Fitness space roughness evaluation method and device, electronic equipment and storage medium

Abstract

The application provides a roughness evaluation method and device of fitness space, electronic equipment and storage medium, and relates to the technical field of biology. The method comprises the following steps: acquiring a data set; the data set comprises a plurality of biological sequences; based on the mutation sites between each biological sequence in the data set, a plurality of adjacent sequence pairs in the data set are determined; the mutation sites between the two biological sequences in each adjacent sequence pair meet the set number of mutation sites; and the roughness of the fitness space is estimated according to each adjacent sequence pair, to obtain a roughness evaluation result. The application solves the problem of inaccurate roughness evaluation of the fitness space in the related art.

Description

Technical Field

[0001] This application relates to the field of biotechnology, and more specifically, to a method, apparatus, electronic device, and storage medium for evaluating the roughness of fitness space. Background Technology

[0002] In biological research, an increasing amount of data is being used to study the complexity and diversity of organisms. Biological sequences are a crucial part of this, including DNA sequences, protein sequences, and RNA sequences. These biological sequences describe the spatial and temporal variations of molecules such as genes and proteins within an organism, and are of great significance for understanding the nature of life and the function and evolution of biological systems.

[0003] A biological sequence fitness space is used to describe the adaptability of biological sequences in different environments. Roughness assessment of the fitness space quantifies fitness to better understand the behavior of biological sequences in their adapted environments. However, due to the complexity of the biological sequence fitness space and the uncertainty of the data, roughness assessments are often prone to errors.

[0004] Incomplete or erroneous datasets can lead to inaccurate representations of the fitness space, resulting in errors in the roughness evaluation. For example, missing key biological sequences in the dataset can cause errors in the roughness evaluation of the fitness space. When the biological sequences are long, their number can be astronomical. Therefore, the resulting fitness space will have a large number of missing biological sequences compared to the complete space. Due to these missing sequences, the fitness space itself has already lost a significant amount of information contained in the biological sequences compared to the original fitness space without missing sequences. In other words, the systematic error in the roughness evaluation result cannot be eliminated, but can only be appropriately reduced.

[0005] As can be seen from the above, the large error in the roughness evaluation results of the fitness space has become an urgent problem to be solved. Summary of the Invention

[0006] This application provides a method, apparatus, electronic device, and storage medium for evaluating the roughness of fitness space, which can solve the problem of large errors in the roughness evaluation results of biological sequence fitness space in related technologies. The technical solutions are as follows:

[0007] According to one aspect of this application, a dataset is obtained; the dataset includes multiple biological sequences; multiple adjacent sequence pairs in the dataset are determined based on mutation sites between the biological sequences in the dataset; the mutation sites between the two biological sequences in each adjacent sequence pair meet a set number of mutation sites; the roughness of the fitness space is estimated based on each adjacent sequence pair to obtain a roughness evaluation result.

[0008] According to one aspect of this application, a dataset acquisition module is used to acquire a dataset; the dataset includes multiple biological sequences; a sequence pair confirmation module is used to determine multiple adjacent sequence pairs in the dataset based on mutation sites between the biological sequences in the dataset; the mutation sites between the two biological sequences in each adjacent sequence pair conform to a set number of mutation sites; and a roughness evaluation module is used to estimate the roughness of the fitness space based on each adjacent sequence pair to obtain a roughness evaluation result.

[0009] According to one aspect of this application, an electronic device includes at least one processor and at least one memory, wherein program instructions or code are stored in the memory; the program instructions or code are loaded and executed by the processor, causing the electronic device to implement the roughness evaluation method of the fitness space as described above.

[0010] According to one aspect of this application, a storage medium stores program instructions or code thereon, which are loaded and executed by a processor to implement a roughness evaluation method for fitness space as described above.

[0011] According to one aspect of this application, an application product includes program instructions or code stored in a storage medium. The processor of an electronic device reads the program instructions or code from the storage medium, loads and executes the program instructions or code, causing the electronic device to implement the roughness evaluation method of the fitness space as described above.

[0012] The beneficial effects of the technical solution provided in this application are:

[0013] In the above technical solution, by determining the adjacent sequence pairs in the dataset, the roughness evaluation result is obtained based on each adjacent sequence pair, avoiding the abnormal smoothing effect on the edit distance caused by non-adjacent sequences in the dataset when biological sequences are missing; adjacent sequences can maintain an approximate roughness before and after the biological sequence is missing. By calculating adjacent sequences, the roughness evaluation result can be minimized from being affected by the biological sequence missing, ensuring the accuracy of the roughness evaluation result, thereby effectively solving the problem of large error in the roughness evaluation result of the biological sequence fitness space in related technologies. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.

[0015] Figure 1 This is a schematic diagram based on the implementation environment involved in this application;

[0016] Figure 2This is a flowchart illustrating a roughness evaluation method for fitness space according to an exemplary embodiment;

[0017] Figure 3 This is a flowchart illustrating another method for roughness evaluation in fitness space according to an exemplary embodiment;

[0018] Figure 3a yes Figure 3 Schematic diagrams illustrating the specific implementations of the two methods for obtaining adjacent sequence pairs in the corresponding embodiments;

[0019] Figure 3b yes Figure 3 A schematic diagram illustrating the specific implementation of the difference in computational fitness and the standardized fitness involved in the corresponding embodiment;

[0020] Figure 3c Figure 3 A schematic diagram illustrating a specific implementation of grouping based on mutation sites in the corresponding embodiment;

[0021] Figure 4 This is a schematic diagram illustrating the specific implementation of a roughness evaluation method for fitness space in an application scenario.

