Method, device and medium for predicting non-coding RNA coding potential

Abstract

In the disclosed embodiment, a method, device and medium for predicting the coding potential of non-coding RNA are provided, which belongs to the field of bioinformatics technology, specifically comprising: generating a codon vector; obtaining all possible combinations of codons, and generating a non-coding RNA sequence representation matrix using a vocabulary; inputting the non-coding RNA sequence representation matrix into a bidirectional LSTM network and an attention mechanism PBLA to deeply extract the potential feature information of the transcript sequence, and obtain the transcript embedding vector of the non-coding RNA; step 4, calculating the mutual information between the features to be sorted and the prediction results by a correlated redundant feature selection algorithm based on mutual information, obtaining the feature importance ranking by calculating the correlation and redundancy, taking the feature subset with the best performance as the final feature, and splicing the final feature with the transcript embedding vector to predict the coding potential. Through the scheme disclosed in the present invention, the prediction accuracy and generalization performance are improved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of bioinformatics technology, and in particular to a method, device, and medium for predicting the coding potential of non-coding RNA. Background Art

[0002] Currently, noncoding RNA (ncRNA) refers to all RNAs transcribed from the genome but lacking the ability to encode proteins. However, with the rapid development of ribosome profiling, sequencing technologies, and proteomics, a growing number of discoveries suggest that ncRNAs may possess coding potential. The peptides or proteins encoded by these ncRNAs have important biological significance in the physiological processes of cancer development. To date, ncRNAs with known functions include miRNAs, circRNAs, and lncRNAs. Peptides or proteins encoded by lncRNAs can regulate gene expression through alternative RNA splicing, thereby exerting cancer-related functions. For example, HOXB-AS3 is a lncRNA that encodes a conserved 53-aa small peptide that effectively inhibits metabolic processes in colon cancer. circ-SHPRH, circ-PINT, circ-AKT3, and circ-FBXW7 encode SHPRH-146aa, 87aa, 174aa, and 185aa peptides or proteins, respectively, which inhibit glioma cell proliferation and induce cell cycle arrest. The peptide encoded by circ-Gprc5a binds to Gprc5a, which is highly expressed in bladder tumors and bladder cancer stem cells. This peptide is required for the development and metastasis of bladder cancer and has the function of targeted elimination of bladder tumors and bladder cancer stem cells. In CaSki cervical cancer cells, specific disruption of circE7 helps to prevent the translation of the E7 oncoprotein.

[0003] The translation of these ncRNAs has demonstrated the existence of a hidden human proteome, opening up a new perspective for ncRNA research. Peptides encoded by ncRNAs have significant tumor-promoting or tumor-suppressing functions by interfering with cancer metastasis or metabolism. Due to their high tumor specificity, these peptides have great potential as therapeutic targets and are expected to become a new resource for the research of anti-tumor protein drugs. Furthermore, exploring the coding potential of ncRNA sequences will help discover effective tumor biomarkers, which is of great significance for early tumor detection, precision medicine, and prognosis prediction. However, to date, this research is still in its infancy. Due to the unclear characteristics of ncRNAs encoding small peptides or proteins, the relevant bioinformatics prediction and analysis methods are extremely lacking, which has seriously hindered the study of the coding potential of ncRNAs.

[0004] It can be seen that there is an urgent need for a method to predict the coding potential of non-coding RNA with high prediction accuracy and generalization performance. Summary of the Invention

[0005] In view of this, the embodiments of the present disclosure provide a method, device and medium for predicting the coding potential of non-coding RNA, which at least partially solve the problems of poor prediction accuracy and generalization ability in the existing technology.

[0006] In a first aspect, the present disclosure provides a method for predicting the coding potential of non-coding RNA, comprising:

[0007] Step 1: Based on the input non-coding RNA transcript sequence, a short sequence of a preset length is calculated to obtain a vocabulary. The CLRNA2V model is proposed to maximize the consistency between similar codons and minimize the consistency between different codons to generate a codon vector.

