Gene prediction method and apparatus, computer device, and computer readable storage medium

Abstract

Provided are a gene prediction method and apparatus, a computer device, and a computer readable storage medium. The method comprises: acquiring a template gene sequence, and a genetic gene sequence and a free gene sequence which correspond to a subject under test (step S102); determining a target gene site in the genetic gene sequence and the free gene sequence (step S104); extracting feature data corresponding to the target gene site, wherein the feature data is used for representing attribute features of the target gene site, and the feature data comprises first input data and second input data (step S106); acquiring a target gene prediction model, wherein the target gene prediction model comprises a first network, a second network, and a third network, and an output of the first network and the second network is an input of the third network (step S108); and respectively inputting the first input data and the second input data into the first network and the second network to obtain a gene prediction result outputted by the third network and corresponding to said subject (step S110).

Description

Gene prediction methods, devices, computer equipment, and computer-readable storage media Technical Field

[0001] This application relates to the field of bioinformatics, and in particular to a gene prediction method, apparatus, computer equipment, and computer-readable storage medium. Background Technology

[0002] With the development of bioinformatics technology, the detection results of cell-free genes in peripheral plasma of genetic parents can effectively predict the gene sequences of genetic offspring, which has great research value for experimental analysis and practical application of biological genetic characteristics.

[0003] Traditional techniques typically rely on Bayesian probabilistic models to infer the genome of offspring, but these models have low predictive accuracy.

[0004] Summary of the Invention

[0005] This application provides a gene prediction method, apparatus, computer device, and computer-readable storage medium.

[0006] In a first aspect, this application provides a gene prediction method, executed by a computer device, comprising:

[0007] Obtain the template gene sequence and the corresponding genetic gene sequence and free gene sequence of the test object. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the test object.

[0008] Identify target gene loci in genetic sequences and free gene sequences;

[0009] Extract feature data corresponding to the target gene locus. The feature data is used to characterize the attribute features of the target gene locus. The feature data includes the first input data and the second input data.

[0010] Obtain the target gene prediction model, which includes a first network, a second network, and a third network. The outputs of the first and second networks are the inputs of the third network.

[0011] The first input data and the second input data are respectively input into the first network and the second network to obtain the gene prediction results corresponding to the test object output by the third network.

[0012] Secondly, this application also provides a method for generating a gene prediction model, executed by a computer device, comprising:

[0013] Obtain the genetic sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the biological sample. The standard gene sequence corresponds to the genetic sequence and the free gene sequence.

[0014] Identify target gene loci in genetic sequences and free gene sequences; extract feature data corresponding to the target gene loci, which are used to characterize the attribute features of the target gene loci. The feature data includes first input data and second input data;

[0015] The initial gene prediction model is trained based on the first input data and the second input data to obtain the target gene prediction model. The target gene prediction model includes a first network, a second network and a third network. The outputs of the first network and the second network are the inputs of the third network. The target gene prediction model is used to predict the gene results of the corresponding biological sample.

[0016] Thirdly, this application also provides a gene prediction device, comprising:

[0017] The acquisition module is used to acquire the template gene sequence and the genetic gene sequence and free gene sequence corresponding to the test object. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the test object.

[0018] The extraction module is used to determine the target gene loci in the genetic gene sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data is used to characterize the attribute features of the target gene loci, and the feature data includes the first input data and the second input data;

[0019] The prediction module is used to obtain a target gene prediction model, which includes a first network, a second network, and a third network. The outputs of the first network and the second network are the inputs of the third network. The module also inputs the first input data and the second input data into the first network and the second network, respectively, to obtain the gene prediction results corresponding to the test object output by the third network.

[0020] Fourthly, this application also provides a gene prediction model generation apparatus, comprising:

[0021] The acquisition module is used to acquire the genetic gene sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the biological sample. The standard gene sequence corresponds to the genetic gene sequence and the free gene sequence.

[0022] The extraction module is used to determine the target gene loci in the genetic gene sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data is used to characterize the attribute features of the target gene loci, and the feature data includes the first input data and the second input data;

[0023] The training module is used to train the corresponding initial gene prediction model based on feature data to obtain the corresponding target gene prediction model. The gene prediction model is used to predict the gene results of the corresponding biological sample.

[0024] A computer device includes a memory and a processor, the memory storing computer-readable instructions, the processor executing the computer-readable instructions to perform the steps described in any of the above methods.

[0025] A computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by a processor, perform the steps described in any of the preceding methods.

[0026] Details of one or more embodiments of this application are set forth in the following drawings and description. Other features, objects, and advantages of this application will become apparent from the specification, drawings, and claims. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 is a flowchart illustrating a gene prediction method in one embodiment;

[0029] Figure 2 is a flowchart illustrating the steps for determining the target gene site in one embodiment;

[0030] Figure 3 is a flowchart illustrating the process of extracting feature data corresponding to target gene loci in one embodiment;

[0031] Figure 4 is a schematic diagram of the process of filtering feature data in one embodiment;

[0032] Figure 5 is a flowchart illustrating the prediction of gene prediction sequences based on a target gene prediction model in one embodiment.

[0033] Figure 6 is a flowchart illustrating the process of determining the predicted gene sequence of a test subject in one embodiment;

[0034] Figure 7 is a flowchart illustrating the gene prediction model generation method in another embodiment;

[0035] Figure 8 is a flowchart illustrating the process of determining the target gene site in one embodiment;

[0036] Figure 9 is a flowchart illustrating the process of generating feature data corresponding to each target gene locus in one embodiment.

[0037] Figure 10 is a schematic diagram of the process of filtering feature data in one embodiment;

[0038] Figure 11 is a flowchart illustrating the process of training a gene prediction model based on feature data in one embodiment;

[0039] Figure 13 is a technical schematic diagram of a gene prediction method in a specific embodiment;

[0040] Figure 14 shows the neural network model used in the gene prediction method in a specific embodiment;

[0041] Figure 15 is a schematic diagram of the actual prediction application of the gene prediction method in a specific embodiment;

[0042] Figure 16 is a structural block diagram of a gene prediction device in one embodiment;

[0043] Figure 17 is a structural block diagram of a gene prediction model generation device in one embodiment;

[0044] Figure 18 is an internal structure diagram of a computer device in one embodiment;

[0045] Figure 19 is an internal structural diagram of a computer device in another embodiment. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0047] In one embodiment, as shown in Figure 1, a gene prediction method is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0048] Step S102: Obtain the template gene sequence and the corresponding genetic gene sequence and free gene sequence of the test object.

[0049] Among them, the free gene sequence represents the sequence of the free gene corresponding to the test subject in the corresponding genetic object sample (such as peripheral blood, plasma).

[0050] Specifically, the computer equipment can obtain the template gene sequence and the corresponding genetic gene sequence, free gene sequence and concentration of the test object from the database, or it can obtain the above data from third-party equipment, servers, etc.

[0051] Step S104: Determine the target gene loci in the genetic sequence and the free gene sequence.

[0052] The target gene site is a specific base pair.

[0053] Specifically, the computer equipment can compare / match the genetic gene sequence and the free gene sequence with the template gene sequence for each corresponding base pair. The base pairs that fail to be matched / matched in the genetic gene sequence and the free gene sequence are taken as the target gene sites. The template gene sequence is the gene sequence corresponding to the object to be tested, such as the gene sequence template corresponding to a certain animal, plant or microorganism. The genetic gene sequence is the gene sequence corresponding to the genetic parent of the object to be tested.

