CN101082944A - Computer simulation method for protein folding procedure based on synthesis algorithm - Google Patents

Abstract

The invention provides a computer simulation method of protein folding process based on comprehensive algorithm. It includes the steps of choosing a non-lattice protein model, determining the energy function, searching for the protein three-dimensional structure with the smallest free energy, simulating the folding process, and data analysis and improvement. The present invention mainly adopts the prediction method of theoretical research, combines the prediction based on knowledge and the prediction based on simulation, so as to break through respective limitations and achieve higher accuracy. From the perspective of computer computing speed and storage capacity, while retaining the main reason for folding, a large number of specific details in the folding process are ignored, thereby greatly improving the computing speed and becoming an effective method suitable for ordinary computers.

Description

Computer Simulation Method of Protein Folding Process Based on Comprehensive Algorithm

(1) Technical field

The invention relates to a method for simulating the protein folding process, which belongs to biological information technology, calculation method and computer virtual reality technology.

(2) Background technology

There are two theoretical methods for predicting the three-dimensional structure of a protein from its amino acid sequence:

One is knowledge-based protein structure prediction, which uses the existing protein structure data in the protein database to establish appropriate parameters according to the principle of pattern matching, and uses these parameters to evaluate, optimize and predict on the basis of protein primary structure. The knowledge-based prediction method is relatively simple and fast, but it needs to find a template with homology that meets the requirements.

The other is the prediction of protein structure based on simulation, which is used to study the folding process, that is, the process of protein folding from a non-functional, dispersed, unfolded protein structure into a functional, natural protein structure, to carry out protein structure prediction. predict. The prediction of protein structure based on simulation requires a large number of free energy equations, which is computationally intensive. But it is the only option when the former cannot search for a suitable template. The methods used for simulation-based protein structure prediction are:

Genetic algorithm has strong global search ability and the ability to solve problems, and it is also simple, versatile and robust, and is suitable for parallel processing. However, genetic algorithm is prone to premature convergence and the result may be non-global optimal Solution and in the later stage of evolution, the search efficiency is low, the convergence speed is slow, and inbreeding occurs.

Particle swarm optimization algorithm is a new kind of stochastic optimization technology, and its idea comes from artificial life and evolutionary computing theory. Optimization is accomplished by particles following individual positions and global extrema. On the one hand, the algorithm sorts the particles, and each particle follows the previous particle to maintain diversity; on the other hand, each particle uses more useful information of other particles to strengthen the cooperation and competition between particles. The results show that the algorithm not only has better convergence accuracy and faster convergence speed, but also can conduct global search more effectively to obtain the optimal solution.

Fuzzy clustering belongs to unsupervised learning in pattern recognition. It does not require training samples and can directly achieve the purpose of automatic classification through machine learning. The most critical technology in pattern recognition is feature extraction. Fuzzy clustering can not only extract features from the original data, but also optimize the selection and dimensionality reduction of features; after extracting features, fuzzy clustering can also provide the nearest neighbor prototype classification device, as well as space division and extraction of fuzzy rules, which help to classify multiple minimum value solutions and find the optimal solution.

In the patent application number 01814465.9, the patent document titled "A Method for Determining the Three-dimensional Structure of a Protein from the Primary Structural Sequence of a Protein" uses a method for predicting protein structure based on knowledge. There must be known The three-dimensional protein structure was used as a reference template.

(3) Contents of the invention

The purpose of the present invention is to provide a combination of knowledge-based prediction and simulation-based prediction, which can achieve higher accuracy, greatly improve the calculation speed, and is suitable for the computer simulation method of the protein folding process based on the comprehensive algorithm of ordinary computers .

