A method for genetic data analysis based on large language model

Abstract

The present invention relates to the technical field of gene analysis bioinformatics, and specifically discloses a method for gene data analysis based on a large language model, the method comprising: analyzing the quality of gene data to be analyzed, judging whether the gene data to be analyzed is available and determining a gene sequence segmentation scheme for the available gene data, inputting the gene data after the gene sequence segmentation into the large language model, obtaining a gene data association feature data set and a gene data analysis result, determining a gene data analysis accuracy feature value, and judging whether the gene data analysis result meets expectations. The present invention solves the problem that the quality and integrity of traditional biological data pose a challenge to the accuracy of the analysis result, and the effective integration and analysis of multimodal data still need to be improved. The whole process optimizes the efficiency of data processing, reduces the occurrence of repeated calculations, misjudgments and invalid results, improves the analysis efficiency, and reduces the waste of resources.

Description

Technical Field

[0001] The present invention relates to the technical field of gene analysis bioinformatics, and in particular to a gene data analysis method based on a large language model. Background Art

[0002] With the development of sequencing technology, biological data has exploded. Bioinformatics provides methods to parse, store and analyze these big data, enabling scientists to extract useful information from complex biological data and helping people study the structure, function and evolution of the entire genome, as well as how they affect health and disease. Bioinformatics also plays an important role in drug discovery and design. Through bioinformatics methods, researchers can identify potential drug targets and design new drug molecules, which speeds up the process of drug discovery and development and makes it possible to customize treatment plans based on individual genomic information. By analyzing large amounts of biological data, bioinformatics helps to reveal the molecular mechanisms of disease and provide new strategies for disease prevention and treatment. Bioinformatics methods can be used to compare the genomes of different species to study their evolutionary relationships and mechanisms. In the agricultural field, bioinformatics can help understand the genomes of crops, thereby improving varieties and increasing crop yields and disease resistance.

[0003] At present, there are still some shortcomings in the research on genetic data analysis based on big language models. Specifically, the quality and integrity of traditional biological data pose challenges to the accuracy of analysis results. The effective integration and analysis of multimodal data still need to be improved to more comprehensively understand biological systems. Genome sequencing data often contain sequencing errors, missing data or noisy data. If the quality of these data is not high, the model may produce misleading results during training and prediction, affecting the accuracy of gene function prediction. The correlation between multimodal data is often complex and nonlinear. The current big language model may be insufficient in capturing these complex correlations. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method for analyzing genetic data based on a large language model, which can effectively solve the problems involved in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for analyzing genetic data based on a large language model, comprising the following steps: analyzing the quality of genetic data to be analyzed to determine whether the genetic data to be analyzed is available; analyzing the basic information of the available genetic data to determine the genetic data gene sequence segmentation scheme; based on the genetic data gene sequence segmentation scheme, inputting the genetic data after the gene sequence segmentation into the large language model to obtain a genetic data associated feature data set and a genetic data analysis result; based on the genetic data associated feature data set, determining a genetic data analysis accuracy feature value; based on the genetic data analysis accuracy feature value, determining a genetic data analysis accuracy score; based on the genetic data analysis accuracy score, determining whether the genetic data analysis result is qualified.

[0006] As a further method, it is determined whether the gene data to be analyzed is available. The specific analysis process is: obtaining a quality data set of the gene data to be analyzed, and based on the obtained quality data set of the gene data to be analyzed, comprehensively analyzing to obtain a quality assessment value of the gene data to be analyzed, and using the quality assessment value of the gene data to be analyzed as an analysis basis for determining whether the gene data to be analyzed is available; comparing the quality assessment value of the gene data to be analyzed with a quality reference threshold of the gene data to be analyzed stored in a database; if the quality assessment value of the gene data to be analyzed is higher than or equal to the quality reference threshold of the gene data to be analyzed, then the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is available; if the quality assessment value of the gene data to be analyzed is lower than the quality reference threshold of the gene data to be analyzed, then the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is unavailable, and the gene data to be analyzed is re-acquired and re-analyzed.

[0007] As a further method, the quality data set of the gene data to be analyzed specifically includes the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed.

[0008] As a further method, the quality assessment value of the gene data to be analyzed, the specific analysis process is:

[0009] ;

[0010] In the formula, is the quality assessment value of the gene data to be analyzed, is the sequencing error rate of the gene data to be analyzed, is the number of repeated sequences in the gene data to be analyzed, is the sequence depth of the gene data to be analyzed, is the compensation factor for the sequencing error rate of the gene data to be analyzed. is the compensation factor for the number of repeated sequences in the gene data to be analyzed. is the compensation factor of the set depth of the gene data sequence to be analyzed, and e is a natural constant.

[0011] As a further method, a genetic data gene sequence segmentation scheme is determined, and the specific analysis process is: obtaining a genetic data basic information data set, based on the obtained genetic data basic information data set, comprehensive analysis to obtain genetic data basic information characteristic values, the genetic data basic information characteristic values ​​are used as analysis basis for determining the genetic data gene sequence segmentation scheme; the genetic data basic information characteristic values ​​are stored as designated labels, the designated labels are compared with the genetic data gene sequence segmentation schemes corresponding to the designated labels stored in the database, and the genetic data gene sequence segmentation scheme corresponding to the designated labels is obtained.