[0054] Step S106: Extract the feature data corresponding to the target gene loci.

[0055] Among them, the feature data is used to characterize the attribute features of the target gene locus, and the feature data includes the first input data and the second input data.

[0056] Specifically, the computer equipment extracts corresponding feature parameters based on the chromosome number and location information of the target gene locus, as well as the sequencing depth and sequencing quality of the gene sequence containing the target gene locus. These feature parameters can reflect the chromosome number and location information of the corresponding target gene locus, as well as the sequencing depth and sequencing quality of the gene sequence containing the target gene locus.

[0057] Step S108: Obtain the target gene prediction model.

[0058] The target gene prediction model includes a first network, a second network, and a third network. The outputs of the first and second networks are the inputs of the third network. The first, second, and third networks can be any neural network or deep learning model, such as a convolutional neural network or a fully connected neural network.

[0059] Step S110: Input the first input data and the second input data into the first network and the second network respectively to obtain the target gene locus of the gene prediction result corresponding to the test object output by the third network.

[0060] The gene prediction result can be either a genotype or a gene sequence. The gene prediction result can include at least two candidate gene sequences or genotypes. Each candidate gene prediction sequence or genotype includes a corresponding confidence level. The confidence level is used to characterize the probability that the corresponding candidate gene prediction sequence is a true gene sequence. The higher the confidence level, the greater the probability that the corresponding candidate gene prediction sequence is a true gene sequence, and vice versa.

[0061] The aforementioned gene prediction method involves obtaining a template gene sequence and the corresponding genetic and free gene sequences of the target object, determining the target gene loci in the genetic and free gene sequences, extracting feature data corresponding to the target gene loci (which characterizes the attribute features of the target gene loci), and including first and second input data. A target gene prediction model is then obtained, comprising a first network, a second network, and a third network, with the outputs of the first and second networks serving as the input to the third network. The first and second input data are then input into the first and second networks, respectively, to obtain the gene prediction result corresponding to the target object output by the third network. This method, based on the genetic and free gene sequences, can quickly and accurately identify / determine the target gene loci. Furthermore, according to the acquired target gene prediction model, the first and second input data are fed into the first and second networks of the target gene prediction model, respectively, to obtain the gene prediction result corresponding to the test object output by the third network. This completes the prediction of the gene sequence of the test object, combining the characteristic attributes of the sample data itself with a specific model, thus more accurately learning the base characteristics of each gene locus in the sample gene sequence and improving the prediction accuracy of the gene sequence.

[0062] In one embodiment, the genetic sequence corresponding to the test object is the genotype of the genetic object corresponding to the test object.

[0063] In one embodiment, determining the target gene locus in a genetic sequence or a free gene sequence includes:

[0064] The genetic gene sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic gene sequence and the free gene sequence, respectively.

[0065] Specifically, the computer equipment can compare / match the genetic gene sequence and the free gene sequence with the template gene sequence for each corresponding base pair. The base pairs that fail to be matched / matched in the genetic gene sequence and the free gene sequence are taken as the target gene sites. The template gene sequence is the gene sequence corresponding to the object to be tested, such as the gene sequence template corresponding to a certain animal, plant or microorganism. The genetic gene sequence is the gene sequence corresponding to the genetic parent of the object to be tested.

[0066] In this embodiment, target gene loci within these sequences are determined by matching genetic and free gene sequences with template gene sequences. This matching process allows for the precise identification of specific gene locations, providing a clear basis for further analysis or manipulation. This method effectively identifies and distinguishes the presence and location of specific genes, offering significant application value in fields such as gene research, disease diagnosis, and personalized medicine.

[0067] In one embodiment, as shown in Figure 2, the genetic sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence, including:

[0068] In step S202, the genetic sequence and the free gene sequence are matched with the template gene sequence respectively to obtain the matching results for each base pair.

[0069] Specifically, the computer equipment matches the genetic sequence of the subject to be tested with the template gene sequence for each base pair, thereby identifying the target base pairs in the genetic sequence that differ from the corresponding base pairs in the template gene sequence. Similarly, the free gene sequence is matched with the template gene sequence to obtain the target base pairs in the free gene sequence that differ from the corresponding base pairs in the template gene sequence.

[0070] Step S204: Based on the matching results, determine the target base pairs between the genetic sequence and the free gene sequence.

[0071] The target base pair is the base pair in the genetic sequence and the free gene sequence that does not match the corresponding base pair in the template gene sequence.

[0072] Step S206: The target base pair is used as the target gene site for the genetic sequence and the free gene sequence.

[0073] In this embodiment, the genetic sequence and the free gene sequence are matched with the template gene sequence respectively to obtain the matching results of the base pairs. Based on the matching results, each target base pair of the genetic sequence and the free gene sequence is determined. The target base pairs are used as the target gene sites of the genetic sequence and the free gene sequence. Thus, the target gene sites in the genetic sequence and the free gene sequence can be accurately and quickly determined by using the template gene sequence as the comparison standard, thereby improving the efficiency and accuracy of gene site identification.

[0074] In one embodiment, as shown in Figure 3, extracting feature data corresponding to the target gene locus includes:

[0075] Step S302: Obtain the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus.

[0076] Sequencing depth refers to the number of times a genomic region is sequenced (generally, the higher the sequencing depth, the higher the accuracy of the detection results, because higher coverage depth can reduce detection errors and the probability of missed detection); sequencing quality is used to characterize the sequencing accuracy of the sequence sample data of the corresponding gene, and location information indicates the specific location of the corresponding gene locus on the corresponding chromosome.

[0077] Step S304: Based on the sequencing depth, sequencing quality, chromosome number, and location information of the target gene locus, generate feature data corresponding to the target gene locus.

[0078] Specifically, the computer device extracts the corresponding feature parameters based on the sequencing depth, sequencing quality, chromosome number, and location information of the gene loci determined in the above steps. These feature parameters can be used to characterize the aforementioned feature attributes of the target gene loci. Specifically, the computer device can obtain the correspondence between the feature attributes (sequencing depth, sequencing quality, and chromosome number and location information) of gene loci in the database and the feature parameters. Then, based on this correspondence and the feature attributes of the target gene loci, the target feature parameters corresponding to the target gene loci are determined. Finally, feature filtering is performed on the target feature parameters to determine the feature data corresponding to the target gene loci.

[0079] In this embodiment, the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus are obtained respectively. Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, feature data corresponding to the target gene locus is generated. This enables the determination / extraction of the corresponding feature data based on the feature attributes of the target gene locus, removing a large amount of redundant data, retaining the feature data that can characterize the feature attributes of the target gene locus, simplifying the input data of the model, reducing the computational burden of the data, and improving the computational efficiency of the data.

[0080] In one embodiment, as shown in Figure 4, feature data corresponding to the target gene locus is generated based on the sequencing depth, sequencing quality, and chromosome number and location information of the target gene locus. This data includes:

[0081] Step S402: Generate initial feature parameters corresponding to the target gene locus based on the sequencing depth, sequencing quality, chromosome number, and location information of the target gene locus.