[0008] Step 2: By keeping the original sequence unchanged, deleting the first nucleotide, and deleting the first two nucleotides, and obtaining all possible codon combinations according to the non-overlapping trinucleotide algorithm, a non-coding RNA sequence representation matrix is ​​generated using the vocabulary;

[0009] Step 3: Using transfer learning, the non-coding RNA sequence representation matrix is ​​input into a bidirectional LSTM network and attention mechanism PBLA pre-trained by non-coding RNA sequences and messenger RNA sequences to deeply extract the potential feature information of the transcript sequence and obtain the transcript embedding vector of the non-coding RNA;

[0010] In step 4, the mutual information between the features to be sorted and the prediction results is calculated through the mutual information-based correlated redundant feature selection algorithm. The feature importance ranking is obtained by calculating the correlation and redundancy. The feature subset with the best performance is taken as the final feature, and the final feature is spliced ​​with the transcript embedding vector to predict the coding potential.

[0011] According to a specific implementation of the embodiment of the present disclosure, step 1 specifically includes:

[0012] All possible codon types are calculated through a sliding window of preset length. If a new codon is generated, it is added to the vocabulary. Otherwise, the sliding window continues to obtain all codon types. The CLRNA2V model uses contrastive learning to convert RNA sequences into vectors and calculate the codon vector V using the following formula kmer :

[0013] kmer=N1N2N3,N2N3N4,…,N n-2 N n-1 N n

[0014] V kmer =CLRNA2V(kmer)

[0015] Where ncRNA = N1, N2, ..., N n represents non-coding RNA transcripts, and n is the transcript length.

[0016] According to a specific implementation of the embodiment of the present disclosure, step 2 specifically includes:

[0017] By generating three different forms of transcript sequences of the same non-coding RNA sequence, the entire sequence is divided into multiple subsequences of a fixed length of m, and all possible non-overlapping trinucleotides of each subsequence are calculated. The representation matrices of the three different forms of transcripts are processed through a maximum pooling layer to obtain the final non-coding RNA sequence representation matrix.

[0018] According to a specific implementation of the embodiment of the present disclosure, the non-overlapping trinucleotides generated by three different forms of transcript sequences are represented as

[0019] NLOT1=N1N2N3,N4N5N6,…,N m-2 N m-1 N m

[0020] NLOT2=N2N3N4,N5N6N7,…,N m-3 N m-2 N m-1

[0021] NLOT3=N3N4N5,N6N7N8,…,N m-4 N m-3 N m-2 .

[0022] According to a specific implementation of the embodiment of the present disclosure, step 3 specifically includes:

[0023] We used two types of RNA transcripts, non-coding RNA sequences and messenger RNA sequences, to train a BiLSTM based on the attention mechanism. We then migrated the pre-trained model, PBLA, to the task of predicting the coding potential of non-coding RNA. The non-coding RNA sequence representation matrix was fed into PBLA, and the potential sequence features of the sequence were extracted through forward LSTM and backward LSTM. The specific expression is as follows:

[0024]

[0025]

[0026]

[0027] in Represent the forward LSTM and reverse LSTM in PBLA, The forward LSTM and reverse LSTM of PBLA respectively represent the hidden layer vector representation. Represents the vector representation of the i-th subsequence, score t Represents the attention score obtained by the attention mechanism, a t Indicates the final importance.

[0028] According to a specific implementation of the embodiment of the present disclosure, the calculation formula of the correlation and redundancy is:

[0029]

[0030] Among them, F is the feature set, f i ,f j represents the feature in the feature set, c is the predicted category, and I(;) is the mutual information.

[0031] In a second aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0032] at least one processor; and,

[0033] a memory communicatively connected to the at least one processor; wherein,

[0034] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the non-coding RNA coding potential prediction method in the aforementioned first aspect or any implementation of the first aspect.

[0035] In a third aspect, an embodiment of the present disclosure further provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the non-coding RNA coding potential prediction method in the aforementioned first aspect or any implementation of the first aspect.

[0036] In a fourth aspect, an embodiment of the present disclosure further provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the non-coding RNA coding potential prediction method in the aforementioned first aspect or any implementation of the first aspect.

