful tool for therapeutic in vivo gene delivery. However, the transduction efficiency of the natural capsid still limits the therapeutic purposes, and in addition, due to the complexity of the genotype-phenotype relationship and many functional properties that must be optimized at the same time, the engineering of enhanced capsids has proven to be challenging.

[0003] At present, the physical interactions that determine the function of proteins are not well understood, so more means are directed evolution, that is, when the mechanism is limited, repeated application of random mutations and artificial selection is usually the default engineering strategy. But the capsids designed in this way are unusable in a very high proportion, and the production efficiency is very low.

[0004] At present, there is also a new method of training a binary classifier by using a small amount of data with a certain function first (in this method, the function of the capsid is that the virus corresponding to the designed capsid sequence has activity). Then randomly sample in the randomly divided mutant subspace (the construction method of the subspace is shown in the Figure 1 The sequence has activity or no activity). If it is determined by the binary classifier that it has activity, the sample is retained, and if it has no activity, it is deleted. Finally, through such a constantly circulating screening method, a set of capsid sequences with expected activity is screened out. The sequence set constructed by this method has a higher proportion of active sequences than the random mutation sequence set, but the premise of using this method is that there is multi-mutation data, and only single-mutation data cannot train the model, nor can it judge whether the multi-mutation sequence has a specific function. SUMMARY

[0005] The technical problem to be solved by the embodiments of the present application is to provide a multi-site mutant sequence sample acquisition method and device to realize that the samples obtained from the mutation amino acid probability distribution not only control the number of mutations between the natural capsid, but also the proportion of available samples in the sample set is significantly better than random mutation, and the generation efficiency is significantly better than rational artificial design.

[0006] In a first aspect, the embodiments of the present application provide a multi-site mutant sequence sample acquisition method, which comprises:

[0007] The elements of each position in the protein sequence are subjected to saturated single mutation processing to obtain a saturated single mutation sequence, and a function value matrix corresponding to the saturated single mutation sequence is obtained;

[0008] The function value matrix is preprocessed to obtain a probability distribution matrix indicating the amino acid probability distribution of each position;

[0009] The probability distribution matrix is sampled based on a specified sampling method to obtain a multi-site mutation sequence sample.

[0010] Optionally, the elements of each position in the protein sequence are subjected to saturated single mutation processing to obtain a saturated single mutation sequence, and a function value matrix corresponding to the saturated single mutation sequence is obtained, comprising:

[0011] The elements of each position in the protein sequence are subjected to saturated single mutation processing to obtain a saturated single mutation sequence;

[0012] The function value corresponding to the saturated single mutation sequence is calculated;

[0013] According to the function value corresponding to each position, a function value matrix corresponding to the saturated single mutation sequence is generated.

[0014] Optionally, the function value matrix is preprocessed to obtain a probability distribution matrix indicating the amino acid probability distribution of each position, comprising:

[0015] The function value of the amino acid type at the insertion position in the function value matrix is replaced with 0, and the element with a function value less than 0 in the function value matrix is replaced with negative infinity to obtain an intermediate function value matrix;

[0016] For different positions of the intermediate function value matrix, the function values of different amino acids of the intermediate function value matrix are normalized to obtain the amino acid probability distribution at the corresponding position;

[0017] According to the amino acid probability distribution at each position in the intermediate function value matrix, a probability distribution matrix indicating the amino acid probability distribution of each position is generated.

[0018] Optionally, the probability distribution matrix is sampled based on a specified sampling method to obtain the multi-site mutation sequence sample, comprising:

[0019] In the case where the specified sampling method is a mutation number uncontrollable sampling method, each position in the probability distribution matrix is traversed;

[0020] According to the amino acid probability distribution corresponding to each position in the traversal, the probability distribution matrix is sampled to obtain the amino acid type corresponding to each position.

[0021] Splicing the amino acid type corresponding to each position to obtain the multi-site mutation sequence sample.

[0022] Optionally, the sampling processing of the probability distribution matrix based on the specified sampling mode to obtain the multi-site mutation sequence sample comprises:

[0023] In the case that the specified sampling mode is a mutation number controllable sampling mode, the entropy of the amino acid probability distribution under different positions of the probability distribution matrix is calculated to obtain the mutation probability corresponding to each position;

[0024] The probability distribution matrix is sampled according to the set mutation number and the mutation probability to obtain the multi-site mutation sequence sample.

[0025] Optionally, the calculation of the entropy of the amino acid probability distribution under different positions of the probability distribution matrix to obtain the mutation probability corresponding to each position comprises:

[0026] The entropy of the amino acid probability distribution under different positions of the probability distribution matrix is calculated to obtain an entropy vector corresponding to each position;

[0027] The value less than the set value in the entropy vector is set to negative infinity;

[0028] The entropy vector is normalized to obtain the mutation probability under each position.

[0029] Optionally, the sampling processing of the probability distribution matrix according to the set mutation number and the mutation probability to obtain the multi-site mutation sequence sample comprises:

[0030] According to the set mutation number, the set mutation number of positions are extracted on the mutation probability;

[0031] The amino acid type in the set mutation number of positions is sampled on the probability distribution matrix;

[0032] The extracted amino acid type and the natural amino acid type on other positions are spliced according to the position sequence to obtain the multi-site mutation sequence sample.