[0012] As a further method, the basic information data set of gene data specifically includes the number of effective gene data reads, the percentage of gene data N, the GC content of gene data, and the sequence depth of gene data.

[0013] As a further method, based on the genetic data associated feature data set, the genetic data analysis accuracy characteristic value is determined. The specific analysis process is: based on the genetic data associated feature data set, a comprehensive analysis is performed to obtain the genetic data analysis accuracy characteristic value, and the genetic data analysis accuracy characteristic value is used as the analysis basis for determining the genetic data analysis accuracy score.

[0014] As a further method, the gene data is associated with a feature data set, specifically including the gene data sequence length, the gene data GC content, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments.

[0015] As a further method, based on the genetic data analysis accuracy characteristic value, the genetic data analysis accuracy score is determined. The specific analysis process is: the genetic data analysis accuracy characteristic value is compared with each genetic data analysis accuracy reference characteristic value stored in the database to obtain the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value, the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value is recorded as the genetic data analysis accuracy characteristic deviation value, the genetic data analysis accuracy characteristic deviation value is imported into the genetic data analysis accuracy similarity model to obtain the genetic data analysis accuracy similarity characteristic value; the genetic data analysis accuracy similarity characteristic value is stored as a designated label, the designated label is compared with the genetic data analysis accuracy score corresponding to each designated label stored in the database, and the genetic data analysis accuracy score corresponding to the designated label is obtained.

[0016] As a further method, whether the genetic data analysis result is qualified is judged, and the specific analysis process is: compare the genetic data analysis accuracy score with the genetic data analysis accuracy definition score stored in the database; if the genetic data analysis accuracy score is higher than or equal to the genetic data analysis accuracy definition score, then the genetic data analysis result corresponding to the genetic data analysis accuracy score is qualified; if the genetic data analysis accuracy score is lower than the genetic data analysis accuracy definition score, then the genetic data analysis result corresponding to the genetic data analysis accuracy score is unqualified, and the genetic data to be analyzed needs to be re-analyzed.

[0017] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0018] (1) The present invention provides a method for analyzing genetic data based on a large language model to evaluate the quality of the genetic data to be analyzed, thereby ensuring that only high-quality, defect-free genetic data enters the analysis process, avoiding misleading analysis results caused by low-quality data, ensuring the reliability of the data source, and determining the most appropriate genetic sequence segmentation scheme based on the basic information of the genetic data. It can ensure that the input large language model data segments are reasonably divided, and the segmented genetic data are input into the large language model to fully explore the potential correlation between the data. The accuracy feature value is analyzed to determine the accuracy score, and each set of analysis results can be comprehensively scored, thereby measuring the reliability of the analysis results.

[0019] (2) The present invention determines whether the genetic data to be analyzed is usable by analyzing the quality of the genetic data to be analyzed. Low-quality data may not only produce erroneous results, but also waste computing resources and time. Screening out unusable data in advance can avoid the model wasting time and computing power on invalid data, thereby improving the overall efficiency of the analysis process. Low-quality data can easily introduce bias and affect the fairness of statistical results and model predictions. It can also reduce bias and ensure that the analysis results are more objective. Data quality analysis can discover and deal with problems in the data at an early stage, thereby reducing error propagation in subsequent analysis and ensuring the reliability of the entire analysis process.

[0020] (3) The present invention determines the accuracy score of genetic data analysis based on the accuracy characteristic value of genetic data analysis, provides the accuracy characteristic value to quantify the reliability of the analysis result, and generates an accuracy score that can intuitively reflect the credibility of the analysis result. The accuracy score helps to identify potential inaccurate analysis results and discover problems in data processing or model application in advance, thereby avoiding wrong decisions caused by data errors or model misjudgment. By referring to the score, it can ensure that the most reliable genetic data results are used for subsequent analysis or experimental design, reducing uncertainty. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0022] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0024] Reference Figure 1 As shown, the present invention provides a method for analyzing genetic data based on a large language model, including: analyzing the quality of the genetic data to be analyzed and determining whether the genetic data to be analyzed is usable.

[0025] Specifically, it is determined whether the gene data to be analyzed is available. The specific analysis process is: obtaining a quality data set of the gene data to be analyzed, and based on the obtained quality data set of the gene data to be analyzed, comprehensively analyzing to obtain a quality assessment value of the gene data to be analyzed, and using the quality assessment value of the gene data to be analyzed as an analysis basis for determining whether the gene data to be analyzed is available; comparing the quality assessment value of the gene data to be analyzed with a quality reference threshold of the gene data to be analyzed stored in a database; if the quality assessment value of the gene data to be analyzed is higher than or equal to the quality reference threshold of the gene data to be analyzed, then the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is available; if the quality assessment value of the gene data to be analyzed is lower than the quality reference threshold of the gene data to be analyzed, then the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is unavailable, and the gene data to be analyzed is re-acquired and re-analyzed.