[0037] The non-coding RNA coding potential prediction scheme in the embodiment of the present disclosure includes: step 1, calculating a short sequence of a preset length based on the input non-coding RNA transcript sequence to obtain a vocabulary, proposing a CLRNA2V model to maximize the consistency between similar codons and minimize the consistency between different codons, and generating a codon vector; step 2, by keeping the original sequence unchanged, deleting the first nucleotide and deleting the first two nucleotides, and obtaining all possible codon combinations according to the non-overlapping trinucleotide algorithm, using the vocabulary to generate a non-coding RNA sequence representation matrix; step 3, using transfer learning to input the non-coding RNA sequence representation matrix into a bidirectional LSTM network and attention mechanism PBLA pre-trained by non-coding RNA sequences and messenger RNA sequences to deeply extract potential feature information of the transcript sequence, and obtain a transcript embedding vector of the non-coding RNA; step 4, calculating the mutual information between the features to be sorted and the prediction results through a correlated redundant feature selection algorithm based on mutual information, obtaining the feature importance ranking by calculating the correlation and redundancy, taking the feature subset with the best performance as the final feature, and splicing the final feature with the transcript embedding vector to predict the coding potential.

[0038] The beneficial effects of the embodiments of the present disclosure are as follows: through the scheme of the present disclosure, the maximum pooling layer screens the most likely codon combination, and uses Bi LSTM to learn the context information of each nucleotide, and predicts the ncRNA coding potential by screening features with maximum correlation and minimum redundancy and splicing them with sequence representation vectors, effectively solving the problems of complex and blind biological experimental processes and high experimental costs, and improving prediction accuracy and generalization performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0040] Figure 1 A schematic diagram of a process for predicting the coding potential of non-coding RNA provided in an embodiment of the present disclosure;

[0041] Figure 2 A schematic diagram of a specific implementation process of a method for predicting the coding potential of non-coding RNA provided in an embodiment of the present disclosure;

[0042] Figure 3 A schematic diagram of the AUC performance of generating different feature subsets according to the sorting of the mutual information-based correlated redundant feature selection algorithm provided in an embodiment of the present disclosure;

[0043] Figure 4A schematic diagram of the performance of predicting ncRNA coding potential under different model structures provided in the embodiments of the present disclosure;

[0044] Figure 5 A schematic diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0045] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0046] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0047] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement an apparatus and / or practice a method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this apparatus and / or practice this method.

[0048] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The illustrations only show components related to the present disclosure and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0049] Additionally, in the following description, specific details are provided to provide a thorough understanding of the examples. However, one skilled in the art will appreciate that the aspects described can be practiced without these specific details.

[0050] The present disclosure provides a method for predicting the coding potential of non-coding RNA, which can be applied to gene analysis processes in biomedical scenarios.

[0051] See also Figure 1 , is a flow chart of a method for predicting the coding potential of non-coding RNA provided by the embodiment of the present disclosure. Figure 1 and Figure 2 As shown, the method mainly includes the following steps:

[0052] Step 1: Based on the input non-coding RNA transcript sequence, a short sequence of a preset length is calculated to obtain a vocabulary. The CLRNA2V model is proposed to maximize the consistency between similar codons and minimize the consistency between different codons to generate a codon vector.

[0053] Furthermore, the step 1 specifically includes:

[0054] All possible codon types are calculated through a sliding window of preset length. If a new codon is generated, it is added to the vocabulary. Otherwise, the sliding window continues to obtain all codon types. The CLRNA2V model uses contrastive learning to convert RNA sequences into vectors and calculate the codon vector V using the following formula kmer :

[0055] kmer=N1N2N3,N2N3N4,…,N n-2 N n-1 N n

[0056] V kmer =CLRNA2V(kmer)

[0057] Where ncRNA = N1, N2, ..., N n represents non-coding RNA transcripts, and n is the transcript length.