[0082] The initial feature parameters can characterize the feature attributes (sequencing depth, sequencing quality, and chromosome number and location information) corresponding to the target gene locus. The initial feature parameters can be determined by computer equipment based on the mapping / correspondence between the feature attributes of the gene locus and the feature parameters.

[0083] Step S404: Based on the significance analysis model, test each initial feature parameter in sequence to obtain the significance analysis results corresponding to each initial feature parameter.

[0084] The significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding gene prediction sequence. For example, a significance analysis model can be constructed using the linear logistic regression method.

[0085] Specifically, after the computer device constructs the significance analysis model, it sequentially adds each initial feature parameter to the significance analysis model. Each time an initial feature parameter is added, the corresponding significance analysis model is determined. Then, the prediction effect (e.g., prediction accuracy) of the current significance analysis model is verified. In subsequent steps, the feature parameters among the initial feature parameters that can cause a significant improvement in the prediction effect of the corresponding significance analysis model are determined as target feature parameters.

[0086] Step S406: Based on the significance analysis results corresponding to each initial feature parameter, the target feature parameters corresponding to the target gene loci are selected and the target feature parameters are determined as the feature data corresponding to the target gene loci.

[0087] One method of selection is to identify the target feature parameters as those that can significantly improve the prediction performance of the significance analysis model among the initial feature parameters.

[0088] Understandably, in this embodiment, each initial feature parameter is used in sequence to construct a significance analysis model. Then, based on the prediction effect of the significance analysis model corresponding to each added feature parameter, the significance of the feature parameter for the gene sequence prediction result is determined. For feature parameters whose significance exceeds a preset threshold, they have a high significance impact on the gene sequence prediction result, while other initial feature parameters have a low significance impact on the gene sequence prediction result and are considered redundant data and are removed.

[0089] In this embodiment, initial feature parameters corresponding to the target gene locus are generated based on the sequencing depth, sequencing quality, chromosome number, and location information. Based on a significance analysis model, each initial feature parameter is tested sequentially to obtain significance analysis results. Based on these results, target feature parameters corresponding to the target gene locus are selected and defined as the feature data corresponding to the target gene locus. This allows for the determination of initial feature parameters that have a significant impact on the prediction model based on the characteristic attributes of the target gene locus. This process eliminates redundant data, retains feature data that characterizes the feature attributes of the target gene locus, simplifies the model's input data, reduces the computational burden, and improves computational efficiency.

[0090] In one embodiment, the first input data characterizes the base sequence features of the corresponding genetic sequence and free gene sequence, and the second input data characterizes the sequencing quality of the corresponding genetic sequence and free gene sequence.

[0091] In one embodiment, as shown in Figure 5, the first input data and the second input data are respectively input into the first network and the second network to obtain the gene prediction results corresponding to the test object output by the third network, including:

[0092] Step S502: The output of the first network and the output of the second network are merged and input into the third network to output at least two candidate gene prediction results.

[0093] Each candidate gene prediction sequence includes a corresponding confidence level. The confidence level is used to characterize the probability that the corresponding candidate gene prediction sequence is the actual gene sequence. The higher the confidence level, the greater the probability that the corresponding candidate gene prediction result is the actual gene sequence, and vice versa.

[0094] Step S504: Based on the confidence level of each candidate gene prediction sequence, determine the corresponding target gene prediction result from each candidate gene prediction sequence.

[0095] Specifically, based on the confidence levels corresponding to the prediction results of each candidate gene determined in the above steps, the computer device uses the candidate gene prediction result corresponding to the highest confidence level as the corresponding target gene prediction sequence. Optionally, if the candidate gene prediction result corresponding to the highest confidence level violates Mendel's laws of inheritance (i.e., a gene mutation occurs, resulting in a new genotype) and cannot be matched in the free gene sequence, then the candidate gene prediction result corresponding to the highest confidence level is removed, and the target gene prediction result is re-determined from the remaining candidate gene prediction sequences.

[0096] Step S506: Use the predicted sequence of the target gene as the gene prediction result corresponding to the object to be tested.

[0097] In this embodiment, by merging the outputs of the first network and the second network and inputting them into the third network, at least two candidate gene prediction sequences are output. Based on the confidence levels corresponding to each candidate gene prediction sequence, the corresponding target gene prediction result is determined from each candidate gene prediction sequence. Then, the target gene prediction sequence is used as the gene prediction sequence corresponding to the test object. Thus, by introducing multiple output results and their corresponding confidence levels, erroneous prediction results that violate Mendel's laws of inheritance during the prediction process can be identified and eliminated, effectively improving the reliability and accuracy of gene prediction.

[0098] In one embodiment, as shown in Figure 6, obtaining the target gene prediction model includes:

[0099] Step S602: Obtain the concentration corresponding to the free gene sequence.

[0100] Concentration is used to characterize the content of the free gene sequence of the test subject in the plasma of the corresponding genetic object.

[0101] Specifically, the computer equipment can obtain the concentration corresponding to the free gene sequence from a database, or it can obtain the above data from third-party equipment, servers, etc.

[0102] Step S604: Based on the gene type and concentration of the target gene locus, determine the corresponding target gene prediction model.

[0103] The gene type of each gene locus refers to the gene type of the genetic sequence corresponding to each gene locus (i.e., the gene sequence corresponding to the genetic parent of the test subject). The target gene prediction model is used to output the gene sequence prediction result of the test subject based on the genetic sequence and free gene sequence of the test subject. The gene sequence prediction result includes at least two candidate gene prediction sequences and their corresponding confidence scores. The confidence scores are used to characterize the corresponding candidate gene prediction sequences, and the sum of the confidence scores corresponding to all predicted gene sequences is 1.

[0104] Specifically, the computer equipment determines the corresponding target gene prediction model from various gene prediction models based on the gene type of the target gene locus and the concentration of the free gene sequence. Then, the feature data is input into the target gene prediction model to obtain the model output. Based on the model output, the gene prediction sequence corresponding to the test object is determined. Each gene prediction model corresponds to a different gene type within a different free gene sequence concentration range. For example, the concentration of the free gene sequence from the fetal part of the test object is divided into 5 concentration ranges {FF<5%, 5%≤FF<10%, 10%≤FF<15%, 15%≤FF<20%, 20%≤FF}, and the gene types of the gene sequence are divided into ABAB, AAAB, and ABAA. Each gene type under each concentration range corresponds to a gene prediction model.

[0105] In one embodiment, as shown in Figure 7, a method for generating a gene prediction model is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0106] Step S702: Obtain the genetic sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample.

[0107] Among them, the free gene sequence represents the sequence of the free gene corresponding to the biological sample in the sample of the corresponding genetic object (such as peripheral blood or plasma), the standard gene sequence corresponds to the genetic gene sequence and the free gene sequence, and the genetic gene sequence is the gene sequence corresponding to the genetic parent of the object to be tested.

[0108] Step S704: Determine the target gene loci in the genetic sequence and the free gene sequence.

[0109] The target gene site is a specific base pair.

[0110] Specifically, the computer device can compare / match the genetic gene sequence and the free gene sequence with the template gene sequence for each corresponding base pair, and take the base pairs in the genetic gene sequence and the free gene sequence that fail to be matched as their target gene sites.

[0111] Step S706: Extract the feature data corresponding to the target gene loci.

[0112] Among them, feature data is used to characterize the attribute features of the target gene loci.

