Method and device for constructing developmental tree of gene sequence system based on dual crawling parallelism

By combining parallel processing technology with SPR and NNI algorithms, the problems of slow convergence speed and long calculation time in the construction of viral gene sequence phylogenetic tree are solved, and fast and accurate phylogenetic tree construction and prediction are achieved.

CN112242178BActive Publication Date: 2024-11-29GUANGDONG INST OF INTELLIGENT MFG
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202010928396.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-09-07
Publication Date
2024-11-29
Estimated Expiration
2040-09-07

AI Technical Summary

Technical Problem

The prior art when constructing a phylogenetic tree of viral gene sequences, due to data characteristics, the convergence speed is slow, it is easy to fall into local optimal points, and the calculation time is long, making it difficult to complete the precise phylogenetic tree establishment within a limited time.

Method used

The combination of the topological structure algorithm based on SPR and NNI randomly changing phylogenetic tree is adopted to improve the computing efficiency by preprocessing gene sequence data and constructing phylogenetic tree in a parallel computing environment.

Benefits of technology

It has achieved rapid construction of gene sequence phylogenetic trees, accurately predicted the population number and reproductive ability of species in various historical periods, and predicted future development trends, significantly improving the computing speed and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112242178B_ABST
    Figure CN112242178B_ABST
Patent Text Reader

Abstract

The present invention discloses a developmental tree construction method and device based on a gene sequence system with dual crawling parallelism, wherein the method comprises: determining a calculation target, and selecting a corresponding gene sequence based on the calculation target; collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data; preprocessing the gene sequence data to obtain preprocessed gene sequence data; and constructing a developmental tree for the gene sequence data based on a combined parallel mechanism of an SPR and NNI random change algorithm for the topological structure of a phylogenetic tree to obtain a construction result. In an embodiment of the present invention, the population size of the species corresponding to the gene sequence in each historical period can be accurately predicted, the reproductive capacity of the species can also be obtained, and its future development trend can be predicted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of gene sequence data analysis, and in particular to a method and device for constructing a developmental tree of a gene sequence system based on dual crawling parallelism. Background Art

[0002] With the development of modern gene sequencing technology, mathematical models are established to analyze gene sequences, and based on this, the relationship between different species, as well as the number of species and reproduction rate of the same species in different geological periods, are found. For viral gene sequences, the rate of viral reproduction and the relationship between different strains can be obtained by building models and analyzing them, and then its ability to spread within the population can be estimated and the transmission path of the virus in the population can be obtained. All mathematical models describing reproduction and correlation are constructed based on phylogenetic trees of gene sequences. Therefore, the ability to accurately and quickly construct sequence phylogenetic trees can help people better understand biological species, study their evolution and growth laws, and serve industries related to biology.

[0003] Virus gene sequence data has the characteristics of a large number of data, short data length and great similarity between data. It is precisely because of these characteristics that when calculating the information for building the phylogenetic tree of virus sequences, the existing optimization algorithms have the problem of slow convergence and convergence to the local optimal point. In addition, due to the excessive number of data, when selecting an algorithm with higher accuracy (i.e. closer to the global optimal value), the calculation time of each step in the optimization calculation becomes very long, which makes it difficult to complete the construction of a phylogenetic tree within a limited time. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention provides a developmental tree construction method and device based on a double-crawling parallel gene sequence system, which can provide the progress of the analysis when analyzing the gene sequence data, and quickly complete the analysis. It can also accurately predict the population size of the species corresponding to the gene sequence in various historical periods, and can also obtain the reproductive capacity of the species and predict its future development trend.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for constructing a gene sequence phylogenetic tree based on dual crawling parallelism, the method comprising:

[0006] Determine a computational target, and select a corresponding gene sequence based on the computational target;

[0007] Collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data;

[0008] Preprocessing the gene sequence data to obtain preprocessed gene sequence data;

[0009] The gene sequence data is processed for developmental tree construction based on the combined parallel mechanism of SPR and NNI random change of the topological structure algorithm of the phylogenetic tree to obtain a construction result.