[0022] Figure 4a yes Figure 4 Roughness evaluation results of the extreme value method in application scenarios on datasets A0-A5;

[0023] Figure 4b yes Figure 4 Roughness evaluation results of the fitting method in the application scenario on dataset A0-A5;

[0024] Figure 4c yes Figure 4 Roughness evaluation results of a fitness space roughness evaluation method in an application scenario on dataset A0-A5;

[0025] Figure 5 This is a structural block diagram of a roughness evaluation device in fitness space according to an exemplary embodiment;

[0026] Figure 6 This is a hardware structure diagram of an electronic device according to an exemplary embodiment;

[0027] Figure 7 This is a structural block diagram of an electronic device according to an exemplary embodiment. Detailed Implementation

[0028] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0029] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0030] The following is an introduction and explanation of several terms used in this application:

[0031] Protein fitness refers to the strength of a protein's functional characteristics as measured under a certain functional characteristic standard; it is generally a linear and continuous value.

[0032] The protein fitness space refers to the dataset obtained by combining all mutations of a protein sequence of the same length with their fitness values ​​into a complete sequence space-fitness key-value pair.

[0033] A protein fitness dataset refers to a collection of data sampled from the protein fitness space, and its size is usually much smaller than the entire protein fitness space.

[0034] Protein roughness refers to the ruggedness of a protein dataset. While protein function is generally determined by one or more sites, complex higher-order interactions may exist between these sites. Roughness is used to assess the magnitude of these higher-order interactions within a protein's fitness space. Higher roughness indicates more complex higher-order interactions within the protein's fitness space.

[0035] The abnormal smoothness effect ignores the differences between biological sequences when evaluating the roughness of biological sequence fitness space, resulting in inaccurate roughness evaluation results.

[0036] As mentioned earlier, the complexity of the biological sequence fitness space and the uncertainty of the data can lead to errors in the roughness assessment results.

[0037] Taking proteins as an example, proteins are organic macromolecules, the most basic building blocks of cells, and a major component of the human body and an important part of food. In nature, naturally occurring proteins are called wild-type proteins. Because some functional properties of natural proteins are unsatisfactory, mutations are induced at certain sites in wild-type proteins to obtain new protein sequences. The functional properties of these mutated proteins are then determined; these proteins are called mutant proteins corresponding to the wild type. The desired functional properties are called the protein's fitness.

[0038] Combining wild-type and mutant proteins, along with their functional characteristics, yields the protein sequence fitness space. Because the number of different mutants can be astronomical when a protein sequence is long, the resulting fitness space contains a large number of missing protein sequences compared to the complete space.

[0039] Currently, either the extreme value method or the fitting method can be used to evaluate the roughness of the protein sequence fitness space. The extreme value method uses the ratio of the number of local extreme sequences to the total number of sequences in the protein sequence fitness space to obtain the roughness evaluation result. The fitting method constructs a linear regression model for the protein sequence fitness space and uses the fitting error to obtain the roughness evaluation result.

[0040] However, because these methods are designed for short protein sequences, they can only obtain accurate roughness assessment results when the protein sequence fitness space is complete. When the protein sequence fitness space is missing, the roughness assessment results will contain a large number of systematic errors. However, the true protein fitness space must contain a large number of missing protein sequences; therefore, the roughness assessment results of the above methods are unreliable.

[0041] As can be seen from the above, the related technologies still suffer from large errors and inaccuracies in the roughness evaluation results of the fitness space.

[0042] Therefore, the fitness space roughness evaluation method provided in this application can effectively reduce the error in fitness space roughness evaluation and improve the accuracy of roughness evaluation results. Accordingly, the fitness space roughness evaluation method is applicable to fitness space roughness evaluation devices, and the fitness space roughness evaluation devices can be deployed on electronic devices configured with von Neumann architecture, such as desktop computers, laptops, servers, etc.

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0044] Figure 1 This is a schematic diagram of an implementation environment involved in a roughness evaluation method for fitness space. It should be noted that this implementation environment is merely an example adapted to the present invention and should not be considered as providing any limitation on the scope of the invention.

[0045] The implementation environment includes a data acquisition terminal 110 and a server terminal 130.

[0046] Specifically, the acquisition end 110 can also be considered as a device capable of acquiring biological sequences, including but not limited to protein sequencers, DNA sequence analysis electrophoresis instruments, etc. For example, the acquisition end 110 is a protein sequencer.

[0047] Server 130 can be an electronic device such as a desktop computer, laptop computer, or server, or a computer cluster consisting of multiple servers, or even a cloud computing center consisting of multiple servers. Server 130 is used to provide backend services, such as, but not limited to, roughness evaluation services for the fitness space.

[0048] The server 130 and the acquisition terminal 110 establish a network communication connection in advance via wired or wireless means, and data transmission between the server 130 and the acquisition terminal 110 is realized through this network communication connection. The transmitted data includes, but is not limited to, datasets containing multiple biological sequences, etc.

[0049] In one application scenario, through the interaction between the acquisition terminal 110 and the server terminal 130, the acquisition terminal 110 collects biological sequences to obtain a dataset and uploads the dataset to the server terminal 130 to request the server terminal 130 to provide a roughness evaluation service for the fitness space.

[0050] For server 130, after obtaining the dataset uploaded by acquisition terminal 110, it calls the roughness evaluation service of fitness space. Based on the mutation sites between biological sequences in the dataset, it determines multiple adjacent sequence pairs in the dataset, estimates the roughness of fitness space based on each adjacent sequence pair, and obtains the roughness evaluation result. This solves the problem of large error and inaccuracy in the roughness evaluation result of biological sequence fitness space in related technologies.