[0033] In a second aspect, an embodiment of the present application provides a multi-site mutation sequence sample acquisition device, the device comprising:

[0034] A function value matrix acquisition module is configured to perform saturated single mutation processing on elements at each position in a protein sequence to obtain a saturated single mutation sequence, and acquire a function value matrix corresponding to the saturated single mutation sequence.

[0035] The probability distribution matrix acquisition module is configured to preprocess the function value matrix to obtain a probability distribution matrix used for indicating an amino acid probability distribution of each position.

[0036] The multi-site mutation sequence sample acquisition module is configured to sample the probability distribution matrix based on a specified sampling manner to obtain a multi-site mutation sequence sample.

[0037] Optionally, the function value matrix acquisition module comprises:

[0038] The saturated single-mutation sequence acquisition unit is configured to perform mutation processing on elements of each position in the protein sequence to any amino acid to obtain a saturated single-mutation sequence.

[0039] The function value calculation unit is configured to calculate a function value corresponding to the saturated single-mutation sequence.

[0040] The function value matrix generation unit is configured to generate a function value matrix corresponding to the saturated single-mutation sequence according to the function value corresponding to each position.

[0041] Optionally, the probability distribution matrix acquisition module comprises:

[0042] The intermediate function value matrix acquisition unit is configured to replace a function value of an empty amino acid type at an insertion position in the function value matrix with 0 and replace an element with a function value less than 0 in the function value matrix with negative infinity to obtain an intermediate function value matrix.

[0043] The amino acid probability distribution acquisition unit is configured to perform normalization processing on function values of different amino acids in the intermediate function value matrix for different positions of the intermediate function value matrix to obtain an amino acid probability distribution at a corresponding position.

[0044] The probability distribution matrix generation unit is configured to generate a probability distribution matrix used for indicating an amino acid probability distribution of each position according to the amino acid probability distribution at each position in the intermediate function value matrix.

[0045] Optionally, the multi-site mutation sequence sample acquisition module comprises:

[0046] The matrix position traversal unit is configured to traverse each position in the probability distribution matrix in a case where the specified sampling manner is a sampling manner in which the number of mutations is uncontrollable.

[0047] The amino acid type acquisition unit is configured to sample the probability distribution matrix according to an amino acid probability distribution corresponding to each position that is traversed to obtain an amino acid type corresponding to each position.

[0048] The first sequence sample obtaining unit is configured to splice the amino acid types corresponding to the positions to obtain the multi-site mutant sequence sample.

[0049] Optionally, the multi-site mutant sequence sample obtaining module comprises:

[0050] The mutation probability obtaining unit is configured to, in the case that the specified sampling mode is a sampling mode in which the number of mutations is controllable, calculate the entropy of the amino acid probability distribution at different positions of the probability distribution matrix to obtain the mutation probability corresponding to each position.

[0051] The second sequence sample obtaining unit is configured to sample the probability distribution matrix according to the set number of mutations and the mutation probability to obtain the multi-site mutant sequence sample.

[0052] Optionally, the mutation probability obtaining unit comprises:

[0053] The entropy vector obtaining sub-unit is configured to calculate the entropy of the amino acid probability distribution at different positions of the probability distribution matrix to obtain an entropy vector corresponding to each position.

[0054] The entropy vector value adjusting sub-unit is configured to set the values smaller than a set value in the entropy vector to negative infinity.

[0055] The mutation probability obtaining sub-unit is configured to normalize the entropy vector to obtain the mutation probability at each position.

[0056] Optionally, the second sequence sample obtaining unit comprises:

[0057] The mutation position extracting sub-unit is configured to extract the positions of the set number of mutations from the mutation probability according to the set number of mutations.

[0058] The amino acid type obtaining sub-unit is configured to sample the amino acid types at the positions of the set number of mutations from the probability distribution matrix.

[0059] The multi-site mutant sequence sample obtaining sub-unit is configured to splice the extracted amino acid types and the natural amino acid types at other positions according to the position sequence to obtain the multi-site mutant sequence sample.

[0060] In a third aspect, an electronic device is provided, comprising:

[0061] A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the multi-site mutant sequence sample obtaining method according to any one of the preceding aspects when executing the program.

[0062] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the multi-site mutant sequence sample acquisition method described in any of the preceding aspects.

[0063] Compared with the prior art, the embodiments of the present application have the following advantages:

[0064] In the embodiments of the present application, each position in the protein sequence is subjected to saturated single mutation processing to obtain a saturated single mutant sequence, and a functional value matrix corresponding to the saturated single mutant sequence is obtained. The functional value matrix is preprocessed to obtain a probability distribution matrix for indicating the amino acid probability distribution of each position. The probability distribution matrix is subjected to sampling processing based on a specified sampling method to obtain a multi-site mutant sequence sample. According to the existing saturated single mutant data of the required capsid function (the function can include but is not limited to having a capsid activity, having a capsid de-immunogenicity, and having a capsid targeting certain specific cells), the embodiments of the present application construct a series of sampling processes, and finally obtain samples from the amino acid probability distribution. The number of mutations between the obtained samples and the natural capsid can be controlled, the proportion of available samples in the sample set is obviously better than that of random mutations, and the generation efficiency is obviously better than that of rational artificial design.