[0010] Optionally, the collecting of gene sequences based on the selected corresponding gene sequences to obtain gene sequence data includes:

[0011] Based on the selected corresponding gene sequence, gene sequence data is obtained by performing gene sequence collection under natural conditions through experimental means; or,

[0012] Based on the selected corresponding gene sequence, the corresponding gene database is searched and downloaded to obtain gene sequence data.

[0013] Optionally, the gene sequence data meets the diversity of gene sequence collection time, collection location and data source; the corresponding gene database includes GENE BANK and GISAID database.

[0014] Optionally, preprocessing the gene sequence data to obtain preprocessed gene sequence data includes:

[0015] Screening the gene sequence data to obtain screened gene sequence data;

[0016] The screened gene sequence data is aligned to obtain pre-processed gene sequence data.

[0017] Optionally, the screening process of the gene sequence data includes:

[0018] Determining the sequence of the composite standard based on the database using a database search language;

[0019] Screening and processing the gene sequence data based on the sequence of the composite standard;

[0020] The sequence of the composite standard is that the shortest gene sequence is greater than 80% of the longest gene sequence and conforms to the selected corresponding gene sequence.

[0021] Optionally, the aligning and arranging the screened gene sequence data includes:

[0022] Performing alignment and sorting processing on the screened gene sequence data based on industrial software mafft or kalign;

[0023] Among them, after alignment and sorting processing, it is unique for the given gene sequence data.

[0024] Optionally, the parallel mechanism of combining the SPR and NNI random change phylogenetic tree topology algorithm to construct a phylogenetic tree on the gene sequence data includes:

[0025] Determining target parameters of the required calculation based on the calculation target;

[0026] Based on the target parameters, the gene sequence data is processed by constructing a phylogenetic tree using a combination of SPR and NNI algorithms for randomly changing the topological structure of the phylogenetic tree in parallel.

[0027] Optionally, the method of constructing a phylogenetic tree based on the target parameter by using a combination of SPR and NNI algorithms for randomly changing the topological structure of the phylogenetic tree in parallel to construct a phylogenetic tree on the gene sequence data includes:

[0028] On the root thread, from the empirically given distribution function g(x|θ k-1 ) extract a parameter set;

[0029] Broadcast the parameter set from the root thread to n other different threads;

[0030] On the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i=1,2,…,N;

[0031] Randomly select N trees T at a time on the root thread i S A SPR neighbor tree T i * , i=1,2,…,N;

[0032] The SPR neighbor tree T i * Evenly transfer from the root thread to other n different threads;

[0033] In each non-root thread, for each SPR neighbor tree T i * Sliding randomly changes the time of each node of the tree;

[0034] Calculate each SPR neighbor tree T on each non-root thread i * L(T i * ,θ * |D);

[0035] Select such that L(T i* ,θ * |D) The largest tree * max ;

[0036] On the root thread, draw a random number r from the uniform distribution U(0,1);

[0037] Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 ;

[0038] Among them, θ represents the target parameter set; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters; T represents the phylogenetic tree required.

[0039] Optionally, the L(T k-1 ,θ k-1 |D) is the optimization objective function, which is obtained by the Bayesian formula; the distribution function g(x|θ k-1 ) is the normal distribution function or the exponential distribution function.

[0040] In addition, an embodiment of the present invention further provides a gene sequence phylogenetic tree construction device based on dual crawling parallelism, the device comprising:

[0041] Determination module: used to determine the calculation target and select the corresponding gene sequence based on the calculation target;

[0042] Collection module: used for collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data;

[0043] Preprocessing module: used to preprocess the gene sequence data to obtain preprocessed gene sequence data;

[0044] Construction module: used for performing developmental tree construction processing on the gene sequence data based on the combined parallel mechanism of SPR and NNI random change of the topological structure algorithm of the phylogenetic tree to obtain the construction result.