[0051] Please see Figure 2 This application provides a method for roughness evaluation in fitness space, applicable to electronic devices, which may be... Figure 1 The server 130 in the implementation environment is shown.

[0052] In the following method embodiments, for ease of description, the execution subject of each step of the method is an electronic device, but this does not constitute a specific limitation.

[0053] like Figure 2 As shown, the method may include the following steps:

[0054] Step 310: Obtain the dataset.

[0055] The dataset includes multiple biological sequences, all of which belong to the same type. The type of biological sequence can be protein sequence, DNA sequence, or RNA sequence, and is not limited here.

[0056] In one possible implementation, if the experimenter wants to perform a roughness assessment of the protein fitness space, then the dataset includes multiple protein sequences.

[0057] It should be noted that the dataset can include not only multiple biological sequences, but also the fitness of each biological sequence. In one possible implementation, the dataset is a tabular dataset with at least two columns. If the biological sequences in the dataset are protein sequences, then the table columns are the amino acid composition (sequence) and the fitness (fitness) of the protein sequence. After the dataset is read in, it will be stored in list form. Each list element is a tuple consisting of the sequence and fitness.

[0058] Step 330: Based on the mutation sites between biological sequences in the dataset, determine multiple adjacent sequence pairs in the dataset.

[0059] In each adjacent sequence pair, the mutation sites between the two biological sequences meet the set number of mutation sites. The adjacent sequence pair includes two biological sequences, which can be composed of the biological sequence before mutation and the biological sequence after mutation, without limitation.

[0060] First, it should be noted that biological sequences are not entirely different; mutation sites can be considered as the points of difference between biological sequences. Mutation sites can arise from natural variation or from artificial intervention. Taking protein sequences as an example, because some functional characteristics of naturally occurring protein sequences are not satisfactory, researchers induce mutations at certain sites in wild-type proteins to obtain new mutant protein sequences to meet human needs. In other words, the differences between the natural protein sequence and the mutant protein are the sites where induced mutations occur, i.e., mutation sites. It should be noted that the above methods are provided as examples only and do not constitute a limitation of the present invention.

[0061] In other words, when the mutation site between two biological sequences is 1, it means that the two biological sequences have one different site. However, since the roughness evaluation result is calculated using non-adjacent matrix pairs, an abnormal smoothing effect may occur, affecting the accuracy of the roughness evaluation result. Therefore, in order to avoid the abnormal smoothing effect, the number of mutation sites can be set as small as possible. In one possible implementation, the number of mutation sites is set to 1. When the mutation site between two biological sequences is 1, these two biological sequences can form an adjacent sequence pair.

[0062] Edit distance is used to indicate the similarity between biological sequences. In one possible implementation, the edit distance can represent the number of mutation sites between two biological sequences. For example, if biological sequence A is "AAA" and biological sequence B is "AAB", then the edit distance between them is 1. Furthermore, a distance threshold is set to determine adjacent sequence pairs. If the distance threshold is set to "edit distance equals 1", then the edit distance between two biological sequences is 1, and adjacent sequence pairs can be obtained from the two biological sequences.

[0063] Step 350: Estimate the roughness of the fitness space based on each adjacent sequence pair to obtain the roughness evaluation result.

[0064] First, it should be noted that fitness space refers to the dataset obtained by forming a complete sequence space-fitness key-value pairs consisting of all mutation sites of biological sequences of the same length and their fitness; roughness refers to the degree of surface features of fitness space, that is, the degree of irregularity or complexity of the surface of fitness space. The higher the roughness, the more complex the higher-order interactions will be in fitness space.

[0065] For example, if the biological sequences in the dataset are protein sequences, then the protein fitness space can be obtained based on all mutation sites of protein sequences of the same length and the fitness of each protein sequence. The roughness of the protein fitness space is the degree of irregularity or complexity of the protein fitness space, which is estimated from the protein fitness space.

[0066] As mentioned earlier, the dataset can include not only multiple biological sequences, but also the fitness of each biological sequence. Based on this, the fitness space of the dataset can be estimated based on each adjacent sequence pair and its fitness to obtain the roughness evaluation result.

[0067] Through the above process, by identifying adjacent sequence pairs in the dataset, roughness evaluation results are obtained based on each adjacent sequence pair. This avoids the abnormal smoothing effect on the edit distance caused by non-adjacent sequences in the dataset when biological sequences are missing. Adjacent sequences can maintain an approximate roughness before and after the biological sequence is missing. The roughness evaluation results obtained through adjacent sequences can minimize the impact of biological sequence missing on the roughness evaluation results, ensuring the accuracy of the roughness evaluation results. This effectively solves the problem of large errors in the roughness evaluation results of the biological sequence fitness space in related technologies.

[0068] Please see Figure 3 This application provides a method for roughness evaluation in fitness space, applicable to electronic devices, which can be... Figure 1 The server 130 in the implementation environment is shown.

[0069] Step 4010: Obtain the dataset.

[0070] First, it should be noted that, depending on the mutation status, the biological sequence may be either unmutated or mutated. In order to enrich the dataset and make the roughness assessment results more accurate, in one possible implementation, the dataset may include a mutation dataset, which may be a protein mutation dataset from deep mutation scanning, such as avgfp, gb1, etc.

[0071] It should be noted that this method can be used to evaluate the roughness of various types of biological sequences. In one possible implementation, the biological sequence can be evaluated using this method if it meets the following conditions:

[0072] 1. Biological sequences are composed of substructures.