[0058] In specific implementation, the dataset of the embodiment of the present invention can be composed of two main databases: SPENCER and LNCipedia5. SPENCER has verified 15 cancer types through an MS-based proteomics analysis pipeline, containing 29,526 small peptides encoded by ncRNAs. These small peptides are translated from 6,842 ncRNA sequences collected from RNAcentral, including 6,803 lncRNAs, 34 misc-RNAs, 2 rRNAs, 2 sRNAs and 1 snRNA. Among them, 22,060 small peptides have been experimentally verified in other studies. This invention uses these 6,842 ncRNA sequences as positive samples. A subset of LNCipedia5 is called the high-confidence set, which includes 107,039 transcripts that lack coding potential based on any indicator assessment. Therefore, this invention uses the high-confidence set as negative samples. In order to avoid dataset imbalance, some transcripts are randomly selected from the high-confidence dataset as the final negative samples based on the number of SPENCER databases with less data.

[0059] Based on 6842 ncRNA transcript sequences with peptide coding potential and 6842 ncRNA transcript sequences that were identified as not having coding potential by any evaluation method, k-mers of length 3 were calculated, the same k-mers were merged to obtain a vocabulary and codon vectors were generated using the CLRNA2V model; if the non-coding RNA transcript is represented as ncRNA = N1, N2, ..., N n , where n is the transcript length, CLRNA2V is a model that converts RNA sequences into vectors using contrastive learning, then the codon vector V kmer It can be calculated by the following formula:

[0060] kmer=N1N2N3,N2N3N4,…,N n-2 N n-1 N n (1)

[0061] V kmer =CLRNA2V(kmer) (2)

[0062] Step 2: By keeping the original sequence unchanged, deleting the first nucleotide, and deleting the first two nucleotides, and obtaining all possible codon combinations according to the non-overlapping trinucleotide algorithm, a non-coding RNA sequence representation matrix is ​​generated using the vocabulary;

[0063] Based on the above embodiment, step 2 specifically includes:

[0064] By generating three different forms of transcript sequences of the same non-coding RNA sequence, the entire sequence is divided into multiple subsequences of a fixed length of m, and all possible non-overlapping trinucleotides of each subsequence are calculated. The representation matrices of the three different forms of transcripts are processed through a maximum pooling layer to obtain the final non-coding RNA sequence representation matrix.

[0065] Furthermore, the non-overlapping trinucleotides generated by the three different forms of transcript sequences are represented as

[0066] NLOT1=N1N2N3,N4N5N6,…,N m-2 N m-1 N m

[0067] NLOT2=N2N3N4,N5N6N7,…,N m-3 N m-2 N m-1

[0068] NLOT3=N3N4N5,N6N7N8,…,N m-4 N m-3 N m-2 .

[0069] In specific implementation, three different transcript sequences of the same ncRNA sequence can be generated by keeping the original sequence unchanged, deleting the first nucleotide, and deleting the second nucleotide. The entire sequence is then divided into 64 subsequences of a fixed length of m. All possible non-overlapping trinucleotides (NLOTs) of each subsequence are calculated, and the final sequence representation matrix is ​​obtained through the maximum pooling layer. The obtained representation matrix fully considers all possible codon combinations and generates a sequence representation that is closest to the actual sequence codon division method. The NLOT generated by the three different transcript sequences can be expressed as:

[0070] NLOT1=N1N2N3,N4N5N6,…,N m-2 N m-1 N m (3)

[0071] NLOT2=N2N3N4,N5N6N7,…,N m-3 N m-2 N m-1 (4)

[0072] NLOT3=N3N4N5,N6N7N8,…,N m-4 N m-3 N m-2 (5)

[0073] Among them, NLOT2 represents the codon combination generated by the sequence after deleting the first nucleotide, and NLOT3 represents the codon combination generated by the sequence after deleting the first and second nucleotides. All codons are represented by the vocabulary generated in step 1 and combined and spliced. The three forms of sequences are converted into corresponding representation matrices, and then the maximum pooling layer is used to obtain the representation matrix of a complete sequence.