[0045] In an embodiment of the present invention, the collected gene sequence data is preprocessed, and then the developmental tree is constructed by combining the parallel mechanism of SPR and NNI random change of the topological structure algorithm of the phylogenetic tree. When analyzing the gene sequence data, the progress of the analysis can be provided, and the analysis can be completed quickly. The population size of the species corresponding to the gene sequence in each historical period can be accurately predicted, the reproductive capacity of the species can be obtained, and its future development trend can be predicted. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a flow chart of a method for constructing a developmental tree based on a double-crawling parallel gene sequence system in an embodiment of the present invention;

[0048] Figure 2 It is a schematic diagram of the structural composition of a developmental tree construction device based on a dual-crawling parallel gene sequence system in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Example

[0051] See also Figure 1 , Figure 1 It is a flow chart of a method for constructing a developmental tree based on a double-crawling parallel gene sequence system in an embodiment of the present invention.

[0052] like Figure 1 As shown, a method for constructing a gene sequence phylogenetic tree based on dual crawling parallelism, the method comprising:

[0053] S11: determining a calculation target, and selecting a corresponding gene sequence based on the calculation target;

[0054] In the specific implementation process of the present invention, it is first necessary to determine the calculation target, and then select the corresponding gene sequence according to the calculation target; for example, in virology, the virus transmission rate parameter, i.e., the basic reproduction number R, is estimated based on the sequence. 0 For target research, it is necessary to collect the corresponding viral gene sequence. Common ones, such as influenza virus, need to collect HA or NA gene sequence, and the new coronavirus commonly needs to collect S gene sequence.

[0055] S12: collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data;

[0056] In the specific implementation process of the present invention, the collection of gene sequences based on the selection of corresponding gene sequences to obtain gene sequence data includes: collecting gene sequences under natural conditions through experimental means based on the selection of corresponding gene sequences to obtain gene sequence data; or, searching and downloading in a corresponding gene database based on the selection of corresponding gene sequences to obtain gene sequence data.

[0057] Furthermore, the gene sequence data meets the diversity of gene sequence collection time, collection location and data source; the corresponding gene databases include GENE BANK and GISAID databases.

[0058] In the specific implementation process of the present invention, when collecting gene sequences, it is necessary to collect them under natural conditions through experimental means according to the selection of corresponding gene sequences, or to search and download corresponding gene sequence data through databases such as GENE BANK and GISAID; the collection of gene sequence data needs to meet the diversity of gene sequence collection time, collection location and data source to ensure that the sampling is statistically significant.

[0059] S13: preprocessing the gene sequence data to obtain preprocessed gene sequence data;

[0060] In the specific implementation process of the present invention, the preprocessing of the gene sequence data to obtain the preprocessed gene sequence data includes: screening the gene sequence data to obtain the screened gene sequence data; aligning the screened gene sequence data to obtain the preprocessed gene sequence data.

[0061] Furthermore, the screening and processing of the gene sequence data includes: determining a composite standard sequence based on a database using a database search language; screening and processing the gene sequence data based on the composite standard sequence; the composite standard sequence is that the shortest gene sequence is greater than 80% of the longest gene sequence and conforms to the selected corresponding gene sequence.

[0062] Furthermore, the alignment and arrangement processing of the screened gene sequence data includes: performing alignment and sorting processing on the screened gene sequence data based on industrial software mafft or kalign; wherein the alignment and sorting processing is unique for the given gene sequence data.

[0063] Specifically, the collected gene sequence data is preprocessed to ensure the effectiveness and rationality of the next gene calculation; the specific preprocessing method includes: screening the gene sequence data, using a database search language based on the database to determine a specific composite standard, wherein the composite standard sequence is that the shortest gene sequence is greater than 80% of the longest gene sequence and meets the selection of the corresponding gene sequence; the difference between the lengths of the screened gene sequence data is within a controllable range, that is, the shortest sequence is greater than 80% of the longest sequence; then the screened gene sequence data is aligned, and the alignment is generally performed using mature industrial software mafft or kalign; the alignment result is theoretically unique for the given gene sequence data; the corresponding fragments are correctly corresponding, and all gene sequences are aligned using position symbols.

[0064] S14: Based on the parallel mechanism of combining SPR and NNI algorithms for randomly changing the topological structure of the phylogenetic tree, a phylogenetic tree construction process is performed on the gene sequence data to obtain a construction result.

