Function generalization and attractor determination method for gene regulatory network

Abstract

The invention provides a function generalization and attractor determination method for a gene regulatory network, which comprises the following steps: for each target gene in the gene regulatory network, calculating a Cramer correlation coefficient between the target gene and all regulatory genes corresponding to the target gene, and selecting the regulatory gene with the highest Cramer correlation coefficient value as guidance; setting other unobserved values in the truth table of the gene regulation and control network; and based on a gene regulatory network truth table, obtaining attractors by traversing each state of the gene. According to the method, the correlation between the genes is calculated by adopting the Cramer correlation coefficient, the problem of low accuracy of the existing algorithm is effectively improved, and the algorithm using the Cramer correlation coefficient and the algorithm proposed in the prior art are subjected to a large amount of experimental comparison; it is proved that the average sensitivity error and the steady-state distribution distance of the method are both reduced, and therefore the effect that truth table prediction is more accurate is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of biological sequencing data processing, in particular to a function generalization method for gene regulatory networks and a method for rapidly determining attractors. Background Art

[0002] In recent years, the rapid development of gene sequence group recognition technology has provided a large amount of experimental data for the study of gene networks. The gene regulatory network is essentially a continuous and complex dynamic system, that is, a complex dynamic system network. The genetic information and genes of any cell in an organism are the same, but the same gene behaves differently in different tissues and cells. The expression of a gene not only affects other genes, but is also affected by other genes. Genes promote and inhibit each other, and are activated in a large environment such as a comprehensive environment in a specific cell and at a specific time, forming a complex gene regulatory network.

[0003] The prediction of the truth table of the gene regulatory network is a very important step in the study of biological sequencing data. The research methods proposed so far, such as the majority rule, the generalization method based on mutual information and the skewness, are all based on biological sequencing data and gene regulatory network models to predict the truth table. Among them, the majority rule is to replace all unknown values ​​with the larger value according to the ratio of 0 and 1 in the data in the existing truth table; the specific implementation of the algorithm based on the skewness is to calculate the ratio of 0 and 1 in the observed data of the target gene, and randomly generalize the unobserved part according to this ratio; the generalization algorithm based on mutual information is to find a gene that has the greatest impact on the regulated gene, and supplement the remaining part according to its criteria. However, after reproducing these three methods, it was found that the performance in the two indicators of average sensitivity error and average steady-state distribution distance was not ideal, which shows that the current method has low prediction accuracy and poor effect.

[0004] In addition, determining the attractor of a gene regulatory network is also a very important step in the research. The attractor is the stable state of the gene during the evolution process. Only by determining the attractor of the genome can it be intervened, so as to finally achieve the purpose of artificially intervening in the steady state of the gene and making it evolve in the expected direction. At present, the method of determining attractors is to use machine learning algorithms to identify attractors in the network. However, machine learning methods may be affected by data noise, and the biological understanding of gene regulatory networks may not be as in-depth as the analysis based on dynamic systems. There are also methods based on network reconstruction to determine attractors. This type of method reconstructs the network and identifies attractors by comparing experimental data with data predicted by computational models, such as Bayesian network reconstruction. However, the network reconstruction method requires a large amount of experimental data to ensure the accuracy of the model, and when the network scale is large, the computational complexity is extremely high.

[0005] Currently, there is no method that can both predict the truth table of the gene regulatory network and quickly determine the attractor based on the prediction. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention proposes a method for function generalization and attractor determination for gene regulatory networks, which can achieve efficient prediction of the truth table of the gene regulatory network and rapid determination of attractors.

[0007] The technical solution of the present invention is:

[0008] A method for function generalization and attractor determination of a gene regulatory network comprises the following steps:

[0009] Step 1: For each target gene in the gene regulatory network, calculate the Cramer correlation coefficient between the target gene and all its corresponding regulatory genes, and select the regulatory gene with the highest Cramer correlation coefficient value as a guide to set other unobserved values ​​in the truth table of the gene regulatory network;

[0010] Step 2: Based on the truth table of the gene regulatory network obtained in step 1, the attractor is obtained by traversing each state of the gene.