[0065] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 A schematic diagram of a subspace construction method provided by an embodiment of the present application;

[0067] Figure 2 A step flowchart of a multi-site mutant sequence sample acquisition method provided by an embodiment of the present application;

[0068] Figure 3 A schematic diagram of the appearance of a folded capsid protein sequence provided by an embodiment of the present application;

[0069] Figure 4 A schematic diagram of the genome structure of an AAV vector provided by an embodiment of the present application;

[0070] Figure 5 A schematic diagram of a saturated single mutation provided by an embodiment of the present application;

[0071] Figure 6 A structural schematic diagram of a multi-site mutant sequence sample acquisition device provided by an embodiment of the present application;

[0072] Figure 7 A structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0073] In order to make the above objectives, features and advantages of the present application more clear and easy to understand, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0074] The terminology used in the embodiments of the present application is merely for the purpose of describing specific embodiments and is not intended to limit the present application. The singular forms "a," "an," and "the" used in the embodiments of the present application and the appended claims are intended to include plural forms as well, unless the context clearly indicates otherwise.

[0075] It should be noted that, in this document, the terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or terminal including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or terminal. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or terminal including the element.

[0076] First, the following can be combined Figure 3 The coat protein sequence is described as follows.

[0077] Referring to Figure 3 , a schematic diagram of the appearance of a coat protein sequence after folding is shown. As Figure 3 indicated, the coat of an adeno-associated virus is a T=1 icosahedron (an icosahedron is a convex polyhedron made up of 20 triangular faces, each of which encloses a pentagonal vertex, with a five-fold rotation axis between each pair of opposite pentagonal vertices. There is a three-fold rotation axis between each pair of opposite triangular centers; there is a two-fold rotation axis between each pair of opposite edge midpoints). It is assembled by 60 VP monomers through the interaction of these rotation axes, which can all be VP3 or can be composed of VP1, VP2 and VP3. Figure 4 A schematic diagram of a saturated single mutation is shown, as Figure 4As shown, the VP1, VP2, VP3 monomer protein sequence correlation, from the figure can be seen that VP3 is contained in VP2 and VP1, therefore, VP3 is the common region of the three VPs, also known as VP3 common region; VP2 is longer than VP3 by about 57 amino acid sequences, and the sequence is contained in VP1, therefore, also called VP1 / VP2 common region; only VP1 N-terminal has a sequence of about 138 amino acids, therefore, also called VP1 unique region (VP1u). Previous studies have shown that the sequence mutation active region of VP monomer is generally in the VP3 region overlapping different monomers. Therefore, the possible mutation region is selected in the VP3 region, and then saturated single mutation is performed on the mutation region. The mutation form of the saturated single mutation also includes inserting any amino acid between two sites. The saturated single mutation can be as follows Figure 5 As shown, wherein "_" represents insertion at this position, and "*" represents the type of replaced amino acid as empty. Assuming that the natural sequence is "DEEEIR", when performing saturated single mutation, each position is replaced with any one of the amino acids below the arrow, and the other positions remain unchanged. For the type of inserted amino acid, any one of the 20 common amino acids can be selected. For the replacement of the amino acid at the existing position, any one of the remaining 19 amino acids in addition to the current amino acid plus a deletion type, a total of 20 types can be selected. If the current amino acid is replaced by deletion, it means that the amino acid at the position will be deleted.

[0078] After obtaining the saturated single mutation data, biological experiments are performed, and the function value of the sequence when each position is mutated to any amino acid is calculated. At this time, we obtain a function value matrix of different amino acid types at different positions. Then a series of data processing is started, and finally the ideal distribution is obtained. Sampling is performed on the distribution to obtain the final multi-site mutation sequence sample.

[0079] Next, the acquisition process of the multi-site mutation sequence sample will be described in detail in combination with specific embodiments.

[0080] Referring to Figure 2 , a step flowchart of a multi-site mutation sequence sample acquisition method provided by an embodiment of the present application is shown, as Figure 2 shown, the multi-site mutation sequence sample acquisition method can include the following steps:

[0081] Step 201: Perform saturated single mutation processing on the elements of each position in the protein sequence to obtain a saturated single mutation sequence, and acquire a function value matrix corresponding to the saturated single mutation sequence.

[0082] The embodiments of the present application can be applied to the scenario of generating multi-mutation capsid data from single-mutation capsid data.

[0083] In the embodiment, the saturated protein mutation sequence can be obtained first, wherein the saturation refers to mutation of each position in the protein sequence, and other positions keep the same amino acid type. It can be understood that the number of the saturated single mutation sequence corresponding to a protein sequence is multiple, which is related to the number of elements in the protein sequence. For example, if the number of elements in a protein sequence is 17, the number of single mutation sequences corresponding to the protein sequence is 17*19+17*20=663, and the 663 single mutation sequences are the saturated single mutation sequences.

[0084] It can be understood that the above examples are only examples for better understanding the technical solutions of the embodiments of the present application, and are not the only limitation of the embodiments.

[0085] In the process of generating the data of the multi-mutation capsid by using the data of the single-mutation capsid, the saturated single mutation sequence corresponding to the protein sequence can be obtained first. The specific process can be described above, and the embodiments will not be described here.

[0086] After obtaining the protein sequence, the elements at each position in the protein sequence can be mutated to obtain the function value matrix corresponding to the saturated single mutation sequence. Specifically, after obtaining the saturated single mutation sequence, biological experiments can be performed on the saturated single mutation sequence, i.e., amino acid mutation processing is performed on each position in the protein sequence, and then the function value of the sequence when the mutation is any amino acid at each position can be calculated. Further, the function value matrix corresponding to the saturated single mutation sequence can be generated according to the function value corresponding to each position. The implementation process can be described in detail in combination with the specific implementation mode described below.