[0026] It needs to be explained that the above-mentioned quality assessment and reference value comparison of genetic data can ensure that only genetic data that meets the quality standards enters the subsequent analysis steps, avoids the participation of low-quality data, and improves the overall quality of the analysis data. Data with a quality assessment value lower than the reference threshold will be automatically marked as unavailable, thereby avoiding low-quality data from interfering with or misleading subsequent analysis results. It is compared with the quality reference threshold stored in the preset database to ensure that the analysis process is based on objective and quantitative standards, reducing interference from human judgment or subjective factors. If the data assessment value is lower than the reference value, the system will automatically prompt to re-acquire the data, avoiding the waste of resources for further analysis of low-quality or invalid data, thereby optimizing the use of computing resources and time. Low-quality genetic data may lead to erroneous biological conclusions or misleading results. Through the quality assessment mechanism, these low-quality data can be effectively prevented from entering the analysis process, reducing the deviation and risk of the analysis results from the source.

[0027] Furthermore, the quality data set of the gene data to be analyzed specifically includes the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed.

[0028] In a specific embodiment, the sequencing error rate refers to the ratio of the sequencing results to the actual DNA sequence that does not match the gene sequence during the gene sequencing process. Errors may occur in steps such as base recognition, insertion or deletion, resulting in differences from the actual gene sequence. A high error rate may affect the accuracy of subsequent analysis. Therefore, reducing the sequencing error rate is the key to ensuring data quality. It is obtained based on analysis of quality control tools (such as FastQC and MultiQC). The number of repeated sequences refers to the fact that during the sequencing process, due to PCR amplification or other technical steps, some DNA fragments are sequenced multiple times, resulting in multiple identical sequences appearing in the sequencing data. Repeated sequencing does not represent repetition in the biological sense, but technical redundancy, which may lead to data deviation. It is obtained using FastQC. Sequence depth (also known as coverage) refers to the number of times the DNA base at the same position is read in the sequencing data. The higher the depth, the more times the base at that position is sequenced, which usually means that the data is more accurate and has higher confidence. Low depth may cause some variants or features to be unable to be accurately identified. FastQC provides average coverage information during quality control.

[0029] It needs to be explained that the above-mentioned high-throughput sequencers will produce a certain error rate during sequencing, especially errors may occur in base recognition. Each single base position sequenced may cause a certain proportion of errors due to the limitations of the instrument. If a gene site is sequenced multiple times (that is, the sequence depth is high), the impact of a single sequencing error can be eliminated through subsequent comparison and analysis. Mutual verification of multiple read data can increase the accuracy of sequencing at that position. When the same sequence fragment appears multiple times due to technical problems in the gene sequencing process (such as repetitions in PCR amplification), these repeated sequences will artificially increase the sequence depth. This artificially high depth does not reflect the actual biological information and may mask real mutations or other biological characteristics. If certain sequences are excessively repeated during the sequencing process (such as due to deviations in PCR amplification), these repeated sequences may mask sequencing errors because a large number of identical read data mask other possible base changes. Although the sequencing depth appears to be high on the surface, the presence of repeated sequences may not really improve the reliability of the data, but may hide sequencing errors.

[0030] Specifically, the quality assessment value of the gene data to be analyzed, the specific analysis process is:

[0031] ;

[0032] In the formula, is the quality assessment value of the gene data to be analyzed, is the sequencing error rate of the gene data to be analyzed, is the number of repeated sequences in the gene data to be analyzed, is the sequence depth of the gene data to be analyzed, is the compensation factor for the sequencing error rate of the gene data to be analyzed. is the compensation factor for the number of repeated sequences in the gene data to be analyzed. is the compensation factor of the set depth of the gene data sequence to be analyzed, and e is a natural constant.

[0033] It needs to be explained that the above-mentioned quality assessment value of the gene data to be analyzed is calculated through the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed. The sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed are normalized. The sequencing error rate, the number of repeated sequences, and the sequence depth reflect the quality of the gene data from different angles, respectively. The quality of the gene data can be considered in an all-round way to ensure that the analysis results are more representative and accurate. A high sequencing error rate, a large number of repeated sequences, or insufficient sequence depth will directly affect the reliability of the data analysis. By combining these parameters into a quality assessment value, low-quality data sets can be quickly identified, thereby avoiding the use of these data for subsequent gene analysis. By calculating the sequencing error rate, the number of repeated sequences, and the sequence depth, an automated data quality control process can be realized. In large-scale gene sequencing projects, automated quality assessment can quickly detect problematic data sets, automatically process them, or mark them as low quality, thereby greatly improving analysis efficiency.

[0034] , , The compensation factors corresponding to the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed are preset in the database, respectively. The numerical value represents the influence of the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed on the quality assessment value of the gene data to be analyzed. When used, the compensation factors corresponding to the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed can be directly obtained from the database. The corresponding relationship can be a preset mapping relationship, in which the mapping relationship can be a one-to-one correspondence or a many-to-one relationship. For example, the number of repeated sequences of the gene data to be analyzed A mapping set is formed between the sequencing error rate and the compensation factor of the sequencing error rate of the gene data to be analyzed, and the real-time sequencing error rate of the gene data to be analyzed is input into the mapping set to obtain the compensation factor of the sequencing error rate of the gene data to be analyzed; for example, a mapping set is formed between the number of repeated sequences of the gene data to be analyzed and the compensation factor of the number of repeated sequences of the gene data to be analyzed, and the real-time number of repeated sequences of the gene data to be analyzed is input into the mapping set to obtain the compensation factor of the number of repeated sequences of the gene data to be analyzed; for example, a mapping set is formed between the sequence depth of the gene data to be analyzed and the compensation factor of the sequence depth of the gene data to be analyzed, and the real-time sequence depth of the gene data to be analyzed is input into the mapping set to obtain the compensation factor of the sequence depth of the gene data to be analyzed.