[0011] Furthermore, in step 2, the specific process of traversing each state of the gene to obtain the attractor is:

[0012] When the number of nodes in the gene regulatory network is N, the corresponding total number of states is 2 N , traverse all states;

[0013] Assume that the attractor set is S and the current state is x t , according to the rules of the truth table, get the next state x t+1 , if x t =x t+1 , then judge x t Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x to t Put it into S and start the evolution of the next state;

[0014] If x t ≠x t+1 , then continue to obtain the next state x t+2 , and compare x t+1 and x t+2 , if x t+1 =x t+2 Then judge x t+1 Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x tot+1 Put it into S and start the evolution of the next state;

[0015] And so on, after traversing all states, the set of states in S is the entire attractor.

[0016] Beneficial Effects

[0017] The function generalization method for gene regulatory networks proposed in the present invention has the following advantages:

[0018] 1. The present invention uses the Cramer correlation coefficient to calculate the correlation between genes, which effectively improves the low accuracy of the existing algorithm. A large number of experimental comparisons are carried out between the algorithm using the Cramer correlation coefficient and the algorithm proposed in the prior art, which proves that the method of the present invention reduces both the average sensitivity error and the steady-state distribution distance, thereby achieving a more accurate effect of truth table prediction.

[0019] 2. The present invention fully combines the function generalization algorithm and the attractor search algorithm, effectively solves the problem of imperfect functions of the prior art, and realizes more perfect processing of biological sequencing data.

[0020] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0022] Figure 1 This is the overall flow chart for gene regulatory network research;

[0023] Figure 2 A working interface for obtaining a complete truth table and an attractor in Example 2 of the present application;

[0024] Figure 3 This is the result diagram in Example 2 of the present application;

[0025] Figure 4 This is an example of a partial truth table given in the content of the invention.

[0026] Figure 5 This is a comparison chart of the average sensitivity error indicators of the algorithms corresponding to the three correlation coefficients.

[0027] Figure 6 This is a comparison chart of the average steady-state distribution distance indicators of the algorithms corresponding to the three correlation coefficients.

[0028] Figure 7This is a comparison chart of the average sensitivity error index between the method of the present invention and the prior art.

[0029] Figure 8 This is a comparison chart of the average steady-state distribution distance index between the method of the present invention and the prior art. DETAILED DESCRIPTION

[0030] Embodiments of the present invention are described in detail below. The embodiments are exemplary and intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0031] In order to make a more accurate prediction of the truth table of the gene regulatory network, the applicant considered using the correlation between the regulatory genes and the regulated genes in the gene regulatory network as a basis to reflect the strength of the interaction between the two genes, and then determine the regulatory gene with the highest correlation with the regulated gene, and generalize the unknown state in the truth table based on the most likely state of the regulated gene when the state of the regulatory gene is 0 or 1 (the gene state is 0 indicating that the gene is not expressed, and the gene state is 1 indicating that the gene is expressed).

[0032] The applicant selected three correlation coefficients, namely, the Pearson correlation coefficient, the Spearman rank correlation coefficient, and the Cramer correlation coefficient, for research. For each target gene in the gene regulatory network, three correlation coefficients are calculated between the target gene and all corresponding regulatory genes. Based on the three correlation coefficients, the three regulatory genes with the highest correlation with the target gene are obtained respectively. The states of the target gene when the gene states corresponding to the three regulatory genes are 0 or 1 are counted respectively, and the unknown states in the truth table are generalized according to the corresponding states.

[0033] Then, the final generalized truth table of the gene regulatory network corresponding to the three correlation coefficients was evaluated, and the performance of the predicted gene regulatory network was evaluated by two indicators: steady-state distribution distance and sensitivity error. According to the data results, the prediction result of the Cramer correlation coefficient was the best. Finally, it was determined that the Cramer correlation coefficient was used to calculate the correlation between the regulatory gene and the regulated gene. Therefore, the Cramer correlation coefficient was used to calculate the correlation between genes in the present invention. The algorithm evaluation indicators corresponding to the three correlation coefficients are as follows: Figure 5 , Figure 6 shown.

[0034] Based on this, the function generalization and attractor determination method for gene regulatory network proposed in this embodiment includes the following steps:

[0035] Step 1: For each target gene in the gene regulatory network, calculate the Cramer correlation coefficient between the target gene and all its corresponding regulatory genes, and select the regulatory gene with the highest Cramer correlation coefficient value as a guide to set other unobserved values ​​in the truth table of the gene regulatory network.

