CN115691669B - Protein structure classification system based on quantum convolution neural network - Google Patents
Protein structure classification system based on quantum convolution neural network Download PDFInfo
- Publication number
- CN115691669B CN115691669B CN202310000900.5A CN202310000900A CN115691669B CN 115691669 B CN115691669 B CN 115691669B CN 202310000900 A CN202310000900 A CN 202310000900A CN 115691669 B CN115691669 B CN 115691669B
- Authority
- CN
- China
- Prior art keywords
- quantum
- protein
- module
- loss function
- pooling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 102000004169 proteins and genes Human genes 0.000 title claims abstract description 107
- 108090000623 proteins and genes Proteins 0.000 title claims abstract description 107
- 238000013528 artificial neural network Methods 0.000 title abstract description 10
- 238000011176 pooling Methods 0.000 claims abstract description 53
- 150000001413 amino acids Chemical class 0.000 claims abstract description 22
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 12
- 125000003275 alpha amino acid group Chemical group 0.000 claims description 30
- 239000002096 quantum dot Substances 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 14
- 238000012360 testing method Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 4
- 238000005259 measurement Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 238000011161 development Methods 0.000 abstract description 2
- 239000003814 drug Substances 0.000 abstract description 2
- 229940079593 drug Drugs 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 30
- 238000010586 diagram Methods 0.000 description 10
- 238000000034 method Methods 0.000 description 7
- 239000011159 matrix material Substances 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- FFEARJCKVFRZRR-BYPYZUCNSA-N L-methionine Chemical compound CSCC[C@H](N)C(O)=O FFEARJCKVFRZRR-BYPYZUCNSA-N 0.000 description 1
- AYFVYJQAPQTCCC-UHFFFAOYSA-N Threonine Natural products CC(O)C(N)C(O)=O AYFVYJQAPQTCCC-UHFFFAOYSA-N 0.000 description 1
- 239000004473 Threonine Substances 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 229930182817 methionine Natural products 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000013612 plasmid Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000000455 protein structure prediction Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a protein structure classification system based on a quantum convolutional neural network, wherein the system comprises: the coding module of protein sequence amino acid characteristic data is used for extracting and reading protein sequence information and corresponding structural information from the protein structure classification data set; a quantum convolution and pooling module for effecting classification of the protein results by parameterized quantum gates; a build loss function module for obtaining a loss function for characterizing the system performance; an update quantum wire parameter module for updating the quantum wire parameters. Compared with the prior art, the invention realizes an efficient quantum computer convolution neural network system, can efficiently classify protein structures, and greatly accelerates the prediction of the protein structures and the development of drugs by using the model used by the system.
Description
Technical Field
The invention belongs to the technical field of quantum computers, and particularly relates to a protein structure classification system based on a quantum convolutional neural network.
Background
Proteins are the main players of life activities, and their functions and structures are closely related. The accuracy of protein structure prediction can be greatly improved by effectively classifying the protein structures. At present, the existing classical machine learning algorithm (neural network, support vector machine, random forest and the like) does much work in the direction of protein structure classification. These efforts first lead to the preprocessing of data based on a protein data set, typically having sequence information, secondary structure information, mutation information, etc. The traditional computer can store the protein data information to the classical bit by means of one-hot coding and the like. The data set is then divided into a training data set and a test data set. On the training data set, extracting features through a machine learning algorithm, training to obtain a prediction model, and then testing the accuracy of the test model on the data set.
The prior art uses machine learning models on classical computers to classify protein structures. The classical computer uses classical bits for calculation, which is different from the quantum ratio used by the quantum computer. The coding patterns of proteins in classical computers do not reflect well the intrinsic information of proteins. The amino acid sequence in the protein has time sequence property, and if the protein is coded on a quantum bit by using a quantum computer, the time sequence property of the amino acid sequence can be well embodied by the entanglement property of the quantum bit. In addition, protein databases are large in number and variety, and classical computers have limited capabilities of storing data, calculating data and the like, so that excessively large data sets cannot be processed.
Disclosure of Invention
In view of the above-mentioned drawbacks in the prior art, the present invention provides a protein structure classification system based on quantum convolutional neural network, which includes: a coding module of protein sequence amino acid characteristic data, a quantum convolution and pooling module, a construction loss function module and an update quantum line parameter module,
the coding module of the protein sequence amino acid characteristic data is used for extracting and reading protein sequence information and corresponding structural information from the protein structure classification data set;
a quantum convolution and pooling module for effecting classification of the protein results by parameterized quantum gates;
a build loss function module for obtaining a loss function for characterizing the system performance;
an update quantum wire parameter module for updating the quantum wire parameters.