[0087] In a specific implementation mode of the present application, the step 201 can include:

[0088] Sub-step A1: performing mutation processing of any amino acid on the elements at each position in the protein sequence to obtain the saturated single mutation sequence.

[0089] In the embodiment, the saturated single mutation sequence refers to the sequence obtained after mutation processing of any amino acid on the elements at any position in the protein sequence.

[0090] After obtaining the saturated single mutation sequence by performing mutation processing of any amino acid on the elements at each position in the protein sequence, sub-step A2 is executed.

[0091] Sub-step A2: calculating the function value corresponding to the saturated single mutation sequence.

[0092] After the saturation single mutation sequence is obtained by performing the mutation treatment of any amino acid on the element of each position in the protein sequence, the function value corresponding to the saturation single mutation sequence can be calculated.

[0093] It can be understood that the function value of the saturation single mutation sequence can be calculated by experiment.

[0094] After the function value corresponding to the saturation single mutation sequence is calculated, substep A3 is performed.

[0095] Substep A3: generating a function value matrix corresponding to the saturation single mutation sequence according to the function value corresponding to each position.

[0096] After the function value corresponding to the saturation single mutation sequence is calculated, a function value matrix corresponding to the saturation single mutation sequence can be generated according to the function value corresponding to each position.

[0097] Specifically, after obtaining the saturation single mutation data, biological experiments are performed to calculate the function value of the sequence when the element at each position is mutated to any amino acid. At this time, a function value matrix of different amino acid types under different positions can be obtained.

[0098] After the saturation single mutation sequence is obtained by performing the mutation treatment of any amino acid on the element of each position in the protein sequence, and the function value matrix corresponding to the saturation single mutation sequence is obtained, step 202 is performed.

[0099] Step 202: preprocessing the function value matrix to obtain a probability distribution matrix for indicating the amino acid probability distribution of each position.

[0100] After the saturation single mutation sequence is obtained by performing the mutation treatment of any amino acid on the element of each position in the protein sequence, and the function value matrix corresponding to the saturation single mutation sequence is obtained, the function value matrix can be preprocessed to obtain a probability distribution matrix for indicating the amino acid probability distribution of each position. Specifically, the function value of the amino acid type at the insertion position can be replaced from null to 0, and the element with a function value less than 0 in the function value matrix can be replaced by negative infinity. Then, the function values of different amino acids at the same column, i.e., the same position, are normalized to obtain the amino acid probability distribution at the position. In this way, each column is calculated to obtain the amino acid probability distribution of each position, and finally the probability distribution matrix is obtained. The implementation process can be described in detail in combination with the specific implementation mode described below.

[0101] In a specific implementation mode of the present application, the above step 202 can include:

[0102] Sub-step B1: replace the functional value of the amino acid type null at the insertion position in the functional value matrix with 0, and replace the element with a functional value less than 0 in the functional value matrix with negative infinity, to obtain an intermediate functional value matrix.

[0103] In this embodiment, after obtaining the functional value matrix, the functional value of the amino acid type null at the insertion position in the functional value matrix can be replaced from null to 0, and the element with a functional value less than 0 in the functional value matrix can be replaced with negative infinity, to obtain an intermediate functional value matrix. Specifically, after obtaining the functional value matrix, the insertion position can be replaced with a deletion element on the basis of this matrix, and the default element value is 0. Because of the deletion at the insertion position, there is no difference between the representative and the natural capsid, and the functional value of the natural capsid is 0. If the functional value of the mutated capsid sequence is greater than 0, it means that the sequence has the function, and if the functional value is less than 0, it means that the sequence does not necessarily have the function. Different rows of the matrix represent different amino acid types, and the 0th row represents the deletion amino acid type. Different columns represent different positions.

[0104] After the above processing, the matrix is further processed to replace the element with a functional value less than 0 with negative infinity (the calculation of the default functional value has been calculated by the ln function), so that an intermediate functional value matrix can be obtained.

[0105] It can be understood that the scheme of replacing the element with a functional value less than 0 in the functional value matrix with negative infinity can also be to replace the element with a functional value less than 0 in the functional value matrix with a given value, which can be a very small value that can be ignored.

[0106] After obtaining the intermediate functional value matrix, sub-step B2 is performed.

[0107] Sub-step B2: for different positions of the intermediate functional value matrix, the functional values of different amino acids of the intermediate functional value matrix are normalized to obtain the amino acid probability distribution at the corresponding position.

[0108] After obtaining the intermediate functional value matrix, the functional values of different amino acids of the intermediate functional value matrix can be normalized for different positions of the intermediate functional value matrix to obtain the amino acid probability distribution at the corresponding position. Specifically, for the intermediate functional value matrix, the functional values of different amino acids can be normalized along the same column, i.e. the same position, to obtain the amino acid probability distribution at the position.

[0109] In this embodiment, the normalization processing method can be to normalize by using softmax. Alternatively, 2 x / TZ, where x represents a specific functional value, T represents a temperature coefficient, if the temperature coefficient is larger, the probability distribution of the sampling is more uniform, and Z can ensure that the sum of all probabilities is equal to 1.