[0113] Specifically, the computer equipment extracts corresponding feature parameters based on the chromosome number and location information of the target gene locus, as well as the sequencing depth and sequencing quality of the gene sequence containing the target gene locus. These feature parameters can reflect the chromosome number and location information of the corresponding target gene locus, as well as the sequencing depth and sequencing quality of the gene sequence containing the target gene locus.

[0114] Step S708: Train the corresponding initial gene prediction model based on the first input data and the second input data to obtain the corresponding target gene prediction model.

[0115] The target gene prediction model includes a first network, a second network, and a third network. The outputs of the first and second networks are the inputs of the third network. The target gene prediction model is used to predict the gene results of the corresponding biological sample. The first, second, and third networks can adopt any neural network or deep learning model, such as convolutional neural networks, fully connected neural networks, etc.

[0116] The aforementioned gene prediction model generation method involves acquiring the genetic gene sequence, free gene sequence, standard gene sequence, and template gene sequence corresponding to the biological sample; determining the target gene loci in the genetic gene sequence and free gene sequence; extracting feature data corresponding to the target gene loci, which characterizes the attribute features of the target gene loci, including first input data and second input data; training the corresponding initial gene prediction model based on the first and second input data to obtain the corresponding target gene prediction model; the target gene prediction model includes a first network, a second network, and a third network, with the outputs of the first and second networks serving as the input to the third network; the target gene prediction model is used to predict the gene results of the corresponding biological sample, thereby enabling more targeted training of the corresponding model based on data with different attribute features, effectively improving the generalization ability of the model training; and performing feature mining on the different attributes of the feature data, thereby achieving full learning of the global and local features of the feature data, improving the prediction efficiency and accuracy of the model.

[0117] In one embodiment, the standard gene sequence is a high-depth sequencing sequence of the corresponding biological sample. The standard gene sequence includes at least two reference gene sequences, and each reference gene sequence includes a corresponding standard confidence level.

[0118] In one embodiment, determining the target gene locus in the genetic sequence or the free gene sequence includes:

[0119] The genetic sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence, respectively.

[0120] Specifically, the computer equipment matches the genetic sequence of the subject to be tested with the template gene sequence for each base pair, thereby identifying the target base pairs in the genetic sequence that differ from the corresponding base pairs in the template gene sequence. Similarly, the free gene sequence is matched with the template gene sequence to obtain the target base pairs in the free gene sequence that differ from the corresponding base pairs in the template gene sequence.

[0121] In this embodiment, the target gene sites are determined by matching the genetic gene sequence and the free gene sequence with the template gene sequence, respectively, to obtain the matching results for each base pair. Based on the matching results, each target base pair in the genetic gene sequence and the free gene sequence is determined, and each target base pair is used as the target gene site in the genetic gene sequence and the free gene sequence. Thus, the target gene site in the genetic gene sequence and the free gene sequence can be accurately and quickly determined by using the template gene sequence as the comparison standard, thereby improving the efficiency and accuracy of gene site identification.

[0122] In one embodiment, as shown in Figure 8, the genetic sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence, including:

[0123] Step S802: Based on the matching results, determine each target base pair between the genetic sequence and the free gene sequence.

[0124] The target base pair is the base pair in the genetic sequence and the free gene sequence that does not match the corresponding base pair in the template gene sequence.

[0125] Step S804: Each target base pair is used as the target gene site for the genetic sequence and the free gene sequence.

[0126] In this embodiment, the genetic sequence and the free gene sequence are matched with the template gene sequence respectively to obtain the matching results corresponding to each base pair. Based on the matching results, each target base pair in the genetic sequence and the free gene sequence is determined. Each target base pair is used as the target gene locus in the genetic sequence and the free gene sequence. Thus, the target gene locus in the genetic sequence and the free gene sequence can be accurately and quickly determined based on the template gene sequence as the comparison standard, thereby improving the efficiency and accuracy of gene locus identification.

[0127] In one embodiment, as shown in Figure 9, extracting feature data corresponding to the target gene locus includes:

[0128] In step S902, the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus are obtained respectively.

[0129] Sequencing depth refers to the number of times a genomic region is sequenced (generally, the higher the sequencing depth, the higher the accuracy of the detection results, because higher coverage depth can reduce detection errors and the probability of missed detection); sequencing quality is used to characterize the sequencing accuracy of the sequence sample data of the corresponding gene, and location information indicates the specific location of the corresponding gene locus on the corresponding chromosome.

[0130] Step S904: Based on the sequencing depth, sequencing quality, chromosome number, and location information of the target gene locus, generate feature data corresponding to the target gene locus.

[0131] Specifically, the computer device extracts the corresponding feature parameters based on the sequencing depth, sequencing quality, chromosome number, and location information of the gene loci determined in the above steps. These feature parameters can be used to characterize the aforementioned feature attributes of the target gene loci. Specifically, the computer device can obtain the correspondence between the feature attributes (sequencing depth, sequencing quality, and chromosome number and location information) of the gene loci in the database and the feature parameters. Then, based on this correspondence and the feature attributes of the target gene loci, the target feature parameters corresponding to the target gene loci are determined. Finally, feature filtering is performed on the target feature parameters to determine the feature data corresponding to the target gene loci.

[0132] In this embodiment, the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus are obtained respectively. Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, feature data corresponding to the target gene locus is generated. This enables the determination / extraction of the corresponding feature data based on the feature attributes of the target gene locus, removing a large amount of redundant data, retaining the feature data that can characterize the feature attributes of the target gene locus, simplifying the input data of the model, reducing the computational burden of the data, and improving the computational efficiency of the data.

[0133] In one embodiment, as shown in Figure 10, feature data corresponding to the target gene locus is generated based on the sequencing depth, sequencing quality, and chromosome number and location information of the target gene locus, including:

[0134] Step S1002: Generate initial feature parameters corresponding to the target gene locus based on the sequencing depth, sequencing quality, chromosome number, and location information of the target gene locus.

[0135] The initial feature parameters can characterize the feature attributes (sequencing depth, sequencing quality, and chromosome number and location information) corresponding to the target gene locus. The initial feature parameters can be determined by computer equipment based on the mapping / correspondence between the feature attributes of the gene locus and the feature parameters.

[0136] Step S1004: Based on the significance analysis model, test each initial feature parameter in sequence to obtain the significance analysis results corresponding to each initial feature parameter.

[0137] The significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding gene prediction sequence. For example, a significance analysis model can be constructed using the linear logistic regression method.

[0138] Specifically, after the computer device constructs the significance analysis model, it sequentially adds each initial feature parameter to the significance analysis model. Each time an initial feature parameter is added, the corresponding significance analysis model is determined. Then, the prediction effect (e.g., prediction accuracy) of the current significance analysis model is verified. In subsequent steps, the feature parameters among the initial feature parameters that can cause a significant improvement in the prediction effect of the corresponding significance analysis model are determined as target feature parameters.

[0139] Step S1006: Based on the significance analysis results corresponding to each initial feature parameter, the target feature parameters corresponding to the target gene loci are selected and the target feature parameters are determined as the feature data corresponding to the target gene loci.

[0140] One method of selection is to identify the target feature parameters as those that can significantly improve the prediction performance of the significance analysis model among the initial feature parameters.