[0035] Analyze the available basic information of genetic data and determine the genetic data gene sequence segmentation plan.

[0036] Specifically, a genetic data gene sequence segmentation scheme is determined, and the specific analysis process is: obtaining a genetic data basic information data set, and based on the obtained genetic data basic information data set, comprehensively analyzing to obtain a genetic data basic information characteristic value, and using the genetic data basic information characteristic value as an analysis basis for determining the genetic data gene sequence segmentation scheme; storing the genetic data basic information characteristic value as a designated label, and comparing the designated label with the genetic data gene sequence segmentation schemes corresponding to each designated label stored in a database to obtain the genetic data gene sequence segmentation scheme corresponding to the designated label.

[0037] It should be explained that the above-mentioned acquisition of the basic information data set of genetic data and determination of the segmentation scheme based on the eigenvalues ​​can ensure that the segmentation scheme matches the actual characteristics of the data, avoids blindly selecting a fixed segmentation length or strategy, but is tailored to the actual situation of the data, which is more accurate. The segmentation scheme tailored to the data characteristics can effectively avoid information loss (important information loss due to too small fragments) or redundancy (increase in invalid information due to too large fragments), thereby improving the accuracy of the analysis results. Different genetic data sets have different characteristics. For example, there may be significant differences in genome sequence length, repetitive regions, functional intervals, etc. By comparing the eigenvalues ​​with the eigenvalues ​​stored in the database, the segmentation scheme that best suits the current data can be selected, ensuring that the analysis method is flexible enough to cope with different genetic data types. By analyzing the eigenvalues ​​of the basic information of genetic data and comparing them with the database, the optimal segmentation scheme can be automatically determined, reducing manual intervention and trial and error processes, which can greatly improve the efficiency of the data processing process, shorten the time required for analysis, reduce errors and errors, and enhance the reliability and accuracy of the analysis results.

[0038] Furthermore, the basic information data set of gene data specifically includes the number of effective gene data reads, the percentage of gene data N, the GC content of gene data, and the sequence depth of gene data.

[0039] In a specific embodiment, the effective read number refers to the number of high-quality data reads (reads) that can actually be used for subsequent analysis after quality control and redundancy removal steps during the sequencing process. The preliminary sequencing data may contain low-quality reads, repeated sequences or contaminated data. After cleaning and filtering, the remaining high-quality data is called the effective read number. Quality control tools (such as FastQC, MultiQC) and data cleaning tools (such as Trimmomatic, Cutadapt) are used to filter out low-quality reads, adapter contamination or short fragment data to finally obtain the effective read number. The percentage of N refers to the proportion of bases marked as "N" in the gene sequence to the entire sequence. The base "N" represents a base that cannot be determined during the sequencing process, that is, the sequencer failed to accurately read the base at this position. A higher "N" percentage usually reflects poor data quality, which means that the sequencing is not accurate enough or there is missing information. The percentage of N can be evaluated using tools such as FastQC. GC content refers to the percentage of G (guanine) and C (cytosine) base pairs in all bases in a gene sequence. GC content is an important indicator for evaluating gene data quality and genome characteristics. The GC content varies greatly among different species or different gene regions. The genome of a specific species will have a relatively stable GC content. Extreme deviations from expected GC content may indicate data quality issues. The GC content of sequencing data can be directly calculated through tools such as FastQC and MultiQC. Sequence depth (also called coverage) refers to the number of times a specific base position is read during the sequencing process. The higher the depth, the more times the base at that position is sequenced, which usually provides higher accuracy and confidence. For whole genome sequencing or deep sequencing of specific genes, sequence depth is a key indicator for evaluating data reliability and variant detection capabilities. Use tools such as SAMtools, Bedtools, FastQC, etc. to analyze the aligned data to calculate the average depth and local depth at different locations of the genome.

[0040] It should be explained that in the above sequencing process, some bases cannot be read accurately and will be marked as "N", indicating undetermined bases. If the percentage of N in the gene sequence is high, these reads containing a large number of uncertain bases are usually considered to be of low quality. The data cleaning step will remove these low-quality reads, thereby reducing the number of valid reads. Regions with extremely high or low GC content often affect the quality of sequencing. High GC content regions may make it difficult for the sequencer to accurately identify and generate low-quality read data, while low GC content regions may also affect the success rate of sequencing. Extreme GC content may cause more reads to be considered invalid, thereby reducing the number of valid reads. Sequence depth refers to the number of times a certain gene sequence position is sequenced. Therefore, the more valid reads, the higher the number of reads covering each gene position, and the sequence depth increases accordingly. If the number of valid reads is small, the sequence depth will be limited, affecting the accuracy and reliability of subsequent analysis. In some areas, extremely high or low GC content may cause the accuracy of the sequencing instrument to decrease and produce more uncertain bases (N). Especially in high GC content areas, it is difficult for the sequencer to amplify these areas, resulting in more N bases appearing and the percentage of N increasing.