Wherein the protein structure classification dataset is classified according to 99: the scale of 1 is divided into a training data set and a test data set.
Wherein the quantum convolution and pooling module comprises:
the quantum convolution layer basic unit is used for evolving the quantum state loaded with the protein sequence characteristic information;
a quantum-pooling layer basic unit for mapping information of two qubits onto one qubit.
Wherein the quantum convolution and pooling module is further configured to measure the Polly Z expectation of the last qubit as the final predictor of the protein structure classification by alternating the quantum convolution layer and the quantum pooling layer until only one qubit remains.
Wherein the loss function module is used for constructing protein amino acid sequence characteristic data in each batch b
The input is based on the quantum convolution and pooling module, and each protein amino acid sequence obtains a predicted value through the quantum convolution and pooling module
And then obtaining a loss function for characterizing the system performance by solving the mean square error of the predicted value of all protein amino acid sequences of each batch relative to the real label of the predicted value.
Wherein the loss function is expressed by the following equation:
wherein
K is the number of amino acid sequences of the protein contained in the batch b.
The quantum line parameter updating module is specifically used for solving the analytic gradient of the loss function relative to the quantum line parameters based on the parameterized circuit movement rule, and then updating the quantum line parameters.
Wherein, the calculating the analytic gradient of the loss function with respect to the quantum line parameter based on the parameterized circuit movement rule specifically includes:
hypothetical measurement operator
In parametric quantum wires
The expected value of (A) can be expressed as
Wherein,
representing parameterized quantum wires composed of quantum convolutional layers and pooling layers,
representing parameters in the quantum convolutional layer and the pooling layer;
then the expected value function
With respect to parameterized quantum line parameters
Can be expressed as
Wherein the system trains a plurality of epochs using the training data set until a desired accuracy is reached.
Compared with the prior art, the invention realizes an efficient quantum computer convolution neural network system, can efficiently classify protein structures, and the model used by the system can greatly accelerate the prediction of the protein structures and the development of drugs.
Drawings
The above and other objects, features and advantages of exemplary embodiments of the present application will become readily apparent from the following detailed description, which proceeds with reference to the accompanying drawings. Several embodiments of the present application are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which like reference numerals refer to similar or corresponding parts and in which:
FIG. 1 is PSSM matrix data showing amino acid signature properties of proteins according to embodiments of the present invention;
fig. 2 (a) is a diagram showing a quantum wire encoding 20-dimensional data of a single amino acid of a protein sequence onto 10 qubits according to an embodiment of the present invention;
FIG. 2 (b) is a block diagram showing an encoded implementation of amino acid sequence profile data of the entire protein according to an embodiment of the present invention;
FIG. 3 (a) is a block diagram showing a basic cell implementation of a quantum convolutional layer according to an embodiment of the present invention;
FIG. 3 (b) is a block diagram showing a basic cell implementation of a quantum pooling layer according to an embodiment of the present invention;
FIG. 4 is a block diagram illustrating a protein structure classification system based on a quantum convolutional neural network according to an embodiment of the present invention;
fig. 5 is a block diagram illustrating a protein structure classification system based on a quantum convolutional neural network according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail with reference to the accompanying drawings, and it is apparent that the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terminology used in the embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the examples of the present invention and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise, and "a plurality" typically includes at least two.
It should be understood that while the terms first, second, third, etc. may be used in embodiments of the present invention to describe … …, these … … should not be limited to these terms. These terms are used only to distinguish … …. For example, a first … … may also be referred to as a second … …, and similarly, a second … … may also be referred to as a first … …, without departing from the scope of embodiments of the present invention.
It should be understood that the term "and/or" as used herein is merely one type of association that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" herein generally indicates that the former and latter related objects are in an "or" relationship.
The words "if", as used herein, may be interpreted as "at … …" or "at … …" or "in response to a determination" or "in response to a detection", depending on the context. Similarly, the phrase "if determined" or "if detected (a stated condition or event)" may be interpreted as "upon determining" or "in response to determining" or "upon detecting (a stated condition or event)" or "in response to detecting (a stated condition or event)", depending on the context.
It is also noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that an article or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such article or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another like element in a commodity or device comprising the element.