[0065] In the specific implementation process of the present invention, the combined parallel mechanism of the SPR and NNI random change system development tree topology algorithm performs developmental tree construction processing on the gene sequence data, including: determining the required calculation target parameters based on the calculation target; based on the target parameters, using the combined parallel mechanism of the SPR and NNI random change system development tree topology algorithm to construct the gene sequence data.

[0066] Furthermore, the method of constructing a phylogenetic tree based on the target parameter by using a combination of SPR and NNI to randomly change the topological structure algorithm of the phylogenetic tree comprises: on the root thread, from the distribution function g(x|θ k-1 ) extract a parameter set; broadcast the parameter set from the root thread to the other n different threads; on the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i = 1, 2, ..., N; randomly select N trees T on the root thread at a time i S A SPR neighbor tree T i* , i=1,2,…,N; SPR neighbor tree T i * Evenly transfer from the root thread to the other n different threads; in each non-root thread, for each SPR neighbor tree T i * The time for sliding and randomly changing each node of the tree; calculate each SPR neighbor tree T on each non-root thread i * L(T i * ,θ * |D); select such that L(T i * ,θ * |D) The largest tree * max ; Draw a random number r from the uniform distribution U(0,1) on the root thread; Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 ; where θ represents the target parameter set; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters; T represents the phylogenetic tree required.

[0067] Furthermore, the L(T k-1 ,θ k-1 |D) is the optimization objective function, which is obtained by the Bayesian formula; the distribution function g(x|θ k-1 ) is the normal distribution function or the exponential distribution function.

[0068] Specifically, according to the determined calculation target, the target parameter θ of the calculation is determined, and then the number and range of parameters that can be obtained are determined according to the biological scope framed by the calculation target and the corresponding literature information, that is, the target parameter; generally speaking, the target parameters include the mutation rate of the gene sequence and the number of effective populations; specifically, the target is the gene sequence data of the virus to determine the virus transmission speed, and the parameters also include the basic reproduction number R 0 ; For those whose goal is to calculate the population size based on the population genes, the parameter is the effective population size in each historical period (no longer a constant but a piecewise function).

[0069] Therefore, assuming that the maximum number of iterations is M, the following calculation steps are run for each step k = 1, 2, ..., M:

[0070] 1. On the root thread, from the distribution g(x|θ k-1 ) extract θ*;

[0071] 2. Set θ * Broadcast from the root thread to the other n threads;

[0072] 3. On the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i=1,2,…,N;

[0073] 4. Randomly select T on the root thread i S , i=1,2,…,N, a SPR neighbor T i * , i=1,2,…,N;

[0074] 5. T i * , i = 1, 2, ..., N are evenly transferred from the root thread to other n different threads;

[0075] 6. On each non-root thread, for each tree T i * , i=1,2,…,N sliding randomly changes the time of each node of the tree;

[0076] 7. Calculate each tree T on each non-root thread i * ,i=1,2,…,N L(T i * ,θ * |D);

[0077] 8. Select the value that makes L(T i * ,θ * |D) The largest tree * max ;

[0078] 9. Draw a random number r from the uniform distribution U(0,1) on the root thread;

[0079] 10. Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise Tk ←T k-1 ,θ k ←θ k-1 .

[0080] Among them, θ is the target parameter to be obtained; T is the phylogenetic tree to be obtained, and a set of gene sequences corresponds to an optimal phylogenetic tree; g(x|θ k-1 ) is a distribution function given by experience, usually normal distribution or exponential distribution; L(T k-1 ,θ k-1 |D) is the optimization objective function, which is generally given by the Bayesian formula. Specific problems have specific forms; the remaining subscripts and subscripts are the transitional forms of specific parameters in the calculation process; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters;.

[0081] In the above, L(T k-1 ,θ k-1 |D) To optimize the objective function, the objective function expression is abstracted using mathematical methods. Threads are threads used in parallel computing, which can be modified according to different parallel environments; SPR and NNI are the two most common algorithms for randomly changing the topological structure of phylogenetic trees; this algorithm is derived from the classic MCMC phylogenetic tree algorithm. On the basis of the traditional use of SPR or NNI in each step, the two methods are mixed (imitating the climbing mode of taking two steps at a time when climbing a mountain) and the parallel mechanism is added, which can improve the calculation speed and accuracy at the same time.