The purpose of the present invention is achieved like this:

(3) Non-lattice protein model

Choose the HPN non-lattice model. The HPN non-lattice model divides amino acid molecules into H (hydrophobic, hydrophobic), P (Polar, hydrophilic), N (charged), and X (unknown), and these four types of molecules are classified according to Arranged according to certain rules to form a linear chain, the configuration containing n residues is n-2 horizontal angles $\mathrm{qi}(-p\&le;\mathrm{qi}<p)$ and n-3 vertical angles $(-\frac{p}{2}\&le;b_i<\frac{p}{2})$ Exactly; modeling and optimization by using virtual reality technology combined with 3DSMAX and OPENGL, using small balls to represent amino acids, and cylinders to represent bonds between molecules;

(4) Energy function

For each protein molecule, two types of forces are used to represent the potential energy in the molecule: the main chain bending potential (V1) and the non-covalent bond potential (V2), and the energy function is:

$\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span>$

in $V_i\left({\mathrm{\&alpha;}}_i\right)=\frac{1}{4}(1-\mathrm{coa}{\mathrm{\&alpha;}}_i),$ α _i refers to the angle between adjacent bonds, i=2,...,n-1.

V ₂ (r _ij , ξ _i , ξ _j )=4[r _ij ^-12 +C(ξ _i , ξ _j )r _ij ^-16 ]

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 4</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; sign</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

in $\mathrm{sign}\left(x\right)=\left\{\begin{array}{cc}1& x\&Greater\; Equal;0\\ 0& x>0\end{array}\right.;$

(3) Search for the three-dimensional protein structure with the smallest free energy

Based on the energy of each determined three-dimensional space structure calculated above, search the structure space in combination with genetic algorithm, particle swarm algorithm and fuzzy algorithm, and find out the three-dimensional structure corresponding to the minimum energy;

(4) Simulate the folding process

After searching for the natural protein space structure with the smallest free energy, record the space position and properties of each amino acid that makes up the protein, and record the free energy of the protein, and then perform simulations based on the obtained data, combined with VC++ and OPENGL, to The simulated results are rotated, translated, zoomed, and the protein folding process is recorded;

(5) Data analysis and improvement

Compare the obtained results with the existing three-dimensional structure in PDB, adjust the protein structure by using the calculation formula of the root mean square deviation between the predicted conformation and its natural structure, and improve the adopted model and algorithm.

The present invention can also include such features:

1. The specific steps for finding the three-dimensional structure corresponding to the minimum energy are as follows:

Step 1: randomly generate N conformations as the initial population, and determine the initial parameters;

Step 2: Calculate the probability that each conformation is selected for genetic manipulation, and agree to perform random selection based on this probability;

Step 3: Use particle swarm optimization to select crossover and mutation operators;

Step 4: Use the crossover operator and the mutation operator to perform crossover and mutation operations on the selected conformation, and generate a new population;

Step 5: Evaluate the fitness of the group and retain individuals with higher fitness;

Step 6: If the stopping criterion is not met, repeat steps 2 to 5. Otherwise, output the conformation with the smallest energy value in the group;

Step 7: Classify the obtained multiple minimum values to find the minimum value.

2. The formula for calculating the root mean square deviation is:

$\mathrm{<span\; class="notranslate">\; RMSD</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; .</span>$

The present invention mainly adopts the theoretical research prediction method, and combines the prediction based on knowledge and the prediction based on simulation, so as to break through respective limitations and achieve higher accuracy. From the perspective of computer computing speed and storage capacity, while retaining the main reason for folding, a large number of specific details in the folding process are ignored, thereby greatly improving the computing speed and becoming an effective method suitable for ordinary computers.

Using theoretical prediction and computer molecular folding simulation to determine some intermediate conformations, and then based on these conformations, gradually approach the final conformation. Extend the existing computer virtual reality technology modeling theory and method to the microscopic world to realize the simulation of protein folding process.

Here, the genetic algorithm, particle swarm optimization and fuzzy algorithm are combined, the genetic algorithm is used to optimize the population and connection weights, and then the particle swarm optimization algorithm is used to find the optimal solution to search for the space structure with the smallest global energy, and the fuzzy clustering algorithm is used to analyze the solution. Classification and modification.

The main work includes:

· Extract the main parameters that affect the formation process of the molecular space structure.

· Establish an interactive visual environment for the formation process of protein molecular spatial structure.

The present invention designs a dynamic tool for recording protein folding in the system. After obtaining the simulation results of protein folding, the tool can be started to display the dynamic process of folding.

When the protein is synthesized from the ribosome, the amino acid chain in the loose coil state without physiological activity folds into a specific conformation with physiological activity, and the time for this process is extremely short. So that it is impossible to observe its specific dynamic process. Therefore, using computer to simulate the dynamic process of protein folding will help to increase the understanding of protein folding mechanism, thereby providing useful information for protein molecular design.