The related terms of the present application:
PDB (Protein Data Bank) Protein database
NISQ (noise intermediate-scale Quantum) noise-containing mesoscale quantum computer
SCOP (Structural Classification of Proteins) protein structure Classification database
PSSM (Position-specific score matrix) site-specific scoring matrix
The quantum convolution neural network loads the characteristic vector representing the protein sequence amino acid into a quantum state based on an amplitude coding mode, and then processes the quantum state containing the protein sequence amino acid characteristic information through a quantum convolution layer and a quantum pooling layer which respectively correspond to the classical convolution and the pooling. In the process, the dimension of the quantum bit is continuously reduced, finally, one quantum bit is measured, the information obtained by measurement and the real label of the structural classification of the protein are combined into a loss function, and the parameters are continuously updated according to the loss function until a satisfactory threshold value is reached.
Alternative embodiments of the present invention are described in detail below with reference to the accompanying drawings.
The first embodiment,
The invention provides a protein structure classification system based on a quantum convolutional neural network, which comprises the following components: the system comprises a protein sequence amino acid characteristic data coding module, a quantum convolution and pooling module, a loss function construction module and a quantum line parameter updating module.
Wherein, the coding module of the protein sequence amino acid characteristic data is used for extracting protein sequence information and corresponding structure information from a protein structure classification data set (SCOP and the like). The data set was normalized to 99: the scale of 1 is divided into a training data set and a test data set. For amino acid sequence information in the dataset, twenty-dimensional vectors were encoded using the PSSM method. By using
The gyrotron gate acts on qubits (here we use 10 qubits) to load 20-dimensional protein sequence amino acid signature data onto the amplitude of the quantum states.
A quantum convolution and pooling module for building a quantum convolution layer elementary unit and a quantum pooling layer elementary unit, respectively, that can act on two quantum bits, by means of parameterized quantum gates. Then, the quantum convolution layer is formed by the action of the quantum convolution layer basic unit on every two quantum bit pairs of the quantum system, so that the evolution of the quantum state of the previous module loaded with the protein sequence amino acid characteristic information is carried out. Subsequently, a quantum pooling basic unit acts on each quantum bit pair of the quantum system to form a quantum pooling layer, and information of two quantum bits is mapped to one quantum bit, wherein the number of the quantum bits containing high-level information is 5. And then alternately acting the quantum convolution layer and the quantum pooling layer on the remaining 5 qubits, wherein the qubits containing high-level information are 3 qubits, repeating the steps until only one qubit is left by alternately acting the quantum convolution layer and the quantum pooling layer, and finally measuring the Pally Z expected value of the last qubit to be used as a final predicted value of the protein structure classification.
Constructing a loss function module for characterizing the amino acid sequence of the protein in each batch b
Inputting into quantum convolution neural network built based on last module, each proteinThe amino acid sequence of the plasmid can be predicted by the last module
And then obtaining a loss function for characterizing the performance of the model by solving the mean square error of the predicted value of all protein amino acid sequences of each batch relative to the real label of the predicted value.
And the updating quantum line parameter module is used for solving the analytic gradient of the loss function of the last module relative to the quantum line parameters based on the existing parameterized circuit movement rule, then updating the quantum line parameters by using a classical computer, finally training a plurality of epochs by using a protein amino acid sequence training data set, and stopping updating after the expected accuracy is reached.
Example II,
In order to further illustrate the method for predicting the protein structure based on the quantum convolution neural network, the following embodiments are provided:
an encoding module for protein sequence amino acid feature data based on PSSM matrix data characterizing the amino acid feature attributes of each protein, as shown in fig. 1, where each letter represents one of 20 amino acids and each amino acid has a feature vector dimension of 20. By passing
The rotaron gate loads protein amino acid sequence data onto the amplitude of the quantum state (here we use 10 quantum bits) based on the characteristic attribute data of each protein amino acid, with the specific quantum wires as shown in fig. 2. Fig. 2 (a) shows quantum wires encoding 20-dimensional data of single amino acids of a protein sequence onto 10 qubits. FIG. 2 (b) shows the coding implementation of the characteristic data of the entire protein amino acid sequence, here taking the protein amino acid sequence shown in the figure as an example, it can be seen from the figure that firstly the 20-dimensional characteristic data characterizing methionine (M) is coded onto the amplitudes of 10 qubit quantum states in the manner of FIG. 2 (a). Then encoding the characteristic data of threonine (T) to the quantum state, and so on until the wholeUntil the protein sequence is encoded.