[0141] Understandably, in this embodiment, each initial feature parameter is used in sequence to construct a significance analysis model. Then, based on the prediction effect of the significance analysis model corresponding to each added feature parameter, the significance of the feature parameter for the gene sequence prediction result is determined. For feature parameters whose significance exceeds a preset threshold, they have a high significance impact on the gene sequence prediction result, while other initial feature parameters have a low significance impact on the gene sequence prediction result and are considered redundant data and are removed.

[0142] In this embodiment, initial feature parameters corresponding to the target gene locus are generated based on the sequencing depth, sequencing quality, chromosome number, and location information. Based on a significance analysis model, each initial feature parameter is tested sequentially to obtain significance analysis results. Based on these results, target feature parameters corresponding to the target gene locus are selected and defined as the feature data corresponding to the target gene locus. This allows for the determination of initial feature parameters that have a significant impact on the prediction model based on the characteristic attributes of the target gene locus. This process eliminates redundant data, retains feature data that characterizes the feature attributes of the target gene locus, simplifies the model's input data, reduces the computational burden, and improves computational efficiency.

[0143] In one embodiment, as shown in Figure 11, the corresponding initial gene prediction model is trained based on feature data to obtain the corresponding target gene prediction model, including:

[0144] Step S1102: Based on the gene type and free gene sequence concentration of the target gene locus, the matching feature data in the feature data are divided into the same feature sub-data set; and

[0145] Specifically, based on the gene type and free gene sequence concentration of the target gene locus, matching feature data are grouped into the same feature sub-data set. This process involves analyzing the feature data, identifying data points that match a given gene type and free gene sequence concentration, and organizing these data points into a separate feature sub-data set. This method makes the analysis and processing of relevant feature data more focused and efficient, providing a precise data foundation for subsequent gene analysis.

[0146] Step S1104: Based on each feature sub-data set and its corresponding standard gene sequence, determine and train the corresponding initial gene prediction model to obtain the target gene prediction model corresponding to each feature sub-data set.

[0147] Each feature sub-dataset and its corresponding standard gene sequence is used to determine and train an initial gene prediction model. This process involves using data from each feature sub-dataset and the corresponding standard gene sequence as training input to build and optimize the prediction model. In this way, a specific target gene prediction model will be developed for each feature sub-dataset, capable of accurately predicting the expression and variation of specific gene types in that dataset. Thus, models customized for different feature datasets can improve the accuracy and efficiency of gene prediction.

[0148] In this embodiment, firstly, matching data from the feature data are assigned to corresponding feature sub-data sets based on the gene type and free gene sequence concentration of the target gene locus. Then, an initial gene prediction model is determined and trained for each feature sub-data set and its corresponding standard gene sequence, ultimately generating the corresponding target gene prediction model. This method enables accurate processing and prediction of data based on complex gene information, improving prediction accuracy and efficiency, especially when dealing with data with different gene types and concentrations, ensuring the specificity and adaptability of the prediction model.

[0149] In one embodiment, as shown in Figure 12, based on each feature sub-data set and its corresponding standard gene sequence, an initial gene prediction model is determined and trained to obtain the target gene prediction model corresponding to each feature sub-data set, including:

[0150] The initial gene prediction model includes a first network, a second network, and a third network. The first network, the second network, and the third network can be any type of neural network or deep learning network, such as a fully connected layer network or a convolutional layer network, without any specific restrictions.

[0151] Step S1202: For the same feature sub-data set, the corresponding feature sub-data set is divided according to the attribute features to obtain the first input sample and the second input sample.

[0152] The first training sample represents the base sequence characteristics of the corresponding genetic gene sequence and free gene sequence; the second training sample represents the sequencing quality of the corresponding genetic gene sequence and free gene sequence; the attribute characteristics include base sequence characteristics, sequencing quality characteristics, and whether the genetic gene sequence and free gene sequence are homozygous or heterozygous genes.

[0153] Step S1204: Input the first input sample into the first network to obtain the corresponding first network output.

[0154] Step S1206: Input the second input sample into the second network to obtain the corresponding second network output;

[0155] Step S1208: Combine the output of the first network with the output of the second network and input them into the third network to obtain the output of the third network;

[0156] Step S1210: Calculate the model loss based on the difference between the third network output and the corresponding standard gene sequence.

[0157] Step S1212: Adjust the parameter values ​​of the first network, the second network, and the third network based on the model loss until the model training conditions are met, then stop training the corresponding initial gene prediction model to obtain the corresponding target gene prediction model.

[0158] In this embodiment, for the same feature sub-data set, data is partitioned to obtain a first input sample and a second input sample. The first input sample is then input into a first network to obtain the corresponding first network output. The second input sample is then input into a second network to obtain the corresponding second network output. The first and second network outputs are then merged and input into a third network to obtain the third network output. The model loss is calculated based on the difference between the third network output and the corresponding standard gene sequence. The parameter values ​​of the first, second, and third networks are adjusted based on the model loss until the model training conditions are met, at which point the training of the corresponding initial gene prediction model is stopped, resulting in the corresponding target gene prediction model. This allows for more targeted training of corresponding models based on data with different attribute features, effectively improving the generalization ability of model training. Feature mining is performed on the different attributes of the feature data, enabling full learning of the global and local features of the feature data, thereby improving the prediction efficiency and accuracy of the model.

[0159] This application also provides an application scenario in which the above-mentioned gene prediction method is applied, as shown in Figure 13. This method is applied to the prediction of fetal genotype point mutations. Specifically, the gene prediction method is applied in this scenario as follows:

[0160] 1. Construction of gene prediction models

[0161] In this embodiment, high-depth sequencing of cell-free DNA (cf-DNA, corresponding to the aforementioned cell-free gene sequence) from the plasma cells of pregnant women in 10 families was collected in the early stage, as well as whole-genome sequencing data of pregnant women and their husbands (corresponding to the aforementioned genetic sequences), and the standard gene sequence of the fetal genome was obtained through umbilical cord blood.

[0162] Based on the corresponding template gene sequence, base pairs were compared with cf-DNA and the parent genome to obtain each gene locus. Then, 67 initial feature data points containing the sequencing depth, sequencing quality, chromosome number, and location of each gene locus were extracted. Since there are only four base types, a single base pair may have only 4*4=16 point mutations. Ignoring the order of mutations (e.g., AT and TA belong to the same allele), 10 allele base pair classification results were obtained, as shown in Table 1 below.

[0163] Table 1

[0164] Using a linear logistic regression model, significance analysis was performed on the 67 initial feature parameters determined in the above steps to determine the significance of each feature parameter in predicting the effect of gene sequence. Each initial feature parameter with a significance greater than a preset threshold was used as feature data. In this embodiment, significance analysis was performed on the 67 initial feature parameters to obtain 56 target feature parameters, which were then used as feature data corresponding to each gene locus.

[0165] Based on the attribute characteristics of the feature data, the feature data corresponding to the 56 target feature parameters are divided into a first group of sample data (30 target feature parameters) and a second group of sample data (26 target feature parameters). The first group of sample data includes 24 target feature parameters for characterizing base pair sequences, 3 target feature parameters for characterizing the gene type of the parent genome, and 2 target feature parameters for characterizing the data quality of the parent genome and cell-free gene sequences. The second group of sample data characterizes the control parameters related to the sequencing quality of the corresponding parent genome and cell-free gene sequences.