The invention is also applicable to the prediction of the three-dimensional structure of the unknown protein, and can simulate and record the folding dynamic process of the predicted protein.

(4) Description of drawings

Fig. 1 is a process block diagram of the present invention;

Fig. 2 is a system structure diagram of the present invention.

(5) Specific implementation methods

The present invention is described in more detail below in conjunction with accompanying drawing example:

In conjunction with Fig. 1 and Fig. 2, method of the present invention is realized on the system structure built in Fig. 2, and its concrete method is:

1. Non-lattice protein model

The HPN non-lattice model divides amino acid molecules into H (hydrophobic, hydrophobic), P (Polar, hydrophilic), N (charged), X (unknown), and these four types of molecules are arranged according to certain rules. Linear chain, n-2 horizontal angles in configuration of n residues $\mathrm{qi}(-p\&le;\mathrm{qi}<p)$ and n-3 vertical angles $(-\frac{p}{2}\&le;b_i<\frac{p}{2})$ indeed. Modeling and optimization are carried out by using virtual reality technology combined with 3DSMAX and OPENGL. Small balls are used to represent amino acids, and cylinders are used to represent bonds between molecules.

2. Energy function

For each protein molecule, two types of forces are used to represent the potential energy in the molecule: the main chain bending potential (V1) and the non-covalent bond potential (V2). V1 has nothing to do with the sequence itself, and V2 changes with the sequence. , changes in every pair of residues in the backbone that are not in direct contact will affect V2. , the energy function is:

$\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span>$

in $V_i\left({\mathrm{\&alpha;}}_i\right)=\frac{1}{4}(1-\mathrm{coa}{\mathrm{\&alpha;}}_i),$ α _i refers to the angle between adjacent bonds, i=2,...,n-1.

V ₂ (r _ij , ξ _i , ξ _j )=4[r _ij ^-12 +C(ξ _i , ξ _j )r _ij ^-16 ]

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 4</span>}}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; sign</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&xi;</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

in $\mathrm{sign}\left(x\right)=\left\{\begin{array}{cc}1& x\&Greater\; Equal;0\\ 0& x>0\end{array}\right.;$

3. Search for the three-dimensional protein structure with the smallest free energy

Based on the energy of each determined three-dimensional space structure calculated above, search the structure space by combining genetic algorithm, particle swarm algorithm and fuzzy algorithm to find out the three-dimensional structure corresponding to the minimum energy.

Specific steps are as follows:

Step 1: randomly generate N conformations as the initial population, and determine the initial parameters.

Step 2: Calculate the probability that each conformation is selected for genetic manipulation, and agree to random selection based on this probability.

Step 3: Use particle swarm optimization to select crossover and mutation operators.

Step 4: Use the crossover operator and the mutation operator to perform crossover and mutation operations on the selected conformation (namely, the chromosome), and generate a new population.

Step 5: Evaluate the fitness of the group and retain individuals with higher fitness.

Step 6: If the stopping criterion is not met, repeat steps 2 to 5. Otherwise, output the conformation with the smallest energy value in the clique.

The seventh step is to classify the obtained minimum values to find out the minimum values.

4. Simulate the folding process

After searching for the natural protein space structure with the smallest free energy, record the space position and properties of each amino acid that makes up the protein, and record the free energy of the protein, and then perform simulations based on the obtained data, combined with VC++ and OPENGL, to The simulation results are rotated, translated, zoomed, and the protein folding process is recorded.