A quantum convolution and pooling module comprising a quantum convolution layer elementary unit and a quantum pooling layer elementary unit acting on two quantum bits, the corresponding quantum wires of which are shown in fig. 3, fig. 3 (a) being a quantum convolution layer elementary unit implementation and fig. 3 (b) being a quantum pooling layer elementary unit implementation. Based on the quantum convolution layer basic unit and the quantum convolution pooling layer basic unit, the final predicted value can be obtained by alternately acting the quantum convolution layer and the pooling layer. Specifically, as shown in fig. 4, a block C in the figure represents the quantum convolution layer basic unit in fig. 3 (a), a block P represents the quantum pooling layer basic unit, a block C portion in a dotted line frame represents the first quantum convolution layer, and a block P portion in a dotted line frame represents the first quantum pooling layer. As shown in FIG. 4, we obtain the final predicted value as whether the protein amino acid sequence is an alpha helix structure or not by alternately acting quantum convolution and quantum pooling layers so that the information containing the amino acid characteristics of the protein sequence is finally loaded on one qubit and measuring the Pally Z expected value of the last qubit, wherein the truncated qubit indicates the action of the basic unit without the quantum convolution layer and the pooling layer.
Constructing a loss function module which characterizes the amino acid sequence of the protein in each batch b
Inputting the protein into a quantum convolution neural network built by the previous module, and obtaining a corresponding predicted value for each protein amino acid sequence
. Finally, the predicted values of all protein amino acid sequences in the batch are calculated
The mean square error between the predicted value and the true value of the protein amino acid sequence is calculated by combining the true label whether the predicted value is corresponding to the alpha helical structure or not, thereby obtaining the representation quantum volumeThe loss function of the product neural network model performance and the expression of the mean square error loss function are shown as the following formula.
Wherein
K is the number of amino acid sequences of the protein contained in the batch b.
Update quantum line parameter module: firstly, a measuring operator \ hat M \]In parametric quantum wires
The expected value of (A) can be expressed as
Then the expected value function
With respect to parameterized quantum line parameters
Can be expressed as
In the above formula (3)
Representing parameterized quantum wires composed of quantum convolutional layers and pooling layers,
representing parameters in the quantum convolutional layer and the pooling layer.
The above method is called a parameter-shifting rule for solving the gradient of the parametric quantum wire with respect to the analysis of the desired value of the operator.
Through the parameter moving rule, the analytical gradient of the mean square error loss function of the previous module, namely the formula (1), on the quantum circuit parameters of the quantum convolution layer and the pooling layer can be obtained. The parameters are then updated by a gradient descent method using a classical computer. And finally, training a plurality of epochs according to a training data set of protein amino acid sequence characteristic data until the protein structure classification predicted based on the quantum convolution neural network provided by the patent is accurate to a desired degree.
EXAMPLE III
As shown in fig. 5, the present invention provides a protein structure classification system based on quantum convolutional neural network, which includes: a coding module of protein sequence amino acid characteristic data, a quantum convolution and pooling module, a construction loss function module and an update quantum circuit parameter module,
the coding module of the protein sequence amino acid characteristic data is used for extracting and reading protein sequence information and corresponding structural information from the protein structure classification data set;
a quantum convolution and pooling module for effecting classification of the protein results by parameterized quantum gates;
a build loss function module for obtaining a loss function for characterizing the system performance;
an update quantum wire parameter module for updating the quantum wire parameters.
Wherein the protein structure classification dataset is classified according to 99: the scale of 1 is divided into a training data set and a test data set.
Wherein the quantum convolution and pooling module comprises:
the quantum convolution layer basic unit is used for evolving the quantum state loaded with the protein sequence characteristic information;
a quantum-pooling layer basic unit for mapping information of two qubits onto one qubit.
Wherein the quantum convolution and pooling module is further configured to measure the pauli Z expected value of the last qubit as the final prediction value for the protein structure classification by alternating the quantum convolution layer and the quantum pooling layer until only one qubit remains.
Wherein the loss function module is used for constructing protein amino acid sequence characteristic data in each batch b
The input is based on the quantum convolution and pooling module, and each protein amino acid sequence obtains a predicted value through the quantum convolution and pooling module
And then obtaining a loss function for characterizing the system performance by solving the mean square error of the predicted value of all protein amino acid sequences of each batch relative to the real label of the predicted value.