[0166] Based on the concentration of cell-free gene sequences and the genotypes of the parents' genomes, the first and second sets of sample data are further divided to obtain various feature sub-data sets, each of which includes the first and second set of sample data.

[0167] For the same feature sub-data set, the first set of sub-sample data from this feature sub-data set is input into the first network, the second set of sub-sample data is input into the second network, and the outputs of the first and second networks are merged and input into the third network to obtain the output of the third network. Then, based on the output of the third network and the standard gene sequence, a model loss is constructed. The target parameters of the first, second, and third networks are dynamically adjusted based on this model loss until the model loss meets the model training stopping condition, thus obtaining the target gene prediction model. In one embodiment, the first and third networks are both fully connected layer networks, and the second network is a convolutional layer network. Optionally, as shown in Figure 14, the types of the first, second, and third networks may include fully connected layer networks, convolutional layer networks, activation layer networks, etc.

[0168] In this embodiment, the target gene loci are quickly and accurately identified / determined based on the differences between the parent genome, the free gene sequence, and the corresponding template gene sequence. Furthermore, based on the gene type of the target gene loci and the concentration characteristics of the free gene sequence, the corresponding gene prediction model is accurately located, thereby completing the prediction of the gene sequence of the test object. This combines the characteristic attributes of the sample data itself with a specific model, enabling more accurate learning of the base characteristics of each gene locus in the sample gene sequence and improving the prediction accuracy of the gene sequence.

[0169] 2. Application of gene prediction models

[0170] The genomic sequences of the parents of the fetus to be tested and the cell-free gene sequence of the pregnant woman's peripheral blood are obtained to construct the input sample data.

[0171] The input sample data is then compared with the corresponding template gene sequence. All base pairs in the input sample data that do not match the comparison results are taken as target gene sites. Initial feature parameters corresponding to each target gene site are extracted, and then the initial feature parameters are screened using a linear logistic regression model to obtain the input feature data.

[0172] Based on the concentration characteristics of cell-free gene sequences in the peripheral blood of pregnant women and the genotypes of the parents' genome sequences, the corresponding target gene prediction model is determined. Then, the input feature data is input into the target gene prediction model, and each candidate gene prediction sequence is output. Each candidate gene prediction sequence corresponds to a confidence level (the sum of the confidence levels of each gene prediction sequence is 1). Finally, the candidate gene prediction sequence with the highest confidence level is taken as the gene prediction sequence of the fetus to be tested, as shown in Figure 15.

[0173] In this embodiment, feature data corresponding to target gene loci in parental genome sequences and cell-free gene sequences are extracted. Based on the concentration characteristics of cell-free gene sequences and the gene type of target gene loci, the feature data is divided into subsets. Then, a corresponding gene prediction model is trained based on each feature subset. This allows for more targeted training of models based on data with different attribute features, effectively improving the generalization ability of model training. Feature mining is performed on the different attributes of the feature data, thereby enabling full learning of the global and local features of the feature data, improving the prediction efficiency and accuracy of the model.

[0174] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0175] In one embodiment, as shown in FIG16, a gene prediction device is provided. This device can employ software modules, hardware modules, or a combination of both as part of a computer device. Specifically, the device includes: an acquisition module 1602, an extraction module 1604, and a prediction module 1606, wherein:

[0176] The acquisition module 1602 is used to acquire the template gene sequence and the genetic gene sequence and free gene sequence corresponding to the test object. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the test object.

[0177] The extraction module 1604 is used to determine the target gene loci in the genetic gene sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data is used to characterize the attribute features of the target gene loci, and the feature data includes the first input data and the second input data;

[0178] The prediction module 1606 is used to acquire a target gene prediction model, which includes a first network, a second network, and a third network. The outputs of the first network and the second network are the inputs of the third network. The module also inputs the first input data and the second input data into the first network and the second network, respectively, to obtain the gene prediction results corresponding to the test object output by the third network.

[0179] In one embodiment, the genetic sequence corresponding to the test object is the genotype of the genetic object corresponding to the test object.

[0180] In one embodiment, the extraction module 1604 is further configured to match the genetic gene sequence and the free gene sequence with the template gene sequence respectively, and determine the target gene loci in the genetic gene sequence and the free gene sequence.

[0181] In one embodiment, the extraction module 1604 is further configured to match the genetic sequence and the free gene sequence with the template gene sequence respectively to obtain the matching results corresponding to each base pair; based on the matching results, determine the target base pairs of the genetic sequence and the free gene sequence, wherein the target base pairs are the base pairs in the genetic sequence and the free gene sequence that do not match the base pairs at the corresponding positions in the template gene sequence; and use the target base pairs as the target gene sites of the genetic sequence and the free gene sequence.

[0182] In one embodiment, the extraction module 1604 is further configured to obtain the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus, wherein the sequencing quality is used to characterize the sequencing accuracy of the sequence sample data of the corresponding gene; and to generate feature data corresponding to the target gene locus based on the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus.

[0183] In one embodiment, the extraction module 1604 is further configured to generate initial feature parameters corresponding to the target gene locus based on the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus; test each initial feature parameter sequentially based on a significance analysis model to obtain significance analysis results corresponding to each initial feature parameter, wherein the significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding gene prediction sequence; and, based on the significance analysis results corresponding to each initial feature parameter, screen to obtain target feature parameters corresponding to the target gene locus, and determine the target feature parameters as feature data corresponding to the target gene locus.

[0184] In one embodiment, the first input data characterizes the base sequence features of the corresponding genetic sequence and free gene sequence, and the second input data characterizes the sequencing quality of the corresponding genetic sequence and free gene sequence.

[0185] In one embodiment, the prediction module 1606 is further configured to merge the output of the first network with the output of the second network, input the second network into the third network, and output at least two candidate gene prediction sequences, each candidate gene prediction sequence including a corresponding confidence level; determine the corresponding target gene prediction result from each candidate gene prediction sequence based on the confidence level corresponding to each candidate gene prediction sequence; and use the target gene prediction result as the gene prediction result corresponding to the test object.

[0186] In one embodiment, the prediction module 1606 is further configured to obtain the concentration corresponding to the free gene sequence, the concentration being used to characterize the content of the free gene sequence of the test object in the plasma of the corresponding genetic object; and to determine the corresponding target gene prediction model based on the gene type and concentration of the target gene locus.

[0187] The aforementioned gene prediction device obtains the template gene sequence and the corresponding genetic and free gene sequences of the test object, determines the target gene loci in the genetic and free gene sequences, extracts the feature data corresponding to the target gene loci, and uses the feature data to characterize the attribute features of the target gene loci. The feature data includes first input data and second input data. The device then obtains a target gene prediction model, which includes a first network, a second network, and a third network. The outputs of the first and second networks are the inputs of the third network. The first and second input data are respectively input into the first and second networks to obtain the gene prediction result corresponding to the test object output by the third network. This method, based on the genetic and free gene sequences, can quickly and accurately identify / determine the target gene loci. Furthermore, according to the acquired target gene prediction model, the first and second input data are fed into the first and second networks of the target gene prediction model, respectively, to obtain the gene prediction result corresponding to the test object output by the third network. This completes the prediction of the gene sequence of the test object, combining the characteristic attributes of the sample data itself with a specific model, thus more accurately learning the base characteristics of each gene locus in the sample gene sequence and improving the prediction accuracy of the gene sequence.