[0082] After obtaining the target parameter θ and the phylogenetic tree T as described above, the required sequence quantification information can be obtained, and based on this, the corresponding information of the species in each historical period can be obtained and its future changes can be predicted; for example, when the parameter θ includes the number of species, the population size of the species in each historical period (generally mammals can be calculated for 3-5 million years) can be obtained. When the parameter θ includes the basic reproduction number, the reproductive capacity of the virus can be obtained and its future development trend can be predicted.

[0083] In an embodiment of the present invention, the collected gene sequence data is preprocessed, and then the developmental tree is constructed by combining the parallel mechanism of SPR and NNI random change of the topological structure algorithm of the phylogenetic tree. When analyzing the gene sequence data, the progress of the analysis can be provided, and the analysis can be completed quickly. The population size of the species corresponding to the gene sequence in each historical period can be accurately predicted, the reproductive capacity of the species can be obtained, and its future development trend can be predicted.

[0084] Example

[0085] See also Figure 2 , Figure 2 It is a schematic diagram of the structural composition of a developmental tree construction device based on a dual-crawling parallel gene sequence system in an embodiment of the present invention.

[0086] like Figure 2 As shown, a gene sequence phylogenetic tree construction device based on dual crawling parallelism, the device comprising:

[0087] Determination module 21: used to determine a calculation target and select a corresponding gene sequence based on the calculation target;

[0088] In the specific implementation process of the present invention, it is first necessary to determine the calculation target, and then select the corresponding gene sequence according to the calculation target; for example, in virology, the virus transmission rate parameter, i.e., the basic reproduction number R, is estimated based on the sequence. 0 For target research, it is necessary to collect the corresponding viral gene sequence. Common ones, such as influenza virus, need to collect HA or NA gene sequence, and the new coronavirus commonly needs to collect S gene sequence.

[0089] Collection module 22: used to collect gene sequences based on the selected corresponding gene sequences to obtain gene sequence data;

[0090] In the specific implementation process of the present invention, the collection of gene sequences based on the selection of corresponding gene sequences to obtain gene sequence data includes: collecting gene sequences under natural conditions through experimental means based on the selection of corresponding gene sequences to obtain gene sequence data; or, searching and downloading in a corresponding gene database based on the selection of corresponding gene sequences to obtain gene sequence data.

[0091] Furthermore, the gene sequence data meets the diversity of gene sequence collection time, collection location and data source; the corresponding gene databases include GENE BANK and GISAID databases.

[0092] In the specific implementation process of the present invention, when collecting gene sequences, it is necessary to collect them under natural conditions through experimental means according to the selection of corresponding gene sequences, or to search and download corresponding gene sequence data through databases such as GENE BANK and GISAID; the collection of gene sequence data needs to meet the diversity of gene sequence collection time, collection location and data source to ensure that the sampling is statistically significant.

[0093] Preprocessing module 23: used to preprocess the gene sequence data to obtain preprocessed gene sequence data;

[0094] In the specific implementation process of the present invention, the preprocessing of the gene sequence data to obtain the preprocessed gene sequence data includes: screening the gene sequence data to obtain the screened gene sequence data; aligning the screened gene sequence data to obtain the preprocessed gene sequence data.

[0095] Furthermore, the screening and processing of the gene sequence data includes: determining a composite standard sequence based on a database using a database search language; screening and processing the gene sequence data based on the composite standard sequence; the composite standard sequence is that the shortest gene sequence is greater than 80% of the longest gene sequence and conforms to the selected corresponding gene sequence.

[0096] Furthermore, the alignment and arrangement processing of the screened gene sequence data includes: performing alignment and sorting processing on the screened gene sequence data based on industrial software mafft or kalign; wherein the alignment and sorting processing is unique for the given gene sequence data.