Wherein the loss function is expressed by the following equation:
wherein k is the number of amino acid sequences of the protein contained in batch b.
The quantum line parameter updating module is specifically used for solving the analytic gradient of the loss function relative to the quantum line parameters based on the parameterized circuit movement rule, and then updating the quantum line parameters.
Wherein, the calculating the analytic gradient of the loss function with respect to the quantum line parameter based on the parameterized circuit movement rule specifically includes:
hypothetical measurement operator
In parametric quantum wires
The expected value of (A) can be expressed as
Wherein,
representing parameterized quantum wires composed of quantum convolutional layers and pooling layers,
representing parameters in the quantum convolutional layer and the pooling layer;
then the expected value function
With respect to parameterized quantum line parameters
Can be expressed as
Wherein the system trains a plurality of epochs using the training data set until a desired accuracy is reached.
Example four,
Embodiments of the present invention provide a non-volatile computer storage medium, where computer-executable instructions are stored, and the computer-executable instructions may perform the method steps described in the above embodiments.
It should be noted that the computer readable medium mentioned above in the present application may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present application, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In this application, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
The computer readable medium may be embodied in the electronic device; or may be separate and not incorporated into the electronic device.
Computer program code for carrying out operations for aspects of the present application may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local Area Network (AN) or a Wide Area Network (WAN), or the connection may be made to AN external computer (for example, through the internet using AN internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units described in the embodiments of the present invention may be implemented by software or hardware. Where the name of an element does not in some cases constitute a limitation on the element itself.
The foregoing describes preferred embodiments of the present invention, and is intended to provide a clear and concise description of the spirit and scope of the invention, and not to limit the same, but to include all modifications, substitutions, and alterations falling within the spirit and scope of the invention as defined by the appended claims.
Claims (7)
1. A system for protein structure classification based on a quantum convolutional neural network, comprising: a coding module of protein sequence amino acid characteristic data, a quantum convolution and pooling module, a construction loss function module and an update quantum line parameter module,
a coding module for said protein sequence amino acid signature data for reading protein sequence information and corresponding structural information from a protein structure classification dataset;
a quantum convolution and pooling module for effecting classification of the protein structure by a parameterized quantum gate;
a build loss function module for obtaining a loss function for characterizing the system performance;
an update quantum line parameter module for updating the quantum line parameters;
wherein the quantum convolution and pooling module comprises:
the quantum convolution layer basic unit is used for evolving the quantum state loaded with the protein sequence characteristic information;
a quantum-pooling layer basic unit for mapping information of two qubits onto one qubit;
the quantum convolution and pooling module is also used to measure the Poyley Z expectation of the last qubit as the final predictor of the protein structure classification by alternating the quantum convolution layer and the quantum pooling layer until only one qubit remains.
2. The quantum convolutional neural network-based protein structure classification system of claim 1, wherein the protein structure classification dataset is classified according to a 99: the scale of 1 is divided into a training data set and a test data set.
3. The system for classifying protein structures based on quantum convolutional neural network as defined in claim 1, wherein said building loss function module is specifically configured to characterize the amino acid sequence of proteins in each batch b
The input is based on the quantum convolution and pooling module, and each protein amino acid sequence obtains a predicted value through the quantum convolution and pooling module
And then obtaining a loss function for characterizing the system performance by solving the mean square error of the predicted value of all protein amino acid sequences of each batch relative to the real label of the predicted value.
4. The system for classifying protein structures based on quantum convolutional neural network as claimed in claim 3, wherein the loss function is expressed as:
wherein
K is the number of amino acid sequences of the protein contained in the batch b.
5. The system of claim 1, wherein the update quantum wire parameters module is further configured to apply an analytical gradient of the loss function with respect to the quantum wire parameters based on a parameterized circuit motion law, and then update the quantum wire parameters.