[0110] After the functional values of different amino acids in the intermediate functional value matrix are normalized to obtain the amino acid probability distribution at the corresponding position, sub-step B3 is performed.

[0111] Sub-step B3: generating a probability distribution matrix for indicating the amino acid probability distribution of each position according to the amino acid probability distribution at each position in the intermediate functional value matrix.

[0112] After the functional values of different amino acids in the intermediate functional value matrix are normalized to obtain the amino acid probability distribution at the corresponding position, a probability distribution matrix for indicating the amino acid probability distribution of each position can be generated according to the amino acid probability distribution at each position in the intermediate functional value matrix. That is, according to the amino acid probability distribution corresponding to each position, the corresponding probability distribution matrix E ∈ R m×n , where m represents the number of types of amino acids, and n represents the maximum length of the sequence.

[0113] After the functional value matrix is preprocessed to obtain the probability distribution matrix for indicating the amino acid probability distribution of each position, step 203 is performed.

[0114] Step 203: sampling processing the probability distribution matrix based on a specified sampling method to obtain the multi-site mutant sequence sample.

[0115] In this embodiment, the specified sampling method can include at least one of a mutation number uncontrollable sampling method and a mutation number controllable sampling method.

[0116] After the functional value matrix is preprocessed to obtain the probability distribution matrix for indicating the amino acid probability distribution of each position, the probability distribution matrix can be sampled based on a specified sampling method to obtain a multi-site mutant sequence sample.

[0117] The embodiments of the present application construct a series of sampling processes according to the existing saturated single mutation data of the required capsid function (the function can include but is not limited to the capsid having activity, the capsid having deimmunogenicity, and the capsid having targeting to a certain specific cell), sample from the amino acid probability distribution, and finally obtain a sample which not only controls the number of mutations between the natural capsid and the sample, but also has a significantly better proportion of available samples in the sample set than random mutations, and has a significantly better generation efficiency than rational artificial design.

[0118] Next, the sampling process is described in detail as follows in combination with the above-mentioned two sampling methods.

[0119] In an implementation form of the application, the step 203 can comprise:

[0120] Sub-step C1 : in the case that the specified sampling mode is a mutation number uncontrollable sampling mode, traversing each position in the probability distribution matrix.

[0121] In this embodiment, in the case that the specified sampling mode is a mutation number uncontrollable sampling mode, each position in the probability distribution matrix can be traversed first.

[0122] After traversing each position in the probability distribution matrix, sub-step C2 is executed.

[0123] Sub-step C2: sampling the probability distribution matrix according to the amino acid probability distribution corresponding to each traversed position to obtain the amino acid type corresponding to each position.

[0124] After traversing each position in the probability distribution matrix, the probability distribution matrix can be sampled according to the amino acid probability distribution corresponding to each traversed position to obtain the amino acid type corresponding to each position.

[0125] After sampling the probability distribution matrix according to the amino acid probability distribution corresponding to each traversed position to obtain the amino acid type corresponding to each position, sub-step C3 is executed.

[0126] Sub-step C3: concatenating the amino acid types corresponding to each position to obtain the multi-site mutation sequence sample.

[0127] After sampling the probability distribution matrix according to the amino acid probability distribution corresponding to each traversed position to obtain the amino acid type corresponding to each position, the amino acid types corresponding to each position can be concatenated to obtain the multi-site mutation sequence sample. The specific implementation process can be: for the probability distribution matrix, each position can be traversed, and the amino acid probability distribution corresponding to the position is sampled to obtain the type of amino acid at the position. Finally, the types of amino acids obtained at each position are concatenated to obtain the multi-site mutation sequence sample. Because the multi-site mutation sequence sample obtained by sampling at some positions can be a deletion type, although the length of the traversal is fixed, the sequence obtained after concatenation is not necessarily equal in length. Then, the above process is repeatedly repeated to extract more multi-site mutation sequence samples, and the sampling process is stopped after reaching a certain order of magnitude.

[0128] In another implementation form of the application, the step 203 can comprise:

[0129] Sub-step D1: in the case that the specified sampling mode is a mutation number controllable sampling mode, the entropy of the amino acid probability distribution at different positions of the probability distribution matrix is calculated to obtain the mutation probability corresponding to each position.

[0130] In the process of the above sub-step C1 to sub-step C3, the extracted sample does not limit the number of final mutations, so the number of mutations is uncontrollable, and if the entropy of the calculated probability distribution at each position is relatively large, the sequence obtained after sampling for a certain number of positions and time steps will have a large probability of being significantly different from the natural coat, and the number of mutations will be large. However, if a sample with a number of mutations that is relatively small compared to the natural coat, such as 2-8, is desired, the sampling mode of the above sub-step C1 to sub-step C3 may not necessarily be satisfactory. Therefore, a mutation number controllable sampling can be performed.

[0131] In the case that the specified sampling mode is a mutation number controllable sampling mode, the entropy of the amino acid probability distribution at different positions of the probability distribution matrix can be calculated to obtain the mutation probability corresponding to each position. Specifically, the entropy of the amino acid probability distribution at different positions of the probability distribution matrix can be calculated to obtain an entropy vector corresponding to each position. Then, the values in the entropy vector that are less than a set value are set to negative infinity. Finally, the entropy vector can be normalized to obtain the mutation probability at each position.