[0188] For specific limitations regarding the gene prediction device, please refer to the limitations of the gene prediction method above, which will not be repeated here. Each module in the aforementioned gene prediction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0189] In one embodiment, as shown in Figure 17, a gene prediction model generation device is provided. This device can be a software module, a hardware module, or a combination of both integrated into a computer device. Specifically, the device includes: an acquisition module 1702, an extraction module 1704, and a training module 1706, wherein:

[0190] The acquisition module 1702 is used to acquire the genetic gene sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the biological sample. The standard gene sequence corresponds to the genetic gene sequence and the free gene sequence.

[0191] Extraction module 1704 is used to determine the target gene loci in the genetic gene sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data is used to characterize the attribute features of the target gene loci, and the feature data includes the first input data and the second input data;

[0192] Training module 1706 is used to train the corresponding initial gene prediction model based on feature data to obtain the corresponding target gene prediction model. The gene prediction model is used to predict the gene results of the corresponding biological sample.

[0193] In one embodiment, the standard gene sequence is a high-depth sequencing sequence of the corresponding biological sample. The standard gene sequence includes at least two reference gene sequences, and each reference gene sequence includes a corresponding standard confidence level.

[0194] In one embodiment, 1704 is further configured to match the genetic sequence and the free gene sequence with the template gene sequence respectively to determine the target gene loci in the genetic sequence and the free gene sequence.

[0195] In one embodiment, the extraction module 1704 is further configured to match the genetic sequence and the free gene sequence with the template gene sequence respectively to obtain the matching results corresponding to each base pair; based on the matching results, determine each target base pair of the genetic sequence and the free gene sequence, wherein the target base pair is the base pair in the genetic sequence and the free gene sequence that does not match the corresponding base pair in the template gene sequence; and use each target base pair as the target gene site of the genetic sequence and the free gene sequence.

[0196] In one embodiment, the extraction module 1704 is further configured to obtain the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus, wherein the sequencing quality is used to characterize the sequencing accuracy of the sequence sample data of the corresponding gene; and to generate feature data corresponding to the target gene locus based on the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus.

[0197] In one embodiment, the extraction module 1704 is further configured to generate initial feature parameters corresponding to the target gene locus based on the sequencing depth, sequencing quality, and chromosome number and location information corresponding to the target gene locus; test each initial feature parameter sequentially based on a significance analysis model to obtain the significance analysis results corresponding to the initial feature parameters, wherein the significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding gene prediction sequence; and, based on the significance analysis results corresponding to each initial feature parameter, screen to obtain the target feature parameters corresponding to the target gene locus, and determine the target feature parameters as the feature data corresponding to the target gene locus.

[0198] In one embodiment, the extraction module 1604 is further configured to divide the matching feature data in the feature data into the same feature sub-data set based on the gene type and the concentration of the free gene sequence of the target gene locus; and to determine and train the corresponding initial gene prediction model based on each feature sub-data set and its corresponding standard gene sequence, so as to obtain the target gene prediction model corresponding to each feature sub-data set.

[0199] In one embodiment, the extraction module 1704 is further configured to partition the corresponding feature sub-data set according to attribute features for the same feature sub-data set, obtaining a first input sample and a second input sample. The first training sample represents the base sequence features of the corresponding genetic gene sequence and free gene sequence, and the second training sample represents the sequencing quality of the corresponding genetic gene sequence and free gene sequence. The first input sample is input into a first network to obtain the corresponding first network output. The second input sample is input into a second network to obtain the corresponding second network output. The first network output and the second network output are merged and input into a third network to obtain the third network output. The model loss is calculated based on the difference between the third network output and the corresponding standard gene sequence. The parameter values ​​of the first network, the second network, and the third network are adjusted based on the model loss until the model training conditions are met, at which point the training of the corresponding initial gene prediction model is stopped, resulting in the corresponding target gene prediction model. The initial gene prediction model includes the first network, the second network, and the third network.

[0200] The aforementioned gene prediction model generation device acquires the genetic gene sequence, free gene sequence, standard gene sequence, and template gene sequence corresponding to the biological sample; determines the target gene loci in the genetic gene sequence and free gene sequence; then extracts the feature data corresponding to the target gene loci, which is used to characterize the attribute features of the target gene loci. The feature data includes first input data and second input data; and trains the corresponding initial gene prediction model based on the first input data and second input data to obtain the corresponding target gene prediction model. The target gene prediction model includes a first network, a second network, and a third network. The outputs of the first network and the second network are the inputs of the third network. The target gene prediction model is used to predict the gene results of the corresponding biological sample. This allows for more targeted training of the corresponding model based on data with different attribute features, effectively improving the generalization ability of the model training. Feature mining is performed on the different attributes of the feature data, thereby enabling full learning of the global and local features of the feature data, improving the prediction efficiency and accuracy of the model.

[0201] Specific limitations regarding the gene prediction model generation device can be found in the limitations of the gene prediction model generation method described above, and will not be repeated here. Each module in the aforementioned gene prediction model generation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0202] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram is shown in Figure 18. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores feature data corresponding to gene loci. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a gene prediction method or a gene prediction model generation method.

[0203] In an exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in Figure 19. The computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a gene prediction method or a gene prediction model generation method. The display unit of the computer device is used to form a visually visible image and may be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0204] Those skilled in the art will understand that the structures shown in Figures 18 and 19 are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements.

[0205] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0206] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0207] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0208] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0209] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0210] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A gene prediction method, characterized in that, Performed by a computer device, the method includes: Obtain the template gene sequence and the genetic gene sequence and free gene sequence corresponding to the test object. The free gene sequence represents the sequence of the free gene in the genetic sample that corresponds to the test object. Determine the target gene loci in the genetic sequence and the free gene sequence; Extract feature data corresponding to the target gene locus, the feature data being used to characterize the attribute features of the target gene locus, the feature data including first input data and second input data; Obtain a target gene prediction model, wherein the target gene prediction model includes a first network, a second network, and a third network, and the outputs of the first network and the second network are the inputs of the third network; and The first input data and the second input data are respectively input into the first network and the second network to obtain the gene prediction result corresponding to the test object output by the third network.

2. The method according to claim 1, characterized in that, The genetic sequence corresponding to the test object is the genotype of the genetic object corresponding to the test object.

3. The method according to claim 1, characterized in that, Determining the target gene locus in the genetic sequence and the free gene sequence includes: The genetic sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence, respectively.

4. The method according to claim 3, characterized in that, The step of matching the genetic sequence and the free gene sequence with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence includes: The genetic sequence and the free gene sequence are respectively matched with the template gene sequence to obtain the matching results for each base pair; Based on the matching results, target base pairs are determined between the genetic sequence and the free gene sequence. These target base pairs are base pairs in the genetic sequence and the free gene sequence that do not match the corresponding base pairs in the template gene sequence. The target base pair is used as the target gene site for the genetic sequence and the free gene sequence.

5. The method according to claim 1, characterized in that, The extraction of feature data corresponding to the target gene locus includes: Obtain the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus; the sequencing quality is used to characterize the sequencing accuracy of the corresponding gene sequence sample data; and Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, feature data corresponding to the target gene locus is generated.