[0073] 2. The types of substructures are limited.

[0074] 3. Biological sequences have a limited length.

[0075] 4. Each biological sequence possesses a fitness level measured under a certain standard.

[0076] 5. Multiple biological sequences of the same length and their fitness can be sampled to construct a dataset.

[0077] Step 4030: Determine whether the number of biological sequences in the dataset exceeds the set threshold.

[0078] Since the number of biological sequences in the dataset varies, and the number of biological sequences affects the speed of calculating all adjacent sequence pairs in the dataset, in order to speed up the acquisition of adjacent sequence pairs and thus speed up the acquisition of roughness evaluation results, different calculation methods are used to process the dataset depending on whether the number of biological sequences in the dataset exceeds a set threshold.

[0079] The quantity threshold can be adjusted adaptively according to the actual situation. For example, if the electronic device has strong computing power, the quantity threshold can be set to be larger, and conversely, if the electronic device has strong computing power, the quantity can be set to be smaller.

[0080] If the set threshold is not exceeded, proceed to step 4050 to determine adjacent sequence pairs based on the edit distance of each biological sequence in the dataset.

[0081] The dataset contains multiple biological sequences. It can be understood that two biological sequences constitute a biological sequence pair. The edit distance of each biological sequence pair in the dataset can be calculated, and the biological sequence pairs whose edit distance meets the set distance threshold are selected as adjacent sequence pairs, thereby obtaining all adjacent sequence pairs in the dataset.

[0082] The time complexity of this method is related to the number of biological sequences in the dataset, and the time complexity is O(n). , where n is the number of biological sequences in the dataset. This method is suitable for datasets with a small number of biological sequences.

[0083] Specifically, in one possible implementation, combining Figure 3a As described in 3a1, step 4050 may include the following steps:

[0084] Step 4051: Based on the current biological sequence in the dataset, calculate the edit distance between the current biological sequence and the other biological sequences in the dataset to obtain the edit distance between each biological sequence.

[0085] Step 4053: For the current biological sequence, select biological sequences from the remaining biological sequences that have an edit distance that meets the set distance threshold.

[0086] Step 4055: Obtain the neighbor sequence pair corresponding to the current biological sequence from the current biological sequence and the filtered biological sequences.

[0087] Step 4057 continues until all biological sequences have been screened and adjacent sequence pairs are obtained.

[0088] If the set number threshold is exceeded, then step 4070 is executed. If the set number threshold is exceeded, then adjacent sequence pairs are determined based on the candidate sequences corresponding to each biological sequence in the dataset.

[0089] Candidate sequences refer to biological sequences generated using a biological sequence as the parent sequence, whose edit distance from the parent sequence meets a set distance threshold. In other words, by using a biological sequence in the dataset as the parent sequence, multiple candidate sequences with edit distances meeting the set distance threshold can be generated. For example, if the parent sequence is "BAA" and the distance threshold is 1, then three candidate sequences can be generated: "AAA", "BBA", and "BAB".

[0090] It is understandable that candidate sequences may form adjacent sequence pairs with parent biological sequences. However, not all candidate sequences have the same biological sequence in the dataset. If a candidate sequence is successfully found in the dataset, then the same biological sequence exists in the dataset, and adjacent sequence pairs can be obtained based on the candidate sequence. If a candidate sequence fails to be found in the dataset, then the same biological sequence does not exist in the dataset, and adjacent sequence pairs cannot be obtained based on the candidate sequence.

[0091] For example, if the parent sequence is "BAA", and the three candidate sequences are "AAA", "BBA" and "BAB", only "AAA" is successfully found in the dataset, thus obtaining the adjacent sequence pair "BAA-AAA".

[0092] Furthermore, for each biological sequence in the dataset, corresponding candidate sequences are generated, and each candidate sequence is searched in the dataset. Based on the successfully found candidate sequences and their corresponding biological sequences, adjacent sequence pairs are obtained, thus obtaining all adjacent sequence pairs in the dataset.

[0093] The time complexity of this method is related to the number of biological sequences in the dataset, and the time complexity is O(n). Where n is the number of biological sequences in the dataset and k is the length of the biological sequence, the constant term of this method is relatively large and positively correlated with the number of candidate sequences. This method is suitable for datasets with a large number of biological sequences.

[0094] It should be noted that before generating candidate sequences, the dataset can be transformed into a set structure based on the biological sequences. Then, the biological sequences in the dataset are iterated and all possible candidate sequences for each biological sequence are generated. It is understandable that, due to the large size of the dataset, the set structure usually has less space complexity than multidimensional arrays or matrices, because the set only contains different elements. Therefore, when the dataset is large, transforming the dataset into a set structure can effectively reduce processing time and memory consumption, and improve processing efficiency.

[0095] Specifically, in one possible implementation, combining Figure 3a As explained in 3a2, step 4070 includes the following steps:

[0096] Step 4071: For a candidate sequence corresponding to a current biological sequence, search the dataset to see if there is a biological sequence that is the same as the corresponding candidate sequence.

[0097] Step 4073: If the search is successful, then based on the current biological sequence and the found biological sequence, obtain the adjacent sequence pair corresponding to the current biological sequence.

[0098] Step 4075 continues until all biological sequences have been searched, and each adjacent sequence pair is obtained.

[0099] Step 4090: Calculate the first fitness and the second fitness for each adjacent sequence pair to obtain the fitness difference for each adjacent sequence pair.

[0100] As mentioned earlier, a neighboring sequence pair includes two biological sequences. For example, a neighboring sequence pair includes a biological sequence before mutation and a biological sequence after mutation.