6. The system of claim 1, wherein the step of solving the analytical gradient of the loss function with respect to the quantum wire parameters based on parameterized circuit motion rules comprises:
postulated measurement operator
In parametric quantum wires
The expected value of (A) can be expressed as
Wherein,
representing parameterized quantum wires composed of quantum convolutional layers and pooling layers,
representing parameters in the quantum convolutional layer and the pooling layer;
then the expected value function
The gradient for a parametric quantum wire parameter can be expressed as
7. The system for classifying protein structures based on quantum convolutional neural network as claimed in claim 2, wherein said system trains a plurality of epochs using said training data set until a desired accuracy is reached.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310000900.5A CN115691669B (en) | 2023-01-03 | 2023-01-03 | Protein structure classification system based on quantum convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310000900.5A CN115691669B (en) | 2023-01-03 | 2023-01-03 | Protein structure classification system based on quantum convolution neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115691669A CN115691669A (en) | 2023-02-03 |
| CN115691669B true CN115691669B (en) | 2023-03-17 |
Family
ID=85057096
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310000900.5A Active CN115691669B (en) | 2023-01-03 | 2023-01-03 | Protein structure classification system based on quantum convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115691669B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116304883A (en) * | 2023-04-25 | 2023-06-23 | 电子科技大学 | Quantum convolution-based molecular diagram classification method, system, storage medium and terminal |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019094647A1 (en) * | 2017-11-08 | 2019-05-16 | Stc. Unm. | System and methods for graphic encoding of macromolecules for efficient high-throughput analysis |
| CN112289370B (en) * | 2020-12-28 | 2021-03-23 | 武汉金开瑞生物工程有限公司 | Protein structure prediction method and device |
| CN113159239B (en) * | 2021-06-28 | 2021-09-28 | 北京航空航天大学 | Method for processing graph data by quantum graph convolutional neural network |
| CN114444701B (en) * | 2022-02-01 | 2023-10-27 | 上海图灵智算量子科技有限公司 | Training quantum circuit and data embedding method |
-
2023
- 2023-01-03 CN CN202310000900.5A patent/CN115691669B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115691669A (en) | 2023-02-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109657805B (en) | Hyper-parameter determination method, device, electronic equipment and computer readable medium | |
| CN110995459A (en) | Abnormal object identification method, device, medium and electronic equipment | |
| CN115691669B (en) | Protein structure classification system based on quantum convolution neural network | |
| CN117824756B (en) | Building structure monitoring method, equipment and storage medium | |
| CN115618269B (en) | Big data analysis method and system based on industrial sensor production | |
| CN112990530B (en) | Regional population quantity prediction method, regional population quantity prediction device, electronic equipment and storage medium | |
| CN116452333A (en) | Construction method of abnormal transaction detection model, abnormal transaction detection method and device | |
| CN115858388A (en) | Test case priority ordering method and device based on variation model mapping chart | |
| CN118136112A (en) | Polypeptide hemolysis prediction method integrating multi-view feature deep ensemble learning | |
| KR102406375B1 (en) | An electronic device including evaluation operation of originated technology | |
| US20200279148A1 (en) | Material structure analysis method and material structure analyzer | |
| CN115691654B (en) | Method for predicting antibacterial peptide of quantum gate-controlled circulating neural network based on fewer parameters | |
| CN111582313A (en) | Sample data generation method and device and electronic equipment | |
| CN111582647A (en) | User data processing method and device and electronic equipment | |
| JP6701467B2 (en) | Learning device and learning method | |
| US20210295136A1 (en) | Improvement of Prediction Performance Using Asymmetric Tanh Activation Function | |
| CN115456073B (en) | Long-short-term memory-based modeling analysis method for generating type countermeasure network model | |
| Ramachandran et al. | An Accurate Prediction of Medical Insurance Cost Using Forest Regression Algorithms | |
| US11100321B2 (en) | Information processing method and information processing system | |
| CN114678083A (en) | Training method and prediction method of chemical genetic toxicity prediction model | |
| CN111276248B (en) | State determination system and electronic device | |
| Thapa | Using ROC-curve to Illustrate the Use of Binomial Logistic Regression Model in Wine Quality Analysis | |
| CN118821732A (en) | Data processing method, apparatus, computer device and readable storage medium | |
| Subair et al. | Receiver Operating Characteristic Curves in Binary Classification of Protein Secondary Structure Data | |
| CN118861282A (en) | Enterprise financial statement automatic generation analysis system and method based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Hu Yongmei Inventor after: Liu Haijian Inventor after: Geng Yongzhong Inventor after: Li Ning Inventor after: Yang Yusheng Inventor after: Zhao Lixiang Inventor after: Cui Guolong Inventor before: Hu Yongmei Inventor before: Liu Haijian Inventor before: Geng Yongzhong Inventor before: Li Ning Inventor before: Yang Yusheng Inventor before: Zhao Lixiang Inventor before: Cui Guolong |