[0097] Specifically, the collected gene sequence data is preprocessed to ensure the effectiveness and rationality of the next gene calculation; the specific preprocessing method includes: screening the gene sequence data, using a database search language based on the database to determine a specific composite standard, wherein the composite standard sequence is that the shortest gene sequence is greater than 80% of the longest gene sequence and meets the selection of the corresponding gene sequence; the difference between the lengths of the screened gene sequence data is within a controllable range, that is, the shortest sequence is greater than 80% of the longest sequence; then the screened gene sequence data is aligned, and the alignment is generally performed using mature industrial software mafft or kalign; the alignment result is theoretically unique for the given gene sequence data; the corresponding fragments are correctly corresponding, and all gene sequences are aligned using position symbols.

[0098] Construction module 24: for constructing a phylogenetic tree based on the parallel mechanism of the SPR and NNI random change phylogenetic tree topology algorithm to obtain a construction result.

[0099] In the specific implementation process of the present invention, the combined parallel mechanism of the SPR and NNI random change system development tree topology algorithm performs developmental tree construction processing on the gene sequence data, including: determining the required calculation target parameters based on the calculation target; based on the target parameters, using the combined parallel mechanism of the SPR and NNI random change system development tree topology algorithm to construct the gene sequence data.

[0100] Furthermore, the method of constructing a phylogenetic tree based on the target parameter by using a combination of SPR and NNI to randomly change the topological structure algorithm of the phylogenetic tree comprises: on the root thread, from the distribution function g(x|θ k-1 ) extract a parameter set; broadcast the parameter set from the root thread to the other n different threads; on the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i = 1, 2, ..., N; randomly select N trees T on the root thread at a time i S A SPR neighbor tree T i * , i=1,2,…,N; SPR neighbor tree T i * Evenly transfer from the root thread to the other n different threads; in each non-root thread, for each SPR neighbor tree T i * The time for sliding and randomly changing each node of the tree; calculate each SPR neighbor tree T on each non-root thread i * L(T i * ,θ * |D); select such that L(T i * ,θ * |D) The largest tree * max ; Draw a random number r from the uniform distribution U(0,1) on the root thread; Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 ; where θ represents the target parameter set; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters; T represents the phylogenetic tree required.

[0101] Furthermore, the L(T k-1 ,θ k-1 |D) is the optimization objective function, which is obtained by the Bayesian formula; the distribution function g(x|θ k-1) is the normal distribution function or the exponential distribution function.

[0102] Specifically, according to the determined calculation target, the target parameter θ of the calculation is determined, and then the number and range of parameters that can be obtained are determined according to the biological scope framed by the calculation target and the corresponding literature information, that is, the target parameter; generally speaking, the target parameters include the mutation rate of the gene sequence and the number of effective populations; specifically, the target is the gene sequence data of the virus to determine the virus transmission speed, and the parameters also include the basic reproduction number R 0 ; For those whose goal is to calculate the population size based on the population genes, the parameter is the effective population size in each historical period (no longer a constant but a piecewise function).

[0103] Therefore, assuming that the maximum number of iterations is M, the following calculation steps are run for each step k = 1, 2, ..., M:

[0104] 1. On the root thread, from the distribution g(x|θ k-1 ) extract θ*;

[0105] 2. Set θ * Broadcast from the root thread to the other n threads;

[0106] 3. On the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i=1,2,…,N;

[0107] 4. Randomly select T on the root thread i S , i=1,2,…,N, a SPR neighbor T i * , i=1,2,…,N;

[0108] 5. T i * , i = 1, 2, ..., N are evenly transferred from the root thread to other n different threads;

[0109] 6. On each non-root thread, for each tree T i * , i=1,2,…,N sliding randomly changes the time of each node of the tree;

[0110] 7. Calculate each tree T on each non-root thread i * ,i=1,2,…,N L(T i * ,θ * |D);

[0111] 8. Select the value that makes L(T i * ,θ * |D) The largest tree * max ;

[0112] 9. Draw a random number r from the uniform distribution U(0,1) on the root thread;

[0113] 10. Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 .