5. Data analysis and improvement

Compare the obtained results with the existing three-dimensional structure in PDB, adjust the protein structure by using the calculation formula of the root mean square deviation between the predicted conformation and its natural structure, and improve the adopted model and algorithm. The formula for calculating root mean square deviation is:

$\mathrm{<span\; class="notranslate">\; RMSD</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}^{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; .</span>$

Claims (3)

1, a kind of computer simulation method of the protein folding procedure based on integration algorithm is characterized in that:

(1) protein model of the non-lattice point of selection

Select the non-lattice model of HPN, the non-lattice model of HPN is divided into H, P, N, X four classes to amino acid molecular, and by this four kinds of elements according to certain regularly arranged linear chain that forms, the configuration that contains n residue is a n-2 horizontal angle

With n-3 vertical angle $(-\frac{p}{2}\&le;b_i<\frac{p}{2})$ True; By utilization

Virtual reality technology is carried out modeling and optimization in conjunction with 3DSMAX and OPENGL, with bead represented amino acid molecule, represents key between the molecule with cylinder;

(2) determine energy function

To each protein molecule, represent intramolecular potential energy with two class acting forces: crooked gesture (V1) of main chain and non-covalent bond effect gesture (V2), energy function is:

$\mathrm{\&phi;}=\underset{i=2}{\overset{n-1}{\mathrm{\&Sigma;}}}{V}_1\left({\mathrm{\&alpha;}}_i\right)+\underset{i=2}{\overset{n-2}{\mathrm{\&Sigma;}}}\underset{j=i-2}{\overset{n}{\mathrm{\&Sigma;}}}{V}_2(r_{\mathrm{ij}},{\mathrm{\&xi;}}_r,{\mathrm{\&xi;}}_j)$

Wherein $V_1\left({\mathrm{\&alpha;}}_i\right)=\frac{1}{4}(1-\cos {\mathrm{\&alpha;}}_i),$ α _iRefer to the angle between adjacent key, i=2 ..., n-1.

V ₂(r _ij，ξ _i，ξ _j)＝4[r _ij ^-12+C(ξ _i，ξ _j)r _ij ^-16]

$C({\mathrm{\&xi;}}_i,{\mathrm{\&xi;}}_j)=\frac{1}{4}{(2-|{\mathrm{\&xi;}}_i+{\mathrm{\&xi;}}_j\left|\right)}^2+\mathrm{sign}\left(\right|{\mathrm{\&xi;}}_i+{\mathrm{\&xi;}}_j|-4)$

Wherein $\mathrm{sign}\left(x\right)=\left\{\begin{array}{cc}1& x\&GreaterEqual;0\\ 0& x<0\end{array}\right.$

(3) protein three-dimensional structure of search free energy minimum

By the energy of each three-D space structure of determining of calculating above, in conjunction with genetic algorithm, particle cluster algorithm and fuzzy algorithm are searched for structure space, find out the minimum pairing three-dimensional structure of energy;

(4) simulation folding process

After searching out the native protein space structure of free energy minimum, each amino acid locus and attribute of this protein formed in record, and note the free energy of this protein, then according to the data that obtain, carry out emulation in conjunction with VC++ and OPENGL, can be rotated translation Simulation result, convergent-divergent, and the Protein Folding process recorded;

(5) data analysis and improvement

With the result who draws, compare with existing three-dimensional structure among the PDB, utilize the computing formula of the root-mean-square-deviation of prediction conformation and its natural structure, adjust protein structure, and to adopting model and algorithm to improve.

2, the computer simulation method of the protein folding procedure based on integration algorithm according to claim 1 is characterized in that: the described concrete steps of finding out the minimum pairing three-dimensional structure of energy are:

The 1st step: produce N conformation at random as initial population, and definite initial parameter;

The 2nd step: calculate the selected probability that carries out genetic manipulation of each conformation, and agreement is selected at random according to this probability;

The 3rd step: intersect and the choosing of mutation operator with particle cluster algorithm;

The 4th step: utilize crossover operator and mutation operator that selected conformation is intersected and mutation operation, and produce new population;

The 5th step: estimate the fitness of colony, keep the higher individuality of fitness;

The 6th step:, then repeat second and went on foot for the 5th step if do not satisfy stopping criterion.Otherwise, the conformation of energy value minimum in the output group;

The 7th step: a plurality of minimal values that draw are classified with fuzzy algorithm, find out minimal value.

3, the computer simulation method of the protein folding procedure based on integration algorithm according to claim 1 and 2, it is characterized in that: the computing formula of described root-mean-square-deviation is:

$\mathrm{RMSD}={\left(\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{({x_i}^{\left(a\right)}-{x_i}^{\left(b\right)})}^2\right)}^{\frac{1}{2}}$

Images