[0036] like Figure 4 The partial truth table shown in FIG. 1 shows that the output of function f1 corresponds to the state of gene 1. When the input is 00, f1 outputs 0, indicating that the state of gene 1 is 0 at this time. There are two regulatory genes in function f1, namely j1=2 and j2=4. j1=2 means that the first variable of f1 input is the state of gene 2, and j2=4 means that the second variable of f1 input is the state of gene 4. The Cramer correlation coefficients between these two genes and gene 1 are calculated respectively. The Cramer correlation coefficients of the two genes and gene 1 are equal. Then the first regulatory gene j1=2 is selected as the guide by default (if they are not equal, the regulatory gene with a larger coefficient is selected as the guide). According to the observed part of the gene regulatory network truth table, when gene 2 is 0, gene 1 is 0; when gene 2 is 1, gene 1 is 1. Based on this, the function is generalized. When gene 2 is 0, gene 1 is 0 regardless of whether other genes are expressed; when gene 2 is 1, gene 1 is 1 regardless of whether other genes are expressed.

[0037] In order to analyze and compare the proposed algorithm with other existing algorithms, each algorithm is applied to gene regulatory networks with different connectivity, and two indicators, steady-state distribution distance and sensitivity error, are selected as evaluation indicators of the algorithm. It can be seen from the two evaluation indicators that the Cramer generalization algorithm has the best effect. Figure 7 , Figure 8 shown.

[0038] Step 2: Based on the truth table of the gene regulatory network obtained in step 1, the attractor is obtained by traversing each state of the gene.

[0039] The truth table can reflect the overall evolution rules of the gene, so the attractor can be found by traversing each state of the gene based on the truth table.

[0040] When the number of nodes in the gene regulatory network is N, the corresponding total number of states is 2 N , we traverse all states from 000...00 to 111...11: Assume that the attractor set is S and the current state is x t , we obtain its next state x according to the rules of the truth table t+1 , if x t =x t+1 , then judge x t Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x to t Put it into S and start the evolution of the next state; if x t ≠x t+1 , then continue to obtain the next state x t+2, and compare x t+1 and x t+2 , if x t+1 =x t+2 Then judge x t+1 Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x to t+1 Put it into S and start the evolution of the next state; and so on. After traversing all states, the state set in S is the entire attractor.

[0041] Through the above process, we achieved accurate prediction of the truth table of the gene regulatory network and determined the attractor function.

[0042] When the user has a gene regulatory network model and a corresponding truth table, all attractors can be obtained using the method of the present invention, thereby solving the problem of imperfect functions of existing methods and achieving more complete processing of biological sequencing data.

[0043] When the user has a gene regulatory network model and biological sequencing data but does not have a truth table, the method of the present invention is used to predict the truth table through an algorithm based on the Cramer correlation coefficient, and to quickly obtain attractors, and finally directly feed back all the predicted truth tables and attractors to the user.

[0044] In addition, it will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0045] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0048] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0049] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for function generalization and attractor determination in a gene regulatory network, characterized in that: The following steps are involved: Step 1: For each target gene in the gene regulatory network, calculate the Cramer correlation coefficient between the target gene and all its corresponding regulatory genes, and select the regulatory gene with the highest Cramer correlation coefficient value as a guide to set other unobserved values ​​in the truth table of the gene regulatory network; Step 2: Based on the truth table of the gene regulatory network obtained in step 1, the attractor is obtained by traversing each state of the gene.

2. A method for function generalization and attractor determination for gene regulatory networks according to claim 1, characterized in that: In step 2, the specific process of traversing each state of the gene to obtain the attractor is: When the number of nodes in the gene regulatory network is N, the corresponding total number of states is 2 N , traverse all states; Assume that the attractor set is S and the current state is x t , get the next state x according to the rules of the truth table t+1 , if x t =x t+1 , then judge x t Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x to t Put it into S and start the evolution of the next state; If x t ≠x t+1 , then continue to obtain the next state x t+2 , and compare x t+1 and x t+2 , if x t+1 =x t+2 Then judge x t+1 Does it already exist in S? If so, end the evolution of this state and start the evolution of the next state. If not, change state x to t+1 Put it into S and start the evolution of the next state; And so on, after traversing all states, the set of states in S is the entire attractor.

3. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the above-mentioned function generalization and attractor determination method for gene regulatory networks is implemented.

4. A computer device comprising a processor and a memory, characterized in that: When the processor executes the computer program stored in the memory, the above-mentioned function generalization and attractor determination method for gene regulatory networks is implemented.

Images