[0101] In this context, the two biological sequences of an adjacent sequence pair have corresponding fitness values. In one possible implementation, before step 4090, the process may include obtaining a first fitness value and a second fitness value for each adjacent sequence pair, wherein the first fitness value corresponds to one biological sequence of the adjacent sequence pair, and the second fitness value corresponds to the other biological sequence of the adjacent sequence pair.

[0102] The fitness difference refers to the difference in fitness between two biological sequences in an adjacent sequence pair. The calculation formula is shown in equation (1):

[0103] ...Equation (1)

[0104] in, The difference in fitness between adjacent sequence pairs. As the first sequence, For the second sequence, This represents the fitness corresponding to the biological sequence.

[0105] For example, Figure 3b At 3b1, a method for calculating the fitness difference is shown, such as... Figure 3b As shown in 3b1, the fitness difference between adjacent sequence pairs “AAA-AAB” is 1.0, the fitness difference between adjacent sequence pairs “ABA-ABB” is 2.0, and the fitness difference between adjacent sequence pairs “BBA-BBB” is 3.0.

[0106] The fitness difference of each adjacent sequence pair is obtained by calculating according to Equation (1). The roughness evaluation result can be obtained based on the fitness difference of each adjacent sequence pair. However, before the calculation, each adjacent sequence pair can be grouped.

[0107] Step 4110: Generate at least one set of adjacent sequence groups based on the labels corresponding to each adjacent sequence pair.

[0108] In this context, adjacent sequence pairs in a neighboring sequence group share the same tag. The tag is used to indicate the mutation characteristics of each adjacent sequence pair in the neighboring sequence group. In one possible implementation, the tag can be obtained through the mutation site. For example, if the biological sequence is a protein sequence, the corresponding tag consists of "mutation site + pre-mutation amino acid + post-mutation amino acid".

[0109] It is understandable that each pair of adjacent sequences corresponds to a unique label. Conversely, each label may correspond to zero, one, or more adjacent sequence pairs, depending on whether there is a mutation corresponding to the label in the dataset. Based on this, an adjacent sequence group can be obtained from the adjacent sequence pairs of the same label, and the adjacent sequence pairs in the same group will have the same label, thus generating at least one adjacent sequence group from each adjacent sequence pair.

[0110] For example, Figure 3c The diagram illustrates a specific implementation of grouping based on mutation sites, as shown below. Figure 3c As shown, the tag 1 of adjacent sequence group 1 is "site 3 mutates from A to B". Therefore, if the mutation characteristics of the two biological sequences in an adjacent sequence pair match tag 1, then the adjacent sequence pair belongs to group 1. Based on this, adjacent sequence pairs "AAA-AAB", "ABA-ABB", and "BBA-BBB" all belong to group 1. For example... Figure 3c As shown, the tag 2 of adjacent sequence group 2 is "site 2 mutates from B to A". Therefore, if the mutation characteristics of the two biological sequences in the adjacent sequence pair match tag 2, then the adjacent sequence pair belongs to group 2. Based on this, the adjacent sequence pairs "ABA-AAA, BBA-BAB, BBB-BAB" all belong to group 2.

[0111] All adjacent sequence pairs are grouped according to their tags, forming adjacent sequence groups. Adjacent sequence pairs within the same group will have the same tag. Each adjacent sequence group may contain one or more adjacent sequence pairs. In addition to recording the biological sequences of adjacent sequence pairs, the fitness difference between adjacent sequence pairs is also recorded within the same adjacent sequence group.

[0112] Step 4130: Standardize the difference in fitness recorded in each adjacent sequence group to obtain the standardized fitness corresponding to each adjacent sequence group.

[0113] For each adjacent sequence group, the fitness difference recorded in it is standardized. The calculation method is to calculate the average of all fitness differences recorded in the adjacent sequence group, and then subtract this average from each fitness difference to get the updated fitness difference.

[0114] The standardized formulas are shown in equations (2) and (3):

[0115] ...Equation (2)

[0116] ...Equation (3)

[0117] Among them, the adjacent sequence pairs labeled L constitute an adjacent sequence group. The difference in fitness among all adjacent sequence groups. Let be the mean of the fitness differences between adjacent sequence groups, and count be the counting function.

[0118] For example, see 3b, such as Figure 3b As shown in 3b2, the difference in fitness among adjacent sequence pairs is standardized. The standardized fitness of adjacent sequence pair “AAA-AAB” is -1.0, the standardized fitness of adjacent sequence pair “ABA-ABB” is 0.0, and the standardized fitness of adjacent sequence pair “BBA-BBB” is 1.0.

[0119] In one possible implementation, step 4130 may include standardizing the fitness difference of each adjacent sequence pair in each adjacent sequence group, and updating the fitness difference of each adjacent sequence pair using the standardized fitness difference. That is, the updated fitness difference of each adjacent sequence pair is calculated using the above equations (2) and (3).

[0120] Step 4150: Merge each adjacent sequence group and calculate the roughness evaluation result based on the normalized fitness corresponding to each adjacent sequence group.

[0121] After all adjacent sequence groups have been updated as described above, all adjacent sequence groups are merged, and the variance of the difference between all fitness levels is calculated. This variance is the roughness evaluation result of the fitness space. The roughness estimation formula is shown in equation (4):

[0122] ...Equation (4)

[0123] Where E is the roughness evaluation result, and var is the variance function.