[0114] Among them, θ is the target parameter to be obtained; T is the phylogenetic tree to be obtained, and a set of gene sequences corresponds to an optimal phylogenetic tree; g(x|θ k-1 ) is a distribution function given by experience, usually normal distribution or exponential distribution; L(T k-1 ,θ k-1 |D) is the optimization objective function, which is generally given by the Bayesian formula. Specific problems have specific forms; the remaining subscripts and subscripts are the transitional forms of specific parameters in the calculation process; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters;.

[0115] In the above, L(T k-1 ,θ k-1 |D) To optimize the objective function, the objective function expression is abstracted using mathematical methods. Threads are threads used in parallel computing, which can be modified according to different parallel environments; SPR and NNI are the two most common algorithms for randomly changing the topological structure of phylogenetic trees; this algorithm is derived from the classic MCMC phylogenetic tree algorithm. On the basis of the traditional use of SPR or NNI in each step, the two methods are mixed (imitating the climbing mode of taking two steps at a time when climbing a mountain) and the parallel mechanism is added, which can improve the calculation speed and accuracy at the same time.

[0116] After obtaining the target parameter θ and the phylogenetic tree T as described above, the required sequence quantification information can be obtained, and based on this, the corresponding information of the species in each historical period can be obtained and its future changes can be predicted; for example, when the parameter θ includes the number of species, the population size of the species in each historical period (generally mammals can be calculated for 3-5 million years) can be obtained. When the parameter θ includes the basic reproduction number, the reproductive capacity of the virus can be obtained and its future development trend can be predicted.

[0117] In an embodiment of the present invention, the collected gene sequence data is preprocessed, and then the developmental tree is constructed by combining the parallel mechanism of SPR and NNI random change of the topological structure algorithm of the phylogenetic tree. When analyzing the gene sequence data, the progress of the analysis can be provided, and the analysis can be completed quickly. The population size of the species corresponding to the gene sequence in each historical period can be accurately predicted, the reproductive capacity of the species can be obtained, and its future development trend can be predicted.

[0118] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0119] In addition, the above is a detailed introduction to a developmental tree construction method and device based on a dual-crawling parallel gene sequence system provided in an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for a person skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for constructing a phylogenetic tree of gene sequences based on dual crawling parallelism, characterized in that: The method comprises: Determine a computational target, and select a corresponding gene sequence based on the computational target; Collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data; Preprocessing the gene sequence data to obtain preprocessed gene sequence data; Based on the SPR and NNI random change phylogenetic tree topology algorithm combined with a parallel mechanism, the gene sequence data is subjected to a phylogenetic tree construction process to obtain a construction result; The algorithm for randomly changing the topology structure of the phylogenetic tree based on SPR and NNI is combined with a parallel mechanism to construct a phylogenetic tree for the gene sequence data, including: Determining target parameters of the required calculation based on the calculation target; Based on the target parameters, a parallel mechanism combining SPR and NNI algorithms for randomly changing the topological structure of the phylogenetic tree is used to construct a phylogenetic tree for the gene sequence data; The method of constructing a phylogenetic tree based on the target parameter by using a combination of SPR and NNI random change phylogenetic tree topology algorithm and a parallel mechanism comprises: On the root thread, from the empirically given distribution function g(x|θ k-1 ) extract a parameter set; Broadcast the parameter set from the root thread to n other different threads; On the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i=1,2,…,N; Randomly select N trees T at a time on the root thread i S A SPR neighbor tree T i * , i=1,2,…,N; The SPR neighbor tree T i * Evenly transfer from the root thread to other n different threads; In each non-root thread, for each SPR neighbor tree T i * Sliding randomly changes the time of each node of the tree; Calculate each SPR neighbor tree T on each non-root thread i * L(T i * ,θ * |D); Select such that L(T i * ,θ * |D) The largest tree * max ; On the root thread, draw a random number r from the uniform distribution U(0,1); Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 ; Among them, θ represents the target parameter set; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters; T represents the phylogenetic tree required.

2. The method for constructing a phylogenetic tree of a gene sequence according to claim 1, characterized in that: The collecting of gene sequences based on the selected corresponding gene sequences to obtain gene sequence data includes: Based on the selected corresponding gene sequence, gene sequence data is obtained by performing gene sequence collection under natural conditions through experimental means; or, Based on the selected corresponding gene sequence, the corresponding gene database is searched and downloaded to obtain gene sequence data.