[0041] It should be explained that the specific analysis process of the above basic information characteristic values ​​of gene data is as follows:

[0042] ;

[0043] In the formula, is the basic information characteristic value of gene data, is the effective number of gene data reads, For genetic data percentage, For genetic data content, is the sequence depth of the gene data to be analyzed, is the compensation factor for the set number of effective gene data reads, is the compensation factor of the percentage of the set gene data N, Gene data set Compensation factor for content, It is the compensation factor of the sequence depth of the gene data to be analyzed.

[0044] It should be explained that the above-mentioned characteristic values ​​of basic information of genetic data are calculated through the number of effective reads of genetic data, the percentage of genetic data N, the GC content of genetic data, and the sequence depth of genetic data. The number of effective reads of genetic data, the percentage of genetic data N, the GC content of genetic data, and the sequence depth of genetic data are normalized. Through the four dimensions of effective read number, percentage of N, GC content and sequence depth, the quality of genetic data can be comprehensively analyzed, reflecting the information at different levels of the data, thereby improving the accuracy and reliability of genetic analysis. Different genetic data may show different characteristics in terms of effective read number, percentage of N, GC content and sequence depth. By calculating these characteristic values, researchers can choose the segmentation scheme that best suits the current data in a targeted manner to improve the accuracy and efficiency of the analysis.

[0045] , , The compensation factors corresponding to the number of effective gene data reads, the percentage of gene data N, and the GC content of gene data respectively preset in the database represent the numerical values ​​of the influence of the number of effective gene data reads, the percentage of gene data N, and the GC content of gene data on the characteristic value of the basic information of gene data. When used, the compensation factors corresponding to the number of effective gene data reads, the percentage of gene data N, and the GC content of gene data can be directly obtained from the database. The corresponding relationship can be a preset mapping relationship, in which the mapping relationship can be one-to-one or many-to-one. For example, the number of effective gene data reads and the compensation factor of the number of effective gene data reads form a mapping set, and the real-time number of effective gene data reads is input into the mapping set to obtain the compensation factor of the number of effective gene data reads; for example, the percentage of gene data N and the compensation factor of the percentage of gene data N form a mapping set, and the real-time percentage of gene data N is input into the mapping set to obtain the compensation factor of the percentage of gene data N; for example, the GC content of gene data and the compensation factor of the GC content of gene data form a mapping set, and the real-time GC content of gene data is input into the mapping set to obtain the compensation factor of the GC content of gene data.

[0046] Based on the gene data gene sequence segmentation scheme, the gene data after gene sequence segmentation is input into the large language model to obtain the gene data association feature data set and gene data analysis results.

[0047] It needs to be explained that the above-mentioned large language model is a type of model based on deep learning, which usually uses natural language processing (NLP) technology and is trained on large-scale text or other data sets. It can understand, generate, translate and summarize text information. Relying on huge parameters and data, it performs well in capturing data patterns, contextual information and relationships. The most common examples include the GPT series of models. Genetic data is sequence data and has a complex intrinsic structure like natural language. The large language model can capture patterns in the sequence, such as the function of base combinations, key sites in protein structure, etc. By inputting gene sequence fragments, the model can explore potential associations between genes and generate feature data sets for analyzing gene functions, regulatory networks, disease associations, etc. Traditional analysis methods often rely on domain experts to manually set rules to process genetic data, while large language models can automatically learn features from data, greatly reduce interference from human settings, provide more consistent and reliable analysis results, enhance the flexibility of sequence segmentation, and improve computing efficiency and data processing capabilities.

[0048] Specifically, collect genetic data, which are stored in FASTA format and contain sequences of multiple genes, perform feature engineering on the collected genetic data, obtain the length of each gene sequence, calculate the content of cytosine (C) and guanine (G) in each sequence, calculate the average base length and average amino acid length of the segmented fragments, select Transformers or other logical frameworks, including input layer, Transformer layer and output layer, and define the output sequence length, GC content, average base length and average amino acid length of the output layer, and define a loss function, such as mean square error (MSE), to evaluate the difference between the predicted features and actual features output by the model.

[0049] The genetic data to be feature engineered is input through the input layer (the bases (A, T, C, G) of the gene sequence are encoded as numerical values. Usually simple mappings (for example: A=1, T=2, C=3, G=4) or more complex techniques such as one-hot encoding can be used). The Transformer layer includes multiple output heads connected to the output layer, and each head is responsible for outputting different features (sequence length, GC content, average base length, average amino acid length).

[0050] Based on the genetic data associated feature data set, the genetic data analysis accuracy feature value is determined.

[0051] Specifically, based on the genetic data associated feature data set, the genetic data analysis accuracy feature value is determined. The specific analysis process is: based on the genetic data associated feature data set, a comprehensive analysis is performed to obtain the genetic data analysis accuracy feature value, and the genetic data analysis accuracy feature value is used as the analysis basis for determining the genetic data analysis accuracy score.