[0124] Under the above embodiments, firstly, for datasets of different sizes, using different methods to calculate adjacent sequence pairs can reduce the amount of computation and speed up the computation, thereby ensuring the efficiency of roughness evaluation; secondly, by calculating adjacent sequences, the roughness evaluation results can be minimized from being affected by biological sequence deletions, ensuring the accuracy of the roughness evaluation results in the fitness space; and thirdly, this method has good scalability and can be applied to roughness evaluation in the fitness space of various biological sequences, such as DNA sequences, RNA sequences, and protein sequences.

[0125] Figure 4 This is a schematic diagram illustrating a specific implementation of a roughness evaluation method for fitness space in an application scenario. In this application scenario, the dataset includes protein sequences.

[0126] First, it should be noted that in this application scenario, roughness evaluation is performed on the same dataset using two roughness evaluation methods and the roughness evaluation method of this scheme respectively. By comparing the errors of the three roughness evaluation results, the roughness evaluation effect obtained by this method is verified.

[0127] like Figure 4 As shown, the steps for roughness evaluation using the extreme value method are as follows:

[0128] In step 501, all protein sequences are sampled using the data sampling module.

[0129] In step 503, the extreme points are calculated using the calculation module.

[0130] Step 505 calculates the proportion of extreme points.

[0131] Step 507 outputs the roughness evaluation results.

[0132] like Figure 4 As shown, the steps for roughness evaluation using the fitting method are as follows:

[0133] In step 601, all protein sequences are sampled using the data sampling module.

[0134] In step 603, a linear simulation fitting is performed using the calculation module.

[0135] In step 605, the fitting error and slope are calculated.

[0136] Step 507 outputs the roughness evaluation results.

[0137] like Figure 4 As shown, the steps for roughness evaluation using the fitting method are as follows:

[0138] Step 701 involves sampling adjacent sequence pairs in the dataset.

[0139] In step 703, each adjacent sequence pair is grouped based on the label to obtain at least one adjacent sequence group.

[0140] Step 705 standardizes the fitness difference among adjacent sequence groups.

[0141] Step 707 involves combining and calculating the variance.

[0142] Step 507 outputs the roughness evaluation results.

[0143] It should be noted that the dataset was preprocessed before evaluation, as shown in Table 1, where N is the number of protein sequences. A virtual original fitness space was constructed using the synthetic dataset, and the original fitness space was used to form dataset A0 to evaluate the degree of deviation of different schemes from roughness evaluation when data is missing.

[0144] By simulating the random discarding of 10%, 30%, 50%, 70%, and 90% of the sequences in A0, five new datasets A1-A5 were obtained. However, considering that real datasets from deep mutation scans tend to sample more near wild-type proteins and less further away from them, a similar sampling scheme needs to be simulated to construct a dataset suite for testing.

[0145] Furthermore, using the dataset obtained by simulating deep mutation scanning, wild-type proteins were randomly selected, and 0%, 50%, 99%, and 99.99% of the proteins were discarded in subspaces with edit distances of 1, 2, 3, 4, 5, and above from the wild-type proteins, respectively, resulting in a new dataset A6.

[0146] To ensure the reliability of the test structure, multiple different random discarding operations were performed on A1-A5, resulting in 5 different random copies. For A6, 5 random copies were also obtained by changing the specified wild-type protein. The construction descriptions of all datasets are shown in Table 1.

[0147] Table 1

[0148]

[0149] On the original dataset A0, the randomly discarded datasets A1-A4, and the simulated scan dataset A5, the roughness evaluation results of the extreme value method, the roughness evaluation results of the fitting method, and the roughness evaluation results of the proposed method are calculated respectively. The degree of interference of different roughness evaluation results under data missing conditions is compared.

[0150] in, Figure 4a The results of roughness evaluation using the extreme value method on datasets A0-A5 are shown below. Figure 4b The figure shows the roughness evaluation results of the fitting method on the dataset A0-A5. Figure 4c The table shows the roughness evaluation results of this method on datasets A0-A5. In the bar chart, each bar represents a box plot of 5 copies of the same dataset. The Y-axis represents the dimensionless numerical results of the method on different datasets, used to compare relative sizes; it is unitless, or in other words, the unit is "1".

[0151] according to Figure 4a , Figure 4b and Figure 4c The roughness evaluation results are compared with the datasets with missing data (A1-A5). The roughness evaluation results obtained by the three roughness evaluation methods and the dataset corresponding to the original fitness space (A0) are compared. The difference in roughness evaluation between the two is the size of the data on the vertical axis.

[0152] The analysis of the roughness evaluation results above shows that the extreme value method uses the ratio of the number of local extreme value sequences to the total number of sequences to obtain the roughness evaluation result. Sequence A is a local extreme value sequence if and only if A has the highest or lowest fitness among all protein sequences (adjacent sequences) that mutate at one site with A. The larger the proportion of local extreme value sequences among all sequences, the coarser the dataset is considered. However, when protein sequences are missing from the dataset, the extreme value method only calculates whether there is an extreme value among the "adjacent sequences where protein sequence A exists". Therefore, as the deletion rate gradually increases, previously local extreme value sequences remain local extreme value sequences, and other protein sequences will also become new local extreme value sequences as true local extreme values ​​are lost. Therefore, when sequences are missing, the roughness evaluation result obtained by the extreme value method will abnormally increase as the deletion rate increases.

[0153] Regarding the fitting method, a linear model is used to fit the dataset, and the mean squared error R and mean squared slope S of the linear model are calculated. Then, R / S is calculated as the roughness evaluation result of the model. Obviously, because more rugged regions in the fitness space may be missing, the entire sequence space will appear "smoother," but this is just an abnormal smoothing effect caused by incomplete datasets. The real complete dataset will be much rougher. As the missing proportion increases, the roughness evaluation result of the fitting method will show an abnormal decrease.