[0074] Step 3: Using transfer learning, the non-coding RNA sequence representation matrix is ​​input into a bidirectional LSTM network and attention mechanism PBLA pre-trained by non-coding RNA sequences and messenger RNA sequences to deeply extract the potential feature information of the transcript sequence and obtain the transcript embedding vector of the non-coding RNA;

[0075] Based on the above embodiment, step 3 specifically includes:

[0076] We used two types of RNA transcripts, non-coding RNA sequences and messenger RNA sequences, to train a BiLSTM based on the attention mechanism. We then migrated the pre-trained model, PBLA, to the task of predicting the coding potential of non-coding RNA. The non-coding RNA sequence representation matrix was fed into PBLA, and the potential sequence features of the sequence were extracted through forward LSTM and backward LSTM. The specific expression is as follows:

[0077]

[0078]

[0079]

[0080] in Represent the forward LSTM and reverse LSTM in PBLA, The forward LSTM and reverse LSTM of PBLA respectively represent the hidden layer vector representation. Represents the vector representation of the i-th subsequence, score t Represents the attention score obtained by the attention mechanism, a t Indicates the final importance.

[0081] In the specific implementation, the forget gate, input gate, and output gate are introduced in the ncRNA and mRNA classification tasks, and the self-connected hidden unit is replaced with a memory unit. LSTM effectively solves the gradient vanishing or gradient exploding problems existing in RNN. t , input gate i t , unit status update value The working mechanisms are as follows:

[0082] f t =σ(Wsf S t +W h f h t-1 +b f ) (6)

[0083] i t =σ(W si S t +W h i h t-1 +b i ) (7)

[0084]

[0085] Where W represents the self-updated weight of the hidden layer, S represents the embedded representation of the subsequence, h represents the hidden state, b represents the bias vector, and t is the time; the state of the memory unit is determined by the forget gate, input gate, unit state update value and the state of the previous moment:

[0086]

[0087] The output gate outputs the final result of LSTM, the output gate o t The working mechanism is as follows:

[0088] o t =σ(W so S t +W h o h t-1 +b o ) (10)

[0089] And update the state of the hidden layer at the current moment to:

[0090]

[0091] Since RNA sequence transcripts are long, and considering that important sequence information may be forgotten by LSTM, this migration task will use bidirectional LSTM to extract the potential sequential features of the sequence through forward LSTM and backward LSTM. The specific expression is as follows:

[0092]

[0093]

[0094] in Represents forward LSTM and reverse LSTM, H f ,H b Respectively represent the hidden layer vector representation obtained by forward LSTM and reverse LSTM, Represents the vector representation of the i-th subsequence. In addition, since each subsequence contributes differently to the classification task, the invention introduces an attention mechanism to assign different weights to subsequences. The attention mechanism determines and formalizes the representation of the input sequence by approximating the relative contribution of each subsequence. The soft selection mechanism can automatically measure the importance of subsequences, making strongly correlated features easier to extract than weakly correlated features. score t Represents the attention score obtained by the attention mechanism, a t The final importance can be calculated by the following formula:

[0095]

[0096]

[0097] Where V T ,W h ,U o is the weight matrix, b is the bias vector, H f is a subsequence representation, O i To remove H f Finally, the trained PBLA network is applied to the ncRNA coding potential prediction task, and the sequence representation matrix obtained in step 2 is fed into PBLA for further training.

[0098] In step 4, the mutual information between the features to be sorted and the prediction results is calculated through the mutual information-based correlated redundant feature selection algorithm. The feature importance ranking is obtained by calculating the correlation and redundancy. The feature subset with the best performance is taken as the final feature, and the final feature is spliced ​​with the transcript embedding vector to predict the coding potential.

[0099] Furthermore, the calculation formula of the correlation and redundancy is:

[0100]

[0101] Among them, F is the feature set, f i ,f j represents the feature in the feature set, c is the predicted category, and I(;) is the mutual information.

[0102] In the specific implementation, considering that there may be some noise information in these representative features, which will affect the performance of the model, we use the mutual information-based correlation redundant feature selection algorithm to select appropriate features. This invention uses mutual information to calculate the maximum correlation between the features to be sorted and the predicted results and the minimum redundancy between the features. The mutual information I(X; Y) between a feature and the rest of the features is calculated as follows:

[0103]

[0104] The correlation C(F,c) between the feature set F and the prediction result c and the redundancy R(F) between the features are calculated by the following formulas:

[0105]

[0106]

[0107] The calculation process of the final feature score M(F,c) is as follows:

[0108]

[0109] The feature subset obtained by the mutual information-based correlated redundant feature selection algorithm is concatenated with the ncRNA sequence embedding representation obtained in step 3, and the result is fed into the linear layer for coding potential prediction.