[0132] After calculating the entropy of the amino acid probability distribution at different positions of the probability distribution matrix to obtain the mutation probability corresponding to each position, sub-step D2 is performed.

[0133] Sub-step D2: according to the set number of mutations and the mutation probability, the probability distribution matrix is sampled to obtain the multi-site mutation sequence sample.

[0134] After calculating the entropy of the amino acid probability distribution at different positions of the probability distribution matrix to obtain the mutation probability corresponding to each position, the probability distribution matrix can be sampled according to the set number of mutations and the mutation probability to obtain the multi-site mutation sequence sample. The specific implementation process can be as follows: first, the positions of the set number of mutations can be extracted on the mutation probability according to the set number of mutations. Then, the amino acid types at the positions of the set number of mutations can be sampled on the probability distribution matrix. Finally, the extracted amino acid types and the natural amino acid types at other positions are spliced in the order of the positions to obtain the multi-site mutation sequence sample.

[0135] The above process can be described in detail in combination with the following examples.

[0136] The implementation process of the mutation number controllable sampling is as follows:

[0137] 1. According to the obtained probability distribution matrix E, the entropy of the probability distribution at different positions is calculated, and v E R is obtained 1×n ;

[0138] 2. The value less than ε in v is set to negative infinity. ε is a very small value, which can be manually set;

[0139] 3. The v vector is subjected to softmax to obtain the probability P at each position p The smaller the entropy, the more conservative the position is, and the less mutation should be performed at the position. Therefore, it is reasonable to use the size of the entropy to measure the mutation probability of each position selected;

[0140] 4. According to the desired number of mutations n, n positions are extracted on P p ;

[0141] 5. According to the matrix E, the amino acid types at the n corresponding positions are sampled. Then, the natural amino acids at other positions are spliced according to the positions to obtain the final sequence sample;

[0142] 6. The steps 4 and 5 are repeatedly performed to obtain more sequence samples.

[0143] The method for generating multi-mutation capsid sequences from single-mutation capsid sequences proposed in the embodiments of the application can be used in the design of any functional capsid. As long as there is data of a saturated single-mutation capsid under the function, the method can be used for sampling, and after sampling, a capsid containing the function can be generated. Meanwhile, the method for generating multi-mutation capsid sequences from single-mutation capsid sequences can not be limited to the adeno-associated virus capsid sequence, but can also be used in the design of any protein sequence.

[0144] In actual application, after obtaining all multi-site mutation sequence samples, the multi-site mutation sequence samples can be de-duplicated, and the number of mutations between each multi-site mutation sequence sample and the natural sequence can be calculated, and the multi-site mutation sequence samples with a number of mutations less than 2 are filtered out, so as to realize the post-processing process of the multi-site mutation sequence samples.

[0145] The multi-site mutation sequence sample acquisition method provided in this application involves performing saturated single mutation processing on various positions in a protein sequence to obtain saturated single-mutation sequences, and then obtaining the corresponding function value matrix. The function value matrix is ​​preprocessed to obtain a probability distribution matrix indicating the amino acid probability distribution at each position. Based on a specified sampling method, the probability distribution matrix is ​​sampled to obtain multi-site mutation sequence samples. This application embodiment constructs a series of sampling processes based on existing saturated single mutation data for desired capsid functions (functions may include, but are not limited to, active capsids, deimmunogenic capsids, or capsids targeting specific cells). The samples obtained from the amino acid probability distribution not only control the number of mutations between the sample and the natural capsid, but also have a significantly higher proportion of usable samples in the sample set compared to random mutations, and the generation efficiency is significantly better than rational artificial design.

[0146] Reference Figure 6 The diagram shows a schematic representation of a multi-site mutation sequence sample acquisition device provided in an embodiment of this application. Figure 6 As shown, the multi-site mutation sequence sample acquisition device 600 may include the following modules:

[0147] The function value matrix acquisition module 610 is used to perform saturation single mutation processing on the elements at each position in the protein sequence to obtain a saturation single mutation sequence, and to obtain the function value matrix corresponding to the saturation single mutation sequence.

[0148] The probability distribution matrix acquisition module 620 is used to preprocess the function value matrix to obtain a probability distribution matrix that indicates the probability distribution of amino acids at each position.

[0149] The multi-site mutation sequence sample acquisition module 630 is used to sample the probability distribution matrix based on a specified sampling method to obtain multi-site mutation sequence samples.

[0150] Optionally, the function value matrix acquisition module includes:

[0151] The saturated single mutant sequence acquisition unit is used to mutate any amino acid at each position in the protein sequence to obtain a saturated single mutant sequence.

[0152] A function value calculation unit is used to calculate the function value corresponding to the saturated single mutation sequence;

[0153] The function value matrix generation unit is used to generate the function value matrix corresponding to the saturated single mutation sequence based on the function value corresponding to each position.

[0154] Optionally, the probability distribution matrix acquisition module includes:

[0155] The intermediate functional value matrix acquisition unit is configured to replace the functional value of the amino acid type at the insertion position in the functional value matrix with 0, and replace the element with a functional value less than 0 in the functional value matrix with negative infinity, to obtain an intermediate functional value matrix.

[0156] The amino acid probability distribution acquisition unit is configured to normalize the functional value of different amino acids in the intermediate functional value matrix for different positions in the intermediate functional value matrix, to obtain an amino acid probability distribution at the corresponding position.