[0154] In this method, only adjacent sequences are sampled because non-adjacent sequences may produce an abnormally smoothing effect on the edit distance when the protein sequence is deleted. However, adjacent sequences maintain an approximate roughness before and after the protein sequence deletion. Therefore, calculating only adjacent sequences can ensure that the roughness evaluation results are not affected by the protein sequence deletion as much as possible.

[0155] In summary, the roughness evaluation results of this method still have good accuracy and small error even when data is missing, and can more closely approximate the true roughness of the original fitness space.

[0156] In this application scenario, when protein sequences are missing in the protein fitness space, this method can reduce the systematic error in the roughness evaluation results obtained under the existing fitness space compared to the roughness evaluation results obtained under the complete fitness space. Compared to the extreme value method and the fitting method, this method can achieve smaller systematic errors and obtain more accurate roughness evaluation results with similar computational cost, thereby solving the problem of large errors in the roughness evaluation results of biological sequence fitness spaces in related technologies.

[0157] The following are embodiments of the apparatus described in this application, which can be used to execute the roughness evaluation method for the fitness space involved in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments of the roughness evaluation method for the fitness space involved in this application.

[0158] Please see Figure 5 This application provides a roughness evaluation device 900 for fitness space, including but not limited to: a dataset acquisition module 910, a sequence pair confirmation module 930, a sequence group generation module 950, and a roughness evaluation module.

[0159] The dataset acquisition module 910 is used to acquire a dataset, which includes multiple biological sequences.

[0160] The sequence pair confirmation module 930 is used to determine multiple adjacent sequence pairs in the dataset based on the mutation sites between biological sequences in the dataset; the mutation sites between the two biological sequences in each adjacent sequence pair meet the set number of mutation sites.

[0161] Roughness evaluation module 950 is used to estimate the roughness of the fitness space based on each adjacent sequence to obtain the roughness evaluation result.

[0162] It should be noted that the fitness space roughness evaluation device provided in the above embodiments is only illustrated by the division of the above functional modules when performing fitness space roughness evaluation. In actual applications, the above functions can be assigned to different functional modules as needed. That is, the internal structure of the fitness space roughness evaluation device will be divided into different functional modules to complete all or part of the functions described above.

[0163] Furthermore, the embodiments of the fitness space roughness evaluation device and the fitness space roughness evaluation method provided in the above embodiments belong to the same concept, and the specific way in which each module performs its operation has been described in detail in the method embodiments, and will not be repeated here.

[0164] Figure 6 A schematic diagram of the structure of a server is shown according to an exemplary embodiment. This server is suitable for... Figure 1 The server 130 in the implementation environment is shown.

[0165] It should be noted that this server is merely an example adapted to this application and should not be construed as providing any limitation on the scope of use of this application. Nor should this server be interpreted as requiring or depending on any specific feature. Figure 6 One or more components of the exemplary server 2000 shown.

[0166] The hardware architecture of Server 2000 can vary significantly due to differences in configuration or performance, such as... Figure 6 As shown, the server 2000 includes: a power supply 210, an interface 230, at least one memory 250, and at least one central processing unit (CPU) 270.

[0167] Specifically, power supply 210 is used to provide operating voltage for the various hardware devices on server 2000.

[0168] Interface 230 includes at least one wired or wireless network interface for interacting with external devices. For example, to perform... Figure 1 The diagram shows the interaction between the acquisition terminal 110 and the server terminal 130 in the implementation environment.

[0169] Of course, in other examples adapted in this application, interface 230 may further include at least one serial-to-parallel conversion interface 233, at least one input / output interface 235, and at least one USB interface 237, etc. Figure 6 As shown, this does not constitute a specific limitation.

[0170] The memory 250 serves as a carrier for resource storage and can be a read-only memory, random access memory, disk, or optical disk, etc. The resources stored on it include the operating system 251, application programs 253, and data 255, etc., and the storage method can be temporary storage or permanent storage.

[0171] The operating system 251 is used to manage and control the various hardware devices and application programs 253 on the server 200, so as to enable the central processing unit 270 to perform calculations and processing on the massive data 255 in the memory 250. It can be Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0172] Application 253 is a program instruction or code based on operating system 251 that performs at least one specific task, and may include at least one module ( Figure 6 (Not shown), each module can contain program instructions or code for the server 2000. For example, the roughness evaluation device for the fitness space can be considered as application 253 deployed on the server 2000.

[0173] Data 255 can be photos, images, etc. stored on a disk, or recommendation information, etc., stored in memory 250.

[0174] The central processing unit 270 may include one or more processors and is configured to communicate with the memory 250 via at least one communication bus to read program instructions or code stored in the memory 250, thereby performing operations and processing on the massive amount of data 255 in the memory 250. For example, a roughness evaluation method for the fitness space can be completed by the central processing unit 270 reading a series of program instructions or code stored in the memory 250.

[0175] Furthermore, this application can also be implemented through hardware circuits or a combination of hardware circuits and software. Therefore, the implementation of this application is not limited to any specific hardware circuit, software, or combination thereof.

[0176] Please see Figure 7 This application provides an electronic device 4000, which may include: a desktop computer, a laptop computer, a server, etc.

[0177] exist Figure 7 In this context, the electronic device 4000 includes at least one processor 4001 and at least one memory 4003.

[0178] The data interaction between the processor 4001 and the memory 4003 can be achieved through at least one communication bus 4002. This communication bus 4002 may include a path for transmitting data between the processor 4001 and the memory 4003. The communication bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0179] Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of this application.