[0110] The non-coding RNA coding potential prediction method provided in this embodiment screens the most likely codon combination through the maximum pooling layer, and uses BiLSTM to learn the context information of each nucleotide. It predicts the coding potential of ncRNA by screening features with maximum correlation and minimum redundancy and splicing them with sequence representation vectors. It effectively solves the problems of complex and blind biological experimental processes and high experimental costs, and improves prediction accuracy and generalization performance.

[0111] The present invention will be further described below through a specific embodiment. Figure 3 and Figure 4 As shown in Figure 1, the performance of the present invention is compared with that of the classic ncRNA and mRNA classification models to illustrate the effectiveness of the present invention. First, the data set is divided according to an 8:1:1 partition ratio, and the data set of the present invention is used to train the remaining comparison models. The results are shown in Table 1. As can be seen from Table 1, all performances of the present invention are better than those of the other models. The area under the ROC curve (AUC) is 2.3 percentage points higher than that of the currently best classification model RNAsamba. This shows that the method of encoding ncRNA sequences and the network for extracting features of the present invention are significantly effective. By integrating multimodal features, richer embedded information is obtained, which improves the prediction effect of the present invention.

[0112]

[0113]

[0114] Table 1

[0115] To further validate the effectiveness of the present invention, the model was trained using the dataset used by the ncRNA and mRNA classification models to verify the invention's performance in ncRNA and mRNA dataset classification tasks. Unlike the present invention, the positive samples of existing classifiers come from NCBI RefSeq, and the negative samples come from Ensembl. As can be seen in Table 2, the invention far exceeds the performance of other classification models in all evaluation indicators. Although PLEK has a higher SN value than the invention, the overall performance of the present invention is better than PLEK. Therefore, the present invention not only achieves excellent performance in predicting ncRNA coding potential, but also has outstanding results in ncRNA and mRNA classification.

[0116] ACC (%) PRE(%) SN (%) SP(%) F-score MCC AUC CPAT 94.33 94.30 94.58 94.07 0.944 0.887 0.984 CPC2 93.07 95.26 90.92 95.30 0.930 0.862 0.982 PLEK 96.73 98.11 95.42 98.10 0.967 0.935 0.993 CPPred 96.23 97.10 95.44 97.04 0.963 0.925 0.992 Ours 98.61 99.22 98.84 98.01 0.990 0.966 0.996

[0117] Table 2

[0118] See also Figure 5 The present disclosure also provides an electronic device 50, comprising at least one processor and a memory communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the method for predicting the coding potential of non-coding RNAs in the aforementioned method embodiment.

[0119] The embodiments of the present disclosure also provide a non-transitory computer-readable storage medium, which stores computer instructions for causing the computer to execute the non-coding RNA coding potential prediction method in the aforementioned method embodiment.

[0120] An embodiment of the present disclosure also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the non-coding RNA coding potential prediction method in the aforementioned method embodiment.

[0121] Reference below Figure 5 , which shows a schematic structural diagram of an electronic device 50 suitable for implementing the embodiments of the present disclosure. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0122] like Figure 5 As shown, the electronic device 50 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. Various programs and data required for the operation of the electronic device 50 are also stored in the RAM 503. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0123] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 50 to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows the electronic device 50 with various devices, it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0124] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0125] It should be noted that the computer-readable medium mentioned above in the present disclosure 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, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0126] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0127] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device can perform the relevant steps of the method embodiment.

[0128] Alternatively, the computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device can perform the relevant steps of the above method embodiment.

[0129] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0130] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0131] The units involved in the embodiments described in this disclosure may be implemented by software or hardware.

[0132] It should be understood that various parts of the present disclosure can be implemented in hardware, software, firmware, or a combination thereof.