[0157] The probability distribution matrix generation unit is configured to generate a probability distribution matrix for indicating the amino acid probability distribution of each position according to the amino acid probability distribution at each position in the intermediate functional value matrix.

[0158] Optionally, the multi-site mutation sequence sample acquisition module comprises:

[0159] The matrix position traversal unit is configured to traverse each position in the probability distribution matrix in a case where the specified sampling manner is a sampling manner in which the number of mutations is uncontrollable.

[0160] The amino acid type acquisition unit is configured to sample the probability distribution matrix according to the amino acid probability distribution corresponding to each position that is traversed, to obtain an amino acid type corresponding to each position.

[0161] The first sequence sample acquisition unit is configured to splice the amino acid types corresponding to the positions, to obtain the multi-site mutation sequence sample.

[0162] Optionally, the multi-site mutation sequence sample acquisition module comprises:

[0163] The mutation probability acquisition unit is configured to calculate the entropy of the amino acid probability distribution at different positions in the probability distribution matrix in a case where the specified sampling manner is a sampling manner in which the number of mutations is controllable, to obtain a mutation probability corresponding to each position.

[0164] The second sequence sample acquisition unit is configured to sample the probability distribution matrix according to the set number of mutations and the mutation probability, to obtain the multi-site mutation sequence sample.

[0165] Optionally, the mutation probability acquisition unit comprises:

[0166] The entropy vector acquisition subunit is configured to calculate the entropy of the amino acid probability distribution at different positions in the probability distribution matrix, to obtain an entropy vector corresponding to each position.

[0167] The entropy vector value adjustment subunit is configured to set the value less than the set value in the entropy vector to negative infinity.

[0168] The mutation probability obtaining subunit is configured to normalize the entropy vector to obtain a mutation probability at each position.

[0169] Optionally, the second sequence sample obtaining unit comprises:

[0170] The mutation position extracting subunit is configured to extract the set number of positions from the mutation probability according to the set number of mutations.

[0171] The amino acid type obtaining subunit is configured to sample the amino acid types at the set number of positions from the probability distribution matrix.

[0172] The multi-site mutation sequence sample obtaining subunit is configured to splice the extracted amino acid types and the natural amino acid types at other positions according to the position sequence to obtain the multi-site mutation sequence sample.

[0173] The multi-site mutation sequence sample obtaining device provided by the embodiments of the present application obtains a saturated single mutation sequence by performing saturated single mutation processing on each position in a protein sequence, and obtains a function value matrix corresponding to the saturated single mutation sequence. The function value matrix is preprocessed to obtain a probability distribution matrix for indicating the amino acid probability distribution of each position. The probability distribution matrix is sampled based on a specified sampling method to obtain a multi-site mutation sequence sample. According to the existing saturated single mutation data of the required capsid function (the function can include but is not limited to a capsid with activity, a capsid with deimmunogenicity, and a capsid with targeting of a certain specific cell), the embodiments of the present application construct a series of sampling processes, and finally obtain samples from the amino acid probability distribution. The number of mutations between the obtained samples and the natural capsid can be controlled, the proportion of available samples in the sample set is significantly better than that of random mutations, and the generation efficiency is significantly better than that of rational artificial design.

[0174] The embodiments of the present application provide an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the computer program is executed by the processor to implement the multi-site mutation sequence sample obtaining method.

[0175] Figure 7 The structure schematic diagram of an electronic device 700 of the embodiments of the present application is shown. Figure 7As shown, the electronic device 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 702 or loaded into a random access memory (RAM) 703 from a storage unit 708. Various programs and data required by the electronic device 700 to operate can also be stored in the RAM 703. The CPU 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0176] A plurality of components in the electronic device 700 are connected to the I / O interface 705, including an input unit 706, such as a keyboard, a mouse, a microphone, etc., an output unit 707, such as various types of displays, a speaker, etc., a storage unit 708, such as a magnetic disk, an optical disk, etc., and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the electronic device 700 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0177] The various processes and processes described above can be performed by the processing unit 701. For example, the method of any of the embodiments described above can be implemented as a computer software program tangibly embodied in a computer readable medium, such as the storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the CPU 701, one or more actions of the method described above can be performed.

[0178] The embodiments of the present application provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the above-mentioned multi-site mutation sequence sample acquisition method.

[0179] Each embodiment in the present specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same and similar parts between each embodiment can be referred to each other.

[0180] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, device, or computer program product. Therefore, the embodiments of the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0181] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, terminals (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminals to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminals generate a device implemented in the flowcharts Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in one or more flows or blocks.

[0182] These computer program instructions can also be stored in a computer readable memory capable of directing the computer or other programmable data processing terminal to work in a specific manner, so that the instructions stored in the computer readable memory produce a product including instruction devices, which implement the flowcharts Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in one or more flows or blocks.

[0183] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal, so that a series of operation steps are performed on the computer or other programmable terminal to produce a computer implemented process, so that the instructions executed on the computer or other programmable terminal provide a process for implementing the flowcharts Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in one or more flows or blocks.

[0184] Although the preferred embodiments of the present application have been described, those skilled in the art can make further changes and modifications to these embodiments once they know the basic inventive concept. Therefore, the appended claims are intended to be interpreted as including all changes and modifications falling within the scope of the embodiments of the present application.