[0180] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0181] The memory 4003 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program instructions or code in the form of instructions or data structures and accessible by the electronic device 400, but not limited thereto.

[0182] The memory 4003 stores program instructions or code, and the processor 4001 can read the program instructions or code stored in the memory 4003 through the communication bus 4002.

[0183] When the program instructions or code are executed by the processor 4001, the roughness evaluation method of the fitness space in the above embodiments is implemented.

[0184] Furthermore, this application provides a storage medium storing program instructions or code, which are loaded and executed by a processor to implement the roughness evaluation method of the fitness space as described above.

[0185] This application provides an application product, which includes program instructions or code stored in a storage medium. The processor of an electronic device reads the program instructions or code from the storage medium, loads and executes the program instructions or code, so that the electronic device implements the roughness evaluation method of the fitness space as described above.

[0186] Compared with related technologies, firstly, for datasets of different sizes, using different methods to calculate adjacent sequence pairs can reduce computational load and speed up computation, thus ensuring the efficiency of roughness evaluation; secondly, this method has good scalability and can be applied to roughness evaluation in various biological sequence fitness spaces, such as DNA sequences, RNA sequences, and protein sequences; thirdly, the roughness evaluation results obtained through adjacent sequences can minimize the impact of biological sequence deletions on the roughness evaluation results, ensuring the accuracy of the roughness evaluation results, thus effectively solving the problem of large errors in the roughness evaluation results of biological sequence fitness spaces in related technologies.

[0187] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0188] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for evaluating the roughness of a fitness space, characterized in that, include: Obtain the dataset; The dataset includes multiple biological sequences; Based on the mutation sites between the biological sequences in the dataset, multiple adjacent sequence pairs in the dataset are determined; The mutation sites between the two biological sequences in each adjacent sequence pair conform to a set number of mutation sites, wherein the number of mutation sites is 1; Obtain the first fitness and the second fitness of each of the adjacent sequence pairs; The first fitness and the second fitness of each adjacent sequence pair are calculated respectively to obtain the fitness difference corresponding to each adjacent sequence pair; Based on the labels corresponding to each of the adjacent sequence pairs, at least one adjacent sequence group is generated; The fitness difference of each adjacent sequence pair in each of the adjacent sequence groups is standardized respectively; The roughness evaluation result is calculated based on the difference in fitness after standardization.

2. The method as described in claim 1, characterized in that, Before determining multiple adjacent sequence pairs in the dataset based on mutation sites among the biological sequences in the dataset, the method further includes: Determine whether the number of biological sequences in the dataset exceeds a set threshold; If the set threshold is not exceeded, adjacent sequence pairs are determined based on the edit distance of each biological sequence in the dataset.

3. The method as described in claim 2, characterized in that, The step of determining multiple adjacent sequence pairs in the dataset based on mutation sites among the biological sequences in the dataset includes: Based on the current biological sequence in the dataset, the edit distance between the current biological sequence and the other biological sequences in the dataset is calculated to obtain the edit distance between each biological sequence. For a given biological sequence, select from the remaining biological sequences that have an edit distance that meets a set distance threshold; From the current biological sequence and the filtered biological sequences, the adjacent sequence pairs corresponding to the current biological sequence are obtained; The process continues until all biological sequences have been screened, yielding the respective adjacent sequence pairs.

4. The method as described in claim 2, characterized in that, The method further includes: If the set number threshold is exceeded, adjacent sequence pairs are determined based on the candidate sequences corresponding to each biological sequence in the dataset; the candidate sequence refers to a biological sequence generated using the biological sequence as the parent biological sequence and whose edit distance from the parent biological sequence meets the set distance threshold.

5. The method as described in claim 4, characterized in that, The step of determining adjacent sequence pairs based on candidate sequences corresponding to each biological sequence in the dataset includes: For a given biological sequence corresponding to a candidate sequence, search the dataset to see if there exists a biological sequence that is identical to the corresponding candidate sequence. If the search is successful, then based on the current biological sequence and the found biological sequence, the adjacent sequence pair corresponding to the current biological sequence is obtained; This continues until all biological sequences have been searched, yielding each of the aforementioned adjacent sequence pairs.

6. A roughness evaluation device for fitness space, characterized in that, include: A dataset acquisition module is used to acquire a dataset; the dataset includes multiple biological sequences and the fitness of each biological sequence. The sequence pair confirmation module is used to determine multiple adjacent sequence pairs in the dataset based on mutation sites between the biological sequences in the dataset; The mutation sites between the two biological sequences in each of the adjacent sequence pairs conform to the set number of mutation sites; A roughness evaluation module is used to obtain the first fitness and the second fitness of each of the adjacent sequence pairs; The first fitness and the second fitness of each adjacent sequence pair are calculated to obtain the fitness difference corresponding to each adjacent sequence pair; at least one adjacent sequence group is generated based on the label corresponding to each adjacent sequence pair; the fitness difference of each adjacent sequence pair in each adjacent sequence group is standardized; and the roughness evaluation result is calculated based on the standardized fitness difference.

7. An electronic device, characterized in that, include: At least one processor and at least one memory, wherein, The memory stores program instructions or code; The program instructions or code are loaded and executed by the processor, causing the electronic device to implement the roughness evaluation method of the fitness space as described in any one of claims 1 to 5.

8. A storage medium storing program instructions or code thereon, characterized in that, The program instructions or code are loaded and executed by the processor to implement the roughness evaluation method of the fitness space as described in any one of claims 1 to 5.

Images