[0052] It needs to be explained that the accuracy characteristic values ​​extracted from the above-mentioned feature data set associated with genetic data can effectively quantify the reliability of the analysis results. Compared with traditional subjective evaluation methods, quantitative characteristic values ​​can more objectively reflect the accuracy of the analysis and reduce human interference. By calculating the characteristic values, the quality of the analysis results can be evaluated based on clear data standards, thereby ensuring that the data analysis results are more reliable. The accuracy characteristic values ​​are directly derived from the associated features of the genetic data and are based on a data-driven analysis method. The transparency of the entire analysis process is improved, and researchers can clearly see the basis for each step, which helps to interpret the analysis results. By generating accuracy characteristic values ​​and determining scores based on these characteristic values, a standardized analysis framework can be created, and the analysis results of all genetic data sets can be evaluated using the same standards to ensure that comparisons between different data sets are more consistent.

[0053] Furthermore, the gene data associated feature data set specifically includes the gene data sequence length, the gene data GC content, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments.

[0054] In a specific embodiment, the gene data sequence length refers to the total number of base pairs or amino acids in a segment of gene data. For DNA or RNA sequences, the sequence length refers to the number of bases, and for protein sequences, it refers to the number of amino acids. The GC content refers to the percentage of G (guanine) and C (cytosine) base pairs in the gene sequence to the total number of base pairs. The GC content is an important indicator for measuring the characteristics of gene sequences. The GC content of different species or genomic regions may vary greatly. The average base length of the segmented sequence fragments refers to the average number of bases in each fragment generated after the gene data is segmented. The average amino acid length of the segmented sequence fragments refers to the average number of amino acids in each fragment after the protein sequence is segmented. Since proteins are composed of amino acids, the fragment length generated after segmentation is used to evaluate the fragment characteristics in the protein structure to help subsequent function prediction and structural analysis. The gene data sequence length, the gene data GC content, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments are obtained based on the output of the large language model.

[0055] It should be explained that the above GC content is the ratio of the number of G and C bases in the sequence to the total number of bases (i.e., sequence length). The sequence length affects the calculation of the GC content, but the GC content is a relative value. Even if the sequence length is different, as long as the ratio of the GC content remains unchanged, the value is not affected by the length. The DNA or RNA sequence in the genetic data forms a protein sequence through transcription and translation. The segmentation method of the genetic data affects the length of each protein sequence fragment after segmentation. The average base length will affect the amino acid length of the translated fragment, and the amino acid length is a key factor in the prediction of functional protein structure. Reasonable sequence segmentation can ensure that each fragment has sufficient information for protein function and structure analysis.

[0056] It should be explained that the above-mentioned gene data analysis accuracy characteristic value, the specific analysis process is as follows:

[0057] ;

[0058] In the formula, Analyze the accuracy eigenvalues ​​for genetic data, is the length of the gene data sequence, is the average base length of the segmented sequence fragments, is the average amino acid length of the segmented sequence fragments, is the GC content of gene data, is the compensation factor for the set gene data sequence length, is the compensation factor for the average base length of the segmented sequence fragments. is the compensation factor for the average amino acid length of the segmented sequence fragments. It is the compensation factor of the GC content of the set gene data.

[0059] It should be explained that the above-mentioned characteristic values ​​of genetic data analysis accuracy are calculated through the genetic data sequence length, genetic data GC content, average base length of segmented sequence fragments, and average amino acid length of segmented sequence fragments. The genetic data sequence length, genetic data GC content, average base length of segmented sequence fragments, and average amino acid length of segmented sequence fragments are normalized. Combining sequence length, GC content, base length, and amino acid length, the accuracy of genetic data can be comprehensively evaluated from multiple angles. Each dimension reflects different characteristics of genetic data, helping to ensure that the analysis results are more representative. Through the combination of multiple features, changes in data quality can be more sensitively captured. Abnormal sequence length, high or low GC content, and changes in the length of base and amino acid fragments may indicate abnormalities in genetic data in different regions. Changes in these characteristics can help identify potential deviations in analysis results. The characteristic value comprehensively considers different dimensions, which helps to reduce errors in data processing and analysis. The length of segmented sequence fragments can affect the accuracy of the analysis algorithm. By optimizing these parameters, more accurate analysis results can be ensured.

[0060] , , The compensation factors corresponding to the gene data sequence length, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments preset in the database respectively represent the numerical value of the influence of the gene data sequence length, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments on the gene data analysis accuracy characteristic value. When used, the compensation factors corresponding to the gene data sequence length, the average base length of the segmented sequence fragments, and the average amino acid length of the segmented sequence fragments can be directly obtained from the database. The corresponding relationship can be a preset mapping relationship, in which the mapping relationship can be one-to-one or many-to-one. For example, the gene data sequence length A mapping set is formed with the compensation factor of the gene data sequence length, and the real-time gene data sequence length is input into the mapping set to obtain the compensation factor of the gene data sequence length; for example, the average base length of the segmented sequence fragments and the compensation factor of the average base length of the segmented sequence fragments form a mapping set, and the real-time average base length of the segmented sequence fragments is input into the mapping set to obtain the compensation factor of the average base length of the segmented sequence fragments; for example, the average amino acid length of the segmented sequence fragments and the compensation factor of the average amino acid length of the segmented sequence fragments form a mapping set, and the real-time average amino acid length of the segmented sequence fragments is input into the mapping set to obtain the compensation factor of the average amino acid length of the segmented sequence fragments.