[0185] The above provides a multi-site mutation sequence sample acquisition method, a multi-site mutation sequence sample acquisition device, an electronic device, and a computer readable storage medium, which are described in detail in the present application. The principles and implementation modes of the present application are described by applying specific examples in this paper. The above embodiment description is only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for obtaining multi-site mutation sequence samples, characterized in that, The method comprises: Saturated single mutation processing is performed on elements of each position in a protein sequence to obtain a saturated single mutation sequence, and a function value matrix corresponding to the saturated single mutation sequence is obtained; The function value matrix is preprocessed to obtain a probability distribution matrix for indicating an amino acid probability distribution of each position; Sampling processing is performed on the probability distribution matrix based on a specified sampling mode to obtain a multi-site mutation sequence sample; The function value matrix is preprocessed to obtain a probability distribution matrix for indicating an amino acid probability distribution of each position, comprising: The function value of the amino acid type at the insertion position in the function value matrix is replaced with 0, and the elements with a function value less than 0 in the function value matrix are replaced with negative infinity to obtain an intermediate function value matrix; The function values of different amino acids in the intermediate function value matrix are normalized for different positions of the intermediate function value matrix to obtain an amino acid probability distribution at the corresponding position; According to the amino acid probability distribution at each position in the intermediate function value matrix, a probability distribution matrix for indicating the amino acid probability distribution of each position is generated.

2. The method of claim 1, wherein, The saturated single mutation processing is performed on the elements of each position in the protein sequence to obtain a saturated single mutation sequence, and the function value matrix corresponding to the saturated single mutation sequence is obtained, comprising: Arbitrary amino acid mutation processing is performed on the elements of each position in the protein sequence to obtain a saturated single mutation sequence; The function value corresponding to the saturated single mutation sequence is calculated; According to the function value corresponding to each position, a function value matrix corresponding to the saturated single mutation sequence is generated.

3. The method of claim 1, wherein, The sampling processing is performed on the probability distribution matrix based on the specified sampling mode to obtain the multi-site mutation sequence sample, comprising: In the case that the specified sampling mode is a mutation number uncontrollable sampling mode, each position in the probability distribution matrix is traversed; According to the amino acid probability distribution corresponding to each traversed position, the probability distribution matrix is sampled to obtain an amino acid type corresponding to each position; The amino acid types corresponding to the positions are spliced to obtain the multi-site mutation sequence sample.

4. The method of claim 1, wherein, The sampling processing is performed on the probability distribution matrix based on the specified sampling mode to obtain the multi-site mutation sequence sample, comprising: In the case that the specified sampling mode is a mutation number controllable sampling mode, the entropy of the amino acid probability distribution under different positions of the probability distribution matrix is calculated to obtain a mutation probability corresponding to each position; According to the set number of mutations and the mutation probability, the probability distribution matrix is sampled to obtain the multi-site mutation sequence sample.

5. The method of claim 4, wherein, The entropy of the amino acid probability distribution under different positions of the probability distribution matrix is calculated to obtain a mutation probability corresponding to each position, comprising: The entropy of the amino acid probability distribution under different positions of the probability distribution matrix is calculated to obtain an entropy vector corresponding to each position; Values less than a set value in the entropy vector are set to negative infinity; The entropy vector is normalized to obtain a mutation probability under each position.

6. The method of claim 4, wherein, The sampling processing on the probability distribution matrix according to the set number of mutations and the mutation probability includes: extracting the set number of positions on the mutation probability according to the set number of mutations; sampling to obtain the amino acid types at the set number of positions on the probability distribution matrix; splicing the extracted amino acid types and the natural amino acid types at other positions according to the position sequence to obtain the multi-site mutation sequence sample.

7. A multi-site mutant sequence sample acquisition device, comprising: The device includes: a function value matrix acquisition module configured to perform saturated single mutation processing on elements at each position in a protein sequence to obtain a saturated single mutation sequence, and acquire a function value matrix corresponding to the saturated single mutation sequence; a probability distribution matrix acquisition module configured to pre-process the function value matrix to obtain a probability distribution matrix indicating an amino acid probability distribution at each position; a multi-site mutation sequence sample acquisition module configured to perform sampling processing on the probability distribution matrix based on a specified sampling method to obtain the multi-site mutation sequence sample; The probability distribution matrix acquisition module includes: an intermediate function value matrix acquisition unit configured to replace a function value of an empty amino acid type at an insertion position in the function value matrix with 0, and replace elements with function values less than 0 in the function value matrix with negative infinity to obtain an intermediate function value matrix; an amino acid probability distribution acquisition unit configured to perform normalization processing on function values of different amino acids in the intermediate function value matrix for different positions in the intermediate function value matrix to obtain an amino acid probability distribution at a corresponding position; a probability distribution matrix generation unit configured to generate a probability distribution matrix indicating an amino acid probability distribution at each position according to the amino acid probability distribution at each position in the intermediate function value matrix.

8. The apparatus of claim 7, wherein, The function value matrix acquisition module includes: a saturated single mutation sequence acquisition unit configured to perform mutation processing on elements at each position in the protein sequence with any amino acid to obtain a saturated single mutation sequence; a function value calculation unit configured to calculate a function value corresponding to the saturated single mutation sequence; a function value matrix generation unit configured to generate a function value matrix corresponding to the saturated single mutation sequence according to the function values corresponding to the positions.

Images