6. The method according to claim 5, characterized in that, The step of generating feature data corresponding to the target gene locus based on the sequencing depth, sequencing quality, chromosome number, and location information includes: Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, the initial feature parameters corresponding to the target gene locus are generated. Based on the significance analysis model, each initial feature parameter is tested sequentially to obtain the significance analysis results corresponding to each initial feature parameter. The significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding predicted gene sequence. Based on the significance analysis results corresponding to each initial feature parameter, target feature parameters corresponding to the target gene locus are selected and determined as the feature data corresponding to the target gene locus.

7. The method according to claim 1, characterized in that, The first input data characterizes the base sequence features of the corresponding genetic sequence and free gene sequence, and the second input data characterizes the sequencing quality of the corresponding genetic sequence and free gene sequence.

8. The method according to claim 1, characterized in that, The step of inputting the first input data and the second input data into the first network and the second network respectively, and obtaining the gene prediction result corresponding to the test object output by the third network, includes: The outputs of the first network and the second network are merged and input into the third network to output at least two candidate gene prediction sequences, each candidate gene prediction sequence including the corresponding confidence level; Based on the confidence level corresponding to each candidate gene prediction sequence, the corresponding target gene prediction result is determined from each candidate gene prediction sequence; and The target gene prediction result is used as the gene prediction result corresponding to the test object.

9. The method according to claim 1, characterized in that, The acquisition of the target gene prediction model includes: The concentration corresponding to the free gene sequence is obtained, and the concentration is used to characterize the content of the free gene sequence of the test subject in the plasma of the corresponding genetic subject. Based on the gene type and concentration of the target gene locus, a corresponding target gene prediction model is determined.

10. A method for generating a gene prediction model, characterized in that, Performed by a computer device, the method includes: Obtain the genetic sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the biological sample. The standard gene sequence corresponds to the genetic sequence and the free gene sequence. Identify the target gene loci in the genetic sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data being used to characterize the attribute features of the target gene loci, the feature data including first input data and second input data; The initial gene prediction model is trained based on the first input data and the second input data to obtain the corresponding target gene prediction model. The target gene prediction model includes a first network, a second network and a third network. The outputs of the first network and the second network are the inputs of the third network. The target gene prediction model is used to predict the gene results of the corresponding biological sample.

11. The method according to claim 10, characterized in that, The standard gene sequence is a high-depth sequencing sequence of the corresponding biological sample. The standard gene sequence includes at least two reference gene sequences, and each reference gene sequence includes a corresponding standard confidence level.

12. The method according to claim 10, characterized in that, Determining the target gene locus in the genetic sequence and the free gene sequence includes: The genetic sequence and the free gene sequence are matched with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence, respectively.

13. The method according to claim 12, characterized in that, The step of matching the genetic sequence and the free gene sequence with the template gene sequence to determine the target gene loci in the genetic sequence and the free gene sequence includes: The genetic sequence and the free gene sequence are respectively matched with the template gene sequence to obtain the matching results for each base pair; Based on the matching results, target base pairs between the genetic sequence and the free gene sequence are determined, wherein the target base pairs are the base pairs in the genetic sequence and the free gene sequence that do not match the corresponding base pairs in the template gene sequence; and Each target base pair is used as the target gene site for the genetic sequence and the free gene sequence.

14. The method according to claim 10, characterized in that, The extraction of feature data corresponding to the target gene locus includes: Obtain the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus; the sequencing quality is used to characterize the sequencing accuracy of the corresponding gene sequence sample data; and Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, feature data corresponding to the target gene locus is generated.

15. The method according to claim 14, characterized in that, The step of generating feature data corresponding to the target gene locus based on the sequencing depth, sequencing quality, chromosome number, and location information includes: Based on the sequencing depth, sequencing quality, chromosome number, and location information corresponding to the target gene locus, the initial feature parameters corresponding to the target gene locus are generated. Based on the significance analysis model, each initial feature parameter is tested sequentially to obtain the significance analysis results corresponding to the initial feature parameters. The significance analysis model is used to characterize the significance of the influence of each initial feature parameter on the corresponding predicted gene sequence. Based on the significance analysis results corresponding to each initial feature parameter, target feature parameters corresponding to the target gene locus are selected and determined as the feature data corresponding to the target gene locus.

16. The method according to claim 15, characterized in that, The step of training the corresponding initial gene prediction model based on the feature data to obtain the corresponding target gene prediction model includes: Based on the gene type of the target gene locus and the concentration of the free gene sequence, the matching feature data in the feature data are divided into the same feature sub-data set; and Based on each feature sub-data set and its corresponding standard gene sequence, the corresponding initial gene prediction model is determined and trained to obtain the target gene prediction model corresponding to each feature sub-data set.

17. The method according to claim 16, characterized in that, The process of determining and training initial gene prediction models based on each feature sub-data set and its corresponding standard gene sequences, to obtain target gene prediction models for each feature sub-data set, includes: The initial gene prediction model includes a first network, a second network, and a third network; For the same feature sub-data set, the corresponding feature sub-data set is divided according to the attribute features to obtain a first input sample and a second input sample. The first training sample represents the base sequence features of the corresponding genetic gene sequence and free gene sequence, and the second training sample represents the sequencing quality of the corresponding genetic gene sequence and free gene sequence. The first input sample is input into the first network to obtain the corresponding first network output; The second input sample is fed into the second network to obtain the corresponding second network output; The output of the first network and the output of the second network are combined and input into the third network to obtain the output of the third network; The model loss is calculated based on the difference between the output of the third network and the corresponding standard gene sequence; and The parameter values ​​of the first, second, and third networks are adjusted based on the model loss until the model training conditions are met, at which point the training of the corresponding initial gene prediction model is stopped, and the corresponding target gene prediction model is obtained.

18. A gene prediction device, characterized in that, The device includes: The acquisition module is used to acquire the template gene sequence and the genetic gene sequence and free gene sequence corresponding to the test object. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the test object. An extraction module is used to determine the target gene loci in the genetic sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data being used to characterize the attribute features of the target gene loci, the feature data including first input data and second input data; The prediction module is used to acquire a target gene prediction model, which includes a first network, a second network, and a third network, wherein the outputs of the first network and the second network are the inputs of the third network; and to input the first input data and the second input data into the first network and the second network respectively, so as to obtain the gene prediction result corresponding to the test object output by the third network.

19. A gene prediction model generation device, characterized in that, The device includes: The acquisition module is used to acquire the genetic sequence, free gene sequence, standard gene sequence and template gene sequence corresponding to the biological sample. The free gene sequence represents the sequence of the free gene in the corresponding genetic object sample that corresponds to the biological sample. The standard gene sequence corresponds to the genetic sequence and the free gene sequence. An extraction module is used to determine the target gene loci in the genetic sequence and the free gene sequence; extract the feature data corresponding to the target gene loci, the feature data being used to characterize the attribute features of the target gene loci, the feature data including first input data and second input data; The training module is used to train the corresponding initial gene prediction model based on the feature data to obtain the corresponding target gene prediction model, which is used to predict the gene results of the corresponding biological sample.

20. A computer device comprising a memory and a processor, the memory storing computer-readable instructions, characterized in that, When the processor executes the computer-readable instructions, it implements the steps of the method according to any one of claims 1 to 17.

21. A computer-readable storage medium storing computer-readable instructions thereon, characterized in that, When the computer-readable instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 17.

Images