[0133] The above description is merely a specific embodiment of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A method for predicting the coding potential of non-coding RNA, characterized in that: include: Step 1: Based on the input non-coding RNA transcript sequence, a short sequence of a preset length is calculated to obtain a vocabulary. The CLRNA2V model is proposed to maximize the consistency between similar codons and minimize the consistency between different codons to generate a codon vector. Step 2: By keeping the original sequence unchanged, deleting the first nucleotide, and deleting the first two nucleotides, and obtaining all possible codon combinations according to the non-overlapping trinucleotide algorithm, a non-coding RNA sequence representation matrix is ​​generated using the vocabulary; Step 3: Using transfer learning, the non-coding RNA sequence representation matrix is ​​input into a bidirectional LSTM network and attention mechanism PBLA pre-trained by non-coding RNA sequences and messenger RNA sequences to deeply extract the potential feature information of the transcript sequence and obtain the transcript embedding vector of the non-coding RNA; In step 4, the mutual information between the features to be sorted and the prediction results is calculated through the mutual information-based correlated redundant feature selection algorithm. The feature importance ranking is obtained by calculating the correlation and redundancy. The feature subset with the best performance is taken as the final feature, and the final feature is spliced ​​with the transcript embedding vector to predict the coding potential.

2. The method according to claim 1, characterized in that , the step 1 specifically includes: All possible codon types are calculated through a sliding window of preset length. If a new codon is generated, it is added to the vocabulary. Otherwise, the sliding window continues to obtain all codon types. The CLRNA2V model uses contrastive learning to convert RNA sequences into vectors and calculate the codon vector V using the following formula kmer : <h2 style=";text-align:left;direction:ltr">kmer=N1N2N3,N2N3N4,…,N<h2 style=";text-align:left;direction:ltr"> n-2 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> n-1 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> n In kmer =CLRNA2V(Khmer) Where ncRNA = N1, N2, ..., N n represents non-coding RNA transcripts, and n is the transcript length.

3. The method according to claim 2, characterized in that , the step 2 specifically includes: By generating three different forms of transcript sequences of the same non-coding RNA sequence, the entire sequence is divided into multiple subsequences of a fixed length of m, and all possible non-overlapping trinucleotides of each subsequence are calculated. The representation matrices of the three different forms of transcripts are processed through a maximum pooling layer to obtain the final non-coding RNA sequence representation matrix.

4. The method according to claim 3, characterized in that , the non-overlapping trinucleotides generated by the three different forms of transcript sequences are represented as <h2 style=";text-align:left;direction:ltr">NLOT1=N1N2N3,N4N5N6,…,N<h2 style=";text-align:left;direction:ltr"> m-2 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m-1 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m <h2 style=";text-align:left;direction:ltr">NLOT2=N2N3N4,N5N6N7,…,N<h2 style=";text-align:left;direction:ltr"> m-3 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m-2 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m-1 <h2 style=";text-align:left;direction:ltr">NLOT3=N3N4N5,N6N7N8,…,N<h2 style=";text-align:left;direction:ltr"> m-4 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m-3 <h2 style=";text-align:left;direction:ltr"> N<h2 style=";text-align:left;direction:ltr"> m-2 <h2 style=";text-align:left;direction:ltr"> 。 5. The method according to claim 4, characterized in that , the step 3 specifically includes: We used two types of RNA transcripts, non-coding RNA sequences and messenger RNA sequences, to train a BiLSTM based on the attention mechanism. We then migrated the pre-trained model, PBLA, to the task of predicting the coding potential of non-coding RNA. The non-coding RNA sequence representation matrix was fed into PBLA, and the potential sequence features of the sequence were extracted through forward LSTM and backward LSTM. The specific expression is as follows: in Represent the forward LSTM and reverse LSTM in PBLA, The forward LSTM and reverse LSTM of PBLA respectively represent the hidden layer vector representation. Represents the vector representation of the i-th subsequence, score t Represents the attention score obtained by the attention mechanism, a t Indicates the final importance.

6. The method according to claim 5, characterized in that , the calculation formula of the correlation and redundancy is: Among them, F is the feature set, f i ,f j represents the feature in the feature set, c is the predicted category, and I(;) is the mutual information.

7. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the non-coding RNA coding potential prediction method according to any one of claims 1 to 6.

8. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, which are used to enable the computer to execute the non-coding RNA coding potential prediction method according to any one of claims 1 to 6.

Images