3. The method for constructing a gene sequence phylogenetic tree according to claim 2, characterized in that: The gene sequence data meets the diversity of gene sequence collection time, collection location and data source; the corresponding gene databases include GENE BANK and GISAID databases.

4. The method for constructing a gene sequence phylogenetic tree according to claim 1, characterized in that: The preprocessing of the gene sequence data to obtain the preprocessed gene sequence data includes: Screening the gene sequence data to obtain screened gene sequence data; The screened gene sequence data is aligned to obtain pre-processed gene sequence data.

5. The method for constructing a gene sequence phylogenetic tree according to claim 4, characterized in that: The screening process of the gene sequence data includes: Determining the sequence of the composite standard based on the database using a database search language; Screening and processing the gene sequence data based on the sequence of the composite standard; The sequence of the composite standard is that the shortest gene sequence is greater than 80% of the longest gene sequence and conforms to the selected corresponding gene sequence.

6. The method for constructing a phylogenetic tree of a gene sequence according to claim 4, characterized in that: The aligning and arranging the screened gene sequence data comprises: Performing alignment and sorting processing on the screened gene sequence data based on industrial software mafft or kalign; Among them, after alignment and sorting processing, it is unique for the given gene sequence data.

7. The method for constructing a gene sequence phylogenetic tree according to claim 1, characterized in that: The L(T k-1 ,θ k-1 |D) is the optimization objective function, which is obtained by the Bayesian formula; the distribution function g(x|θ k-1 ) is the normal distribution function or the exponential distribution function.

8. A gene sequence phylogenetic tree construction device based on dual crawling parallelism, characterized in that: The device comprises: Determination module: used to determine the calculation target and select the corresponding gene sequence based on the calculation target; Collection module: used for collecting gene sequences based on the selected corresponding gene sequences to obtain gene sequence data; Preprocessing module: used to preprocess the gene sequence data to obtain preprocessed gene sequence data; Construction module: used for constructing a phylogenetic tree based on the SPR and NNI random change phylogenetic tree topology algorithm combined with a parallel mechanism to obtain a construction result; The algorithm for randomly changing the topology structure of the phylogenetic tree based on SPR and NNI is combined with a parallel mechanism to construct a phylogenetic tree for the gene sequence data, including: Determining target parameters of the required calculation based on the calculation target; Based on the target parameters, a parallel mechanism combining SPR and NNI algorithms for randomly changing the topological structure of the phylogenetic tree is used to construct a phylogenetic tree for the gene sequence data; The method of constructing a phylogenetic tree based on the target parameter by using a combination of SPR and NNI random change phylogenetic tree topology algorithm and a parallel mechanism comprises: On the root thread, from the empirically given distribution function g(x|θ k-1 ) extract a parameter set; Broadcast the parameter set from the root thread to n other different threads; On the root thread, from the tree T k-1 Randomly select N trees T without duplication from the NNI neighbors i S , i=1,2,…,N; Randomly select N trees T at a time on the root thread i S A SPR neighbor tree T i * , i=1,2,…,N; The SPR neighbor tree T i * Evenly transfer from the root thread to other n different threads; In each non-root thread, for each SPR neighbor tree T i * Sliding randomly changes the time of each node of the tree; Calculate each SPR neighbor tree T on each non-root thread i * L(T i * ,θ * |D); Select such that L(T i * ,θ * |D) The largest tree * max ; On the root thread, draw a random number r from the uniform distribution U(0,1); Determine whether r is satisfied on the root thread <min(1,L(T k-1 ,θ k-1 |D)) If it satisfies then T←T * max ,θ←θ * , otherwise T k ←T k-1 ,θ k ←θ k-1 ; Among them, θ represents the target parameter set; θ * represents the parameter set; θ k-1 and θ k Both represent a set of historical target parameters; T represents the phylogenetic tree required.

Citation Information

Patent Citations

  • Rebuilding method of phylogenetic tree

    CN103093118A

  • Parallel gene splicing algorithm based on cluster map structure

    CN103761453A