[0061] Based on the genetic data analysis accuracy feature value, a genetic data analysis accuracy score is determined.

[0062] Specifically, based on the genetic data analysis accuracy characteristic value, the genetic data analysis accuracy score is determined, and the specific analysis process is: the genetic data analysis accuracy characteristic value is compared with each genetic data analysis accuracy reference characteristic value stored in the database to obtain the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value, the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value is recorded as the genetic data analysis accuracy characteristic deviation value, the genetic data analysis accuracy characteristic deviation value is imported into the genetic data analysis accuracy similarity model to obtain the genetic data analysis accuracy similarity characteristic value; the genetic data analysis accuracy similarity characteristic value is stored as a designated label, the designated label is compared with the genetic data analysis accuracy score corresponding to each designated label stored in the database, and the genetic data analysis accuracy score corresponding to the designated label is obtained.

[0063] It should be explained that the above-mentioned comparison of the accuracy characteristic value of the genetic data analysis with the reference characteristic value in the database and the use of the absolute value of the minimum difference to calculate the accuracy deviation value can ensure that the analysis process relies on data rather than human subjective judgment. The data-based evaluation method reduces the interference of subjective factors and ensures the objectivity of the analysis. In the process of characteristic value comparison, finding the minimum absolute value of the difference and using it as the accuracy deviation value can ensure the accuracy of the evaluation. This method can reduce the impact of errors and improve the accuracy and reliability of the analysis results. The accuracy deviation value is imported into the similarity model to further calibrate the reliability and quality of the data, which can help ensure that the analysis results are within the standard range and ensure the rigor of the analysis process. By recording the deviation of the characteristic value and the similarity characteristic value, the entire analysis process can be traced. Researchers can compare and analyze historical data and current results to ensure the consistency and interpretability of the data. By importing the characteristic deviation value into the similarity model, automated analysis accuracy evaluation can be achieved. The automated process reduces the complexity of manual intervention and greatly improves the evaluation efficiency.

[0064] It should be explained that the above-mentioned genetic data analysis accuracy similarity model, the specific analysis process is:

[0065] ;

[0066] In the formula, Analyze the accuracy similarity feature value for genetic data, is the characteristic deviation value of genetic data analysis accuracy, and e is a natural constant.

[0067] It needs to be explained that the above-mentioned similarity characteristic value of genetic data analysis accuracy is calculated by the characteristic deviation value of genetic data analysis accuracy. The characteristic deviation value of genetic data analysis accuracy is quantified by comparing the difference between the analysis result of genetic data and the reference value, and can accurately reflect the degree of deviation of the analysis result. The similarity characteristic value calculated by the characteristic deviation value directly reflects the similarity between the genetic data analysis result and the reference value. A high similarity characteristic value indicates that the analysis result is very close to the expectation, which intuitively demonstrates the reliability and accuracy of the result and facilitates the rapid judgment of the quality of the result. By calculating the characteristic deviation value and deriving the similarity characteristic value, the interference of human judgment can be reduced, ensuring that the evaluation of the analysis result is based on objective data, making the evaluation of the analysis accuracy more consistent, and avoiding the evaluation differences caused by different standards of different analysts.

[0068] Based on the genetic data analysis accuracy score, determine whether the genetic data analysis results are qualified.

[0069] Specifically, to determine whether the genetic data analysis result is qualified, the specific analysis process is: compare the genetic data analysis accuracy score with the genetic data analysis accuracy definition score stored in the database; if the genetic data analysis accuracy score is higher than or equal to the genetic data analysis accuracy definition score, then the genetic data analysis result corresponding to the genetic data analysis accuracy score is qualified; if the genetic data analysis accuracy score is lower than the genetic data analysis accuracy definition score, then the genetic data analysis result corresponding to the genetic data analysis accuracy score is unqualified, and the genetic data to be analyzed needs to be re-analyzed.

[0070] It needs to be explained that the above-mentioned comparison of the accuracy score of genetic data analysis with the pre-set accuracy definition score can ensure that each analysis result meets the predetermined quality standards. Only results that meet the standards will be deemed qualified, ensuring the credibility and scientific nature of the data analysis. The automated comparison of the scoring system can more objectively evaluate the reliability of the analysis results, avoid the interference of subjective judgment, and ensure the rigor and scientific nature of the analysis process. The judgment of the analysis results can be completed by an automated system, avoiding the tedious process of manual inspection. Only by comparing the scores can it be automatically determined whether the results are qualified, reducing the time and effort of manual intervention and improving analysis efficiency. Through the preset accuracy definition score, all data analyses are uniformly evaluated under one standard, and the evaluation results of different data sets can be compared under the same framework to ensure the standardization and comparability of the results. It can also prevent data with large errors from entering the downstream analysis process, control data quality at the source, and avoid misleading conclusions and analyses due to data errors.

[0071] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for analyzing genetic data based on a large language model, characterized in that: The following steps are involved: Analyze the quality of the gene data to be analyzed and determine whether the gene data to be analyzed is usable; Analyze the available basic information of genetic data and determine the genetic data gene sequence segmentation plan; Based on the gene data gene sequence segmentation scheme, the gene data after gene sequence segmentation is input into the large language model to obtain the gene data association feature data set and gene data analysis results; Determining a characteristic value of the accuracy of the genetic data analysis based on a genetic data associated characteristic data set; Determining a genetic data analysis accuracy score based on the genetic data analysis accuracy characteristic value; Based on the accuracy score of the genetic data analysis, determine whether the genetic data analysis results are qualified; The gene data associated feature data set specifically includes gene data sequence length, gene data GC content, average base length of segmented sequence fragments, and average amino acid length of segmented sequence fragments.

2. The method for analyzing genetic data based on a large language model according to claim 1, characterized in that: The specific analysis process of judging whether the gene data to be analyzed is available is as follows: Acquire a quality data set of the gene data to be analyzed, and based on the acquired quality data set of the gene data to be analyzed, perform a comprehensive analysis to obtain a quality assessment value of the gene data to be analyzed, and use the quality assessment value of the gene data to be analyzed as an analysis basis for determining whether the gene data to be analyzed is usable; Comparing the quality assessment value of the gene data to be analyzed with the quality reference threshold of the gene data to be analyzed stored in the database; If the quality assessment value of the gene data to be analyzed is higher than or equal to the quality reference threshold of the gene data to be analyzed, the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is available; If the quality assessment value of the gene data to be analyzed is lower than the quality reference threshold of the gene data to be analyzed, the gene data to be analyzed corresponding to the quality assessment value of the gene data to be analyzed is unavailable, and the gene data to be analyzed is acquired again and analyzed again.

3. The method for analyzing genetic data based on a large language model according to claim 2, characterized in that: The quality data set of the gene data to be analyzed specifically includes the sequencing error rate of the gene data to be analyzed, the number of repeated sequences of the gene data to be analyzed, and the sequence depth of the gene data to be analyzed.

4. The method for analyzing genetic data based on a large language model according to claim 2, characterized in that: The specific analysis process of the quality assessment value of the gene data to be analyzed is as follows: ； In the formula, is the quality assessment value of the gene data to be analyzed, is the sequencing error rate of the gene data to be analyzed, is the number of repeated sequences in the gene data to be analyzed, is the sequence depth of the gene data to be analyzed, is the compensation factor for the sequencing error rate of the gene data to be analyzed. is the compensation factor for the number of repeated sequences in the gene data to be analyzed. is the compensation factor of the set depth of the gene data sequence to be analyzed, and e is a natural constant.

5. The method for analyzing genetic data based on a large language model according to claim 1, characterized in that: The specific analysis process of determining the gene data gene sequence segmentation scheme is as follows: Obtaining a genetic data basic information data set, and obtaining a genetic data basic information characteristic value through comprehensive analysis based on the obtained genetic data basic information data set, and using the genetic data basic information characteristic value as an analysis basis for determining a genetic data gene sequence segmentation scheme; The basic information characteristic value of the gene data is stored as a designated label, and the designated label is compared with the gene data gene sequence segmentation scheme corresponding to each designated label stored in the database to obtain the gene data gene sequence segmentation scheme corresponding to the designated label.

6. The method for analyzing genetic data based on a large language model according to claim 5, characterized in that: The basic information data set of gene data specifically includes the number of effective gene data reads, the percentage of gene data N, the GC content of gene data, and the sequence depth of gene data.

7. The method for analyzing genetic data based on a large language model according to claim 1, characterized in that: The gene data analysis accuracy feature value is determined based on the gene data association feature data set. The specific analysis process is as follows: Based on the genetic data association feature data set, a comprehensive analysis is performed to obtain the genetic data analysis accuracy feature value, which is used as the analysis basis for determining the genetic data analysis accuracy score.

8. The method for analyzing genetic data based on a large language model according to claim 1, characterized in that: The gene data analysis accuracy score is determined based on the gene data analysis accuracy feature value, and the specific analysis process is as follows: Compare the genetic data analysis accuracy characteristic value with each genetic data analysis accuracy reference characteristic value stored in the database to obtain the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value, record the minimum value of the absolute value of the difference between the genetic data analysis accuracy characteristic value and each genetic data analysis accuracy reference characteristic value as the genetic data analysis accuracy characteristic deviation value, import the genetic data analysis accuracy characteristic deviation value into the genetic data analysis accuracy similarity model, and obtain the genetic data analysis accuracy similarity characteristic value; The genetic data analysis accuracy similarity feature value is stored as a designated label, and the designated label is compared with the genetic data analysis accuracy score corresponding to each designated label stored in the database to obtain the genetic data analysis accuracy score corresponding to the designated label.

9. The method for analyzing genetic data based on a large language model according to claim 1, characterized in that: The specific analysis process of judging whether the gene data analysis result is qualified is as follows: comparing the genetic data analysis accuracy score with the genetic data analysis accuracy demarcation score stored in a database; If the genetic data analysis accuracy score is higher than or equal to the genetic data analysis accuracy definition score, the genetic data analysis result corresponding to the genetic data analysis accuracy score is qualified; If the genetic data analysis accuracy score is lower than the genetic data analysis accuracy definition score, the genetic data analysis result corresponding to the genetic data analysis accuracy score is unqualified and the genetic data to be analyzed needs to be re-